chore(release): update package version to 1.0.86

Reviewed-on: #56

.gitea/workflows/release.yml

@@ -63,6 +63,8 @@ jobs:
       - name: Build release binaries
         run: |
           VERSION="${{ github.event.inputs.tag || github.ref_name }}"
+          BUILD_DATE="$(date -u +"%Y-%m-%d %H:%M:%S UTC")"
+          LDFLAGS="-X 'git.warky.dev/wdevs/relspecgo/pkg/buildinfo.Version=${VERSION}' -X 'git.warky.dev/wdevs/relspecgo/pkg/buildinfo.BuildDate=${BUILD_DATE}'"
           for target in "linux/amd64" "linux/arm64" "darwin/amd64" "darwin/arm64" "windows/amd64"; do
             GOOS="${target%/*}"
             GOARCH="${target#*/}"
@@ -71,11 +73,18 @@ jobs:
             NAME="relspec-${GOOS}-${GOARCH}${EXT}"
             GOOS="$GOOS" GOARCH="$GOARCH" go build \
               -trimpath \
-              -ldflags "-X git.warky.dev/wdevs/relspecgo/cmd/relspec.version=${VERSION}" \
+              -ldflags "$LDFLAGS" \
               -o "$NAME" ./cmd/relspec
             echo "Built $NAME"
           done
 
+      - name: Build Windows installer
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y nsis
+          VERSION="${{ github.event.inputs.tag || github.ref_name }}"
+          makensis -DVERSION="${VERSION#v}" -DEXE="$PWD/relspec-windows-amd64.exe" -DOUT="$PWD/relspec-setup-windows-amd64.exe" windows/installer.nsi
+
       - name: Create release and upload assets
         run: |
           TAG="${{ github.event.inputs.tag || github.ref_name }}"
@@ -234,11 +243,12 @@ jobs:
         run: |
           VERSION="${{ github.event.inputs.tag || github.ref_name }}"
           PKGVER="${VERSION#v}"
+          BUILD_DATE="$(date -u +"%Y-%m-%d %H:%M:%S UTC")"
 
           for GOARCH in amd64 arm64; do
             GOOS=linux GOARCH=$GOARCH go build \
               -trimpath \
-              -ldflags "-X git.warky.dev/wdevs/relspecgo/cmd/relspec.version=${PKGVER}" \
+              -ldflags "-X 'git.warky.dev/wdevs/relspecgo/pkg/buildinfo.Version=v${PKGVER}' -X 'git.warky.dev/wdevs/relspecgo/pkg/buildinfo.BuildDate=${BUILD_DATE}'" \
               -o relspec ./cmd/relspec
 
             PKGDIR="relspec_${PKGVER}_${GOARCH}"
@@ -309,7 +319,13 @@ jobs:
             rockylinux:9 \
             bash -lc "
               set -euo pipefail
-              dnf install -y rpm-build git &&
+              # Avoid transient/out-of-sync mirrors while bootstrapping the RPM builder.
+              sed -i \
+                -e 's|^mirrorlist=|#mirrorlist=|' \
+                -e 's|^#baseurl=http://dl.rockylinux.org|baseurl=https://dl.rockylinux.org|' \
+                /etc/yum.repos.d/rocky*.repo
+              dnf clean all
+              dnf -y --setopt=timeout=60 --setopt=retries=5 install rpm-build git
               curl -fsSL https://go.dev/dl/go\${GO_VER}.linux-amd64.tar.gz | tar -C /usr/local -xz &&
               export PATH=\$PATH:/usr/local/go/bin &&
               mkdir -p ~/rpmbuild/{BUILD,BUILDROOT,RPMS,SOURCES,SPECS,SRPMS} &&

CONTRIBUTING.md

@@ -1,145 +1,91 @@
 # Contributing to RelSpec
 
-Thank you for your interest in contributing to RelSpec.
+## Setup
 
-## Development Setup
+- Go 1.25+ (see `go.mod`), Git
+- Optional: golangci-lint, Docker/Podman (PostgreSQL integration tests)
 
-### Prerequisites
-- Go 1.21 or higher
-- Git
-- (Optional) golangci-lint for linting
-- (Optional) Docker for database testing
-
-### Getting Started
-
-1. Clone the repository:
 ```bash
-git clone https://github.com/wdevs/relspecgo.git
+git clone git@git.warky.dev:wdevs/relspecgo.git
 cd relspecgo
+make deps
+make build        # outputs build/relspec
 ```
 
-2. Install dependencies:
-```bash
-go mod download
-```
+## Make Targets
 
-3. Run tests:
-```bash
-go test ./...
-```
+| Target                  | Purpose                                              |
+|-------------------------|------------------------------------------------------|
+| `make test`             | Unit tests (race detection, coverage)                |
+| `make test-integration` | Integration tests (needs `RELSPEC_TEST_PG_CONN`)     |
+| `make docker-test`      | PostgreSQL integration tests via Docker/Podman       |
+| `make lint`             | golangci-lint                                        |
+| `make fmt` / `fmt-check`| gofumpt + goimports                                  |
+| `make check`            | vet, fmt-check, staticcheck, govulncheck             |
+| `make coverage`         | Coverage report                                      |
 
-4. Build the project:
-```bash
-go build -o relspec ./cmd/relspec
-```
+Single test: `go test -run TestName ./pkg/readers/dbml`
 
-## Project Structure
+## Layout
 
 ```
-relspecgo/
-├── cmd/              # CLI application entry point
-├── pkg/
-│   ├── readers/      # Input format readers (XML, JSON, DCTX, DB, GORM, Bun)
-│   ├── writers/      # Output format writers (GORM, Bun, JSON, YAML)
-│   ├── models/       # Internal data models for relations
-│   └── transform/    # Transformation and validation logic
-├── examples/         # Usage examples and sample files
-├── tests/            # Integration tests
-└── .claude/          # Claude Code configuration and commands
+cmd/relspec/      CLI commands (convert, diff, merge, split, edit, inspect, job, ...)
+pkg/models/       Core model: Database > Schema > Table > Column/Constraint/Index/Relationship
+pkg/readers/<fmt> One reader per format
+pkg/writers/<fmt> One writer per format
+pkg/diff, merge, inspector, jobs, transform, ui, pgsql, sqltypes ...
+examples/         Sample files
+tests/            Integration tests and assets
+docs/             Feature docs
 ```
 
-## Adding New Readers
+## Adding a Reader
 
-To add a new input format reader:
+1. Create `pkg/readers/<format>/reader.go` with `NewReader(options *readers.ReaderOptions)`.
+2. Implement `readers.Reader`: `ReadDatabase`, `ReadSchema`, `ReadTable`.
+3. Add `reader_test.go` in the same package.
+4. Register the format in the CLI switches (`cmd/relspec/convert.go`, `diff.go`, `edit.go`, ...).
+5. Add a `README.md` in the reader directory.
 
-1. Create a new file in `pkg/readers/` (e.g., `myformat_reader.go`)
-2. Implement the `Reader` interface:
-```go
-type Reader interface {
-    Read(source string) (*models.Schema, error)
-}
-```
-3. Add tests in `pkg/readers/myformat_reader_test.go`
-4. Register the reader in the CLI
+## Adding a Writer
 
-## Adding New Writers
+1. Create `pkg/writers/<format>/writer.go` with `NewWriter(options *writers.WriterOptions)`.
+2. Implement `writers.Writer`: `WriteDatabase`, `WriteSchema`, `WriteTable`.
+3. Add `writer_test.go` in the same package.
+4. Register the format in the CLI switches.
+5. Add a `README.md` in the writer directory.
 
-To add a new output format writer:
+## Code Rules
 
-1. Create a new file in `pkg/writers/` (e.g., `myformat_writer.go`)
-2. Implement the `Writer` interface:
-```go
-type Writer interface {
-    Write(schema *models.Schema, destination string) error
-}
-```
-3. Add tests in `pkg/writers/myformat_writer_test.go`
-4. Register the writer in the CLI
-
-## Code Style
-
-- Follow standard Go conventions
-- Use `gofmt` for formatting
-- Run `go vet` to check for issues
-- Use meaningful variable and function names
-- Add comments for exported functions and types
+- Format with gofumpt/goimports (`make fmt`); `make check` must pass.
+- Iterate `Table.Columns`, `Constraints`, `Indexes`, `Relationships` in sorted order (maps are unordered; output must be deterministic).
+- Every writer stamps `buildinfo.GeneratedComment()` in its file header.
+- Comment exported functions and types.
 
 ## Testing
 
-- Write unit tests for all new functionality
-- Aim for >80% code coverage
-- Use table-driven tests where appropriate
-- Include both positive and negative test cases
+- Tests live in the same package as the code.
+- Table-driven tests; cover positive and negative cases.
+- Reuse existing test data in `tests/` and `examples/` before adding new data.
+- Tests run with `-race`.
 
-### Running Tests
+## Commits
 
-```bash
-# All tests
-go test ./...
+- Format: `type(scope): description`
+- Types: `feat`, `fix`, `docs`, `test`, `refactor`, `chore`, `ci`
+- Keep commits focused. Reference issues where applicable.
 
-# With coverage
-go test -cover ./...
+## Pull Requests
 
-# Verbose output
-go test -v ./...
+1. Branch from `master`.
+2. Add tests; `make test` and `make check` pass.
+3. Update docs/README if behaviour changes.
+4. Open a PR with a clear description.
 
-# Specific package
-go test ./pkg/readers/...
-```
+## Security
 
-## Committing Changes
-
-- Write clear, descriptive commit messages
-- Follow conventional commits format: `type(scope): description`
-  - Types: feat, fix, docs, test, refactor, chore
-  - Example: `feat(readers): add PostgreSQL support`
-- Keep commits focused and atomic
-- Reference issues in commit messages when applicable
-
-## Pull Request Process
-
-1. Create a feature branch from `master`
-2. Make your changes
-3. Add tests for new functionality
-4. Ensure all tests pass
-5. Update documentation if needed
-6. Submit a pull request with a clear description
-
-## Claude Code Commands
-
-This project includes Claude Code slash commands for common tasks:
-
-- `/test` - Run all tests
-- `/build` - Build the binary
-- `/lint` - Run linters
-- `/coverage` - Generate coverage report
-
-## Questions or Issues?
-
-- Open an issue for bugs or feature requests
-- Start a discussion for questions or ideas
-- Check existing issues before creating new ones
+Do not report vulnerabilities in public issues. See [SECURITY.md](SECURITY.md).
 
 ## License
 
-By contributing to RelSpec, you agree that your contributions will be licensed under the Apache License 2.0.
+Contributions are licensed under the Apache License 2.0.

Makefile

@@ -1,4 +1,4 @@
-.PHONY: all build test test-unit test-integration lint coverage clean install help docker-up docker-down docker-test docker-test-integration start stop release release-version godoc vet fmt fmt-check staticcheck govulncheck check
+.PHONY: installer-windows all build test test-unit test-integration lint coverage clean install help docker-up docker-down docker-test docker-test-integration start stop release release-version rerelease godoc vet fmt fmt-check staticcheck govulncheck check
 
 # Binary name
 BINARY_NAME=relspec
@@ -20,9 +20,11 @@ STATICCHECK   = go run honnef.co/go/tools/cmd/staticcheck@latest
 GOVULNCHECK   = go run golang.org/x/vuln/cmd/govulncheck@latest
 
 # Version information
-VERSION := $(shell git describe --tags --always --dirty 2>/dev/null || echo "dev")
+# VERSION file is the single source of truth (bumped by `make release-version`).
+# Falls back to git describe for ad-hoc builds where VERSION hasn't been committed yet.
+VERSION := $(shell [ -f VERSION ] && echo "v$$(cat VERSION | tr -d '[:space:]')" || git describe --tags --always --dirty 2>/dev/null || echo "dev")
 BUILD_DATE := $(shell date -u +"%Y-%m-%d %H:%M:%S UTC")
-LDFLAGS := -X 'main.version=$(VERSION)' -X 'main.buildDate=$(BUILD_DATE)'
+LDFLAGS := -X 'git.warky.dev/wdevs/relspecgo/pkg/buildinfo.Version=$(VERSION)' -X 'git.warky.dev/wdevs/relspecgo/pkg/buildinfo.BuildDate=$(BUILD_DATE)'
 
 # Auto-detect container runtime (Docker or Podman)
 CONTAINER_RUNTIME := $(shell \
@@ -207,7 +209,7 @@ docker-test-integration: docker-up ## Start DB and run integration tests
 		$(GOTEST) -v ./pkg/readers/pgsql/ -count=1 || (make docker-down && exit 1)
 	@make docker-down
 
-release: ## Create and push a new release tag (auto-increments patch version)
+release: lint fmt-check test build ## Run lint, format check, tests, build, then create and push a new release tag
 	@echo "Creating new release..."
 	@latest_tag=$$(git describe --tags --abbrev=0 2>/dev/null || echo ""); \
 	if [ -z "$$latest_tag" ]; then \
@@ -223,6 +225,15 @@ release: ## Create and push a new release tag (auto-increments patch version)
 		echo "Creating new release: $$version"; \
 		commit_logs=$$(git log "$${latest_tag}..HEAD" --pretty=format:"- %s" --no-merges); \
 	fi; \
+	PKGVER=$${version#v}; \
+	echo "Updating package metadata to $$PKGVER..."; \
+	echo "$$PKGVER" > VERSION; \
+	sed -i "s/^pkgver=.*/pkgver=$$PKGVER/" linux/arch/PKGBUILD; \
+	sed -i "s/^Version:.*/Version:        $$PKGVER/" linux/centos/relspec.spec; \
+	git add VERSION linux/arch/PKGBUILD linux/centos/relspec.spec; \
+	if ! git diff --cached --quiet; then \
+		git commit -m "chore(release): update package version to $$PKGVER"; \
+	fi; \
 	if [ -z "$$commit_logs" ]; then \
 		tag_message="Release $$version"; \
 	else \
@@ -232,21 +243,35 @@ release: ## Create and push a new release tag (auto-increments patch version)
 	git push origin "$$version"; \
 	echo "Tag $$version created and pushed to remote repository."
 
-release-version: ## Auto-increment patch version, update package files, commit, tag, and push
-	@CURRENT=$$(git describe --tags --abbrev=0 2>/dev/null || echo "v0.0.0"); \
-	MAJOR=$$(echo $$CURRENT | sed 's/v\([0-9]*\)\.\([0-9]*\)\.\([0-9]*\).*/\1/'); \
-	MINOR=$$(echo $$CURRENT | sed 's/v\([0-9]*\)\.\([0-9]*\)\.\([0-9]*\).*/\2/'); \
-	PATCH=$$(echo $$CURRENT | sed 's/v\([0-9]*\)\.\([0-9]*\)\.\([0-9]*\).*/\3/'); \
-	NEXT="v$$MAJOR.$$MINOR.$$((PATCH + 1))"; \
+release-version: lint fmt-check ## Run lint and format check, then auto-increment patch version, update VERSION/package files, commit, tag, and push
+	@CURRENT=$$([ -f VERSION ] && cat VERSION | tr -d '[:space:]' || (git describe --tags --abbrev=0 2>/dev/null | sed 's/^v//') || echo "0.0.0"); \
+	MAJOR=$$(echo $$CURRENT | cut -d. -f1); \
+	MINOR=$$(echo $$CURRENT | cut -d. -f2); \
+	PATCH=$$(echo $$CURRENT | cut -d. -f3); \
 	PKGVER="$$MAJOR.$$MINOR.$$((PATCH + 1))"; \
-	echo "Current: $$CURRENT → Next: $$NEXT"; \
+	NEXT="v$$PKGVER"; \
+	echo "Current: v$$CURRENT → Next: $$NEXT"; \
+	echo "$$PKGVER" > VERSION; \
 	sed -i "s/^pkgver=.*/pkgver=$$PKGVER/" linux/arch/PKGBUILD; \
 	sed -i "s/^Version:.*/Version:        $$PKGVER/" linux/centos/relspec.spec; \
-	git add linux/arch/PKGBUILD linux/centos/relspec.spec; \
+	git add VERSION linux/arch/PKGBUILD linux/centos/relspec.spec; \
 	git commit -m "chore(release): update package version to $$PKGVER"; \
 	git tag -a "$$NEXT" -m "Release $$NEXT"; \
 	git push origin HEAD "$$NEXT"; \
 	echo "Pushed $$NEXT — release workflow triggered"
 
+rerelease: lint fmt-check ## Move the latest tag to HEAD and force push it
+	@TAG=$$(git describe --tags --abbrev=0 2>/dev/null); \
+	if [ -z "$$TAG" ]; then echo "No existing tags found"; exit 1; fi; \
+	echo "Moving $$TAG to $$(git rev-parse --short HEAD)"; \
+	git tag -f -a "$$TAG" -m "Release $$TAG" HEAD; \
+	git push --force origin "$$TAG"; \
+	echo "Pushed $$TAG — release workflow triggered"
+
 help: ## Display this help screen
 	@grep -E '^[a-zA-Z_-]+:.*?## .*$$' $(MAKEFILE_LIST) | sort | awk 'BEGIN {FS = ":.*?## "}; {printf "\033[36m%-20s\033[0m %s\n", $$1, $$2}'
+
+installer-windows: ## Build the Windows binary and NSIS installer (requires makensis)
+	@echo "Building Windows installer..."
+	GOOS=windows GOARCH=amd64 $(GOBUILD) -trimpath -ldflags "$(LDFLAGS)" -o $(BUILD_DIR)/relspec-windows-amd64.exe ./cmd/relspec
+	makensis -DVERSION=$$(cat VERSION | tr -d '[:space:]') -DEXE=$(CURDIR)/$(BUILD_DIR)/relspec-windows-amd64.exe -DOUT=$(CURDIR)/$(BUILD_DIR)/relspec-setup-windows-amd64.exe windows/installer.nsi

README.md

@@ -16,12 +16,18 @@
 go install -v git.warky.dev/wdevs/relspecgo/cmd/relspec@latest
 ```
 
+Windows: download `relspec-setup-windows-amd64.exe` from the
+[latest release](https://git.warky.dev/wdevs/relspecgo/releases/latest)
+(installs to Program Files and adds `relspec` to `PATH`). See [windows/README.md](windows/README.md).
+
 ## Supported Formats
 
 | Direction | Formats |
 |-----------|---------|
-| **Readers** | `bun` `dbml` `dctx` `drawdb` `drizzle` `gorm` `graphql` `json` `mssql` `pgsql` `prisma` `sqldir` `sqlite` `typeorm` `yaml` |
-| **Writers** | `bun` `dbml` `dctx` `drawdb` `drizzle` `gorm` `graphql` `json` `mssql` `pgsql` `prisma` `sqlexec` `sqlite` `template` `typeorm` `yaml` |
+| **Readers** | `bun` `dbml` `dctx` `drawdb` `drizzle` `gorm` `graphql` `json` `mssql` `mysql` `pgsql` `prisma` `sqldir` `sqlite` `typeorm` `yaml` |
+| **Writers** | `bun` `dbml` `dctx` `drawdb` `drizzle` `gorm` `graphql` `json` `mssql` `mysql` `pgsql` `prisma` `sqlexec` `sqlite` `template` `typeorm` `yaml` |
+
+See [docs/FORMAT_EXAMPLES.md](docs/FORMAT_EXAMPLES.md) for usage examples covering every format.
 
 ## Commands
 
@@ -40,6 +46,26 @@ relspec convert --from pgsql --from-conn "postgres://..." --to sqlite --to-path
 
 # Multiple input files merged
 relspec convert --from json --from-list "a.json,b.json" --to yaml --to-path merged.yaml
+
+# Watch mode: regenerate whenever the source file(s) change (Ctrl-C to stop)
+relspec convert --from dbml --from-path schema.dbml --to gorm --to-path models/ --package models --watch
+```
+
+`--watch` works with `--from-path` and `--from-list` (not live database
+connections or `--dry-run`). Source files are polled every `--watch-interval`
+(default 500ms), a directory source is watched recursively, and the output path
+is ignored so generating into the source tree does not loop. Conversion errors
+are printed and watching continues.
+
+### `batch` — Convert many inputs in one run
+
+Converts each input independently (one output per input, unlike `--from-list`
+which merges). `--input` takes paths or globs; outputs go to `--to-dir`.
+Use `--keep-going` to continue past failures (exit is still non-zero) and
+`--dry-run` to validate without writing. For named workflows see `relspec job run`.
+
+```bash
+relspec batch --from dbml --input "schemas/*.dbml" --to json --to-dir out/
 ```
 
 PostgreSQL connections opened by relspec set `application_name` by default to
@@ -126,7 +152,7 @@ relspec job run build-schema
 version: 1
 jobs:
   build-schema:
-    command: convert            # closed allow-list: convert | merge | scripts-list
+    command: convert            # closed allow-list: convert | merge | split | scripts-list | scripts-exec | templ | inspect | diff
     description: Merge the DBML sources and emit PostgreSQL DDL
     inputs:
       - path: schema/core.dbml
@@ -147,7 +173,9 @@ resolved relative to the job file and may not escape it, and remote database
 credentials are referenced by environment-variable name (`conn_env:`) and
 redacted from logs. The whole plan — unknown commands/formats, duplicate job
 names, missing inputs, path traversal, dependency cycles — is validated before
-any job runs. See [docs/JOB_FILES.md](docs/JOB_FILES.md).
+any job runs. Path-like fields also support `${NAME}` environment-variable
+references, which are expanded and safety-checked during pre-flight. See
+[docs/JOB_FILES.md](docs/JOB_FILES.md).
 
 ### `edit` — Interactive TUI editor
 
@@ -170,6 +198,16 @@ relspec edit --from pgsql --from-conn "postgres://user:pass@localhost/mydb" \
 <img src="./assets/image/screenshots/edit_column.jpg">
 </p>
 
+### `update` — Check for a newer release
+
+```bash
+relspec update            # check, prompt, then download + run installer (Windows)
+relspec update --check    # report only
+relspec update --yes      # skip the prompt
+```
+
+Non-Windows platforms print the release URL. `dev` and commit-hash builds are never reported as outdated.
+
 ## Development
 
 **Prerequisites:** Go 1.24.0+
@@ -180,6 +218,7 @@ make test      # race detection + coverage
 make lint      # requires golangci-lint
 make coverage  # → coverage.html
 make install   # → $GOPATH/bin
+make installer-windows  # → build/relspec-setup-windows-amd64.exe (needs makensis)
 ```
 
 ## Project Structure
@@ -194,6 +233,8 @@ pkg/merge/            Schema merging
 pkg/models/           Internal data models
 pkg/transform/        Transformation logic
 pkg/pgsql/            PostgreSQL utilities
+pkg/updatecheck/      Latest-release lookup and version comparison
+windows/              NSIS installer script
 pkg/sqltypes/         Nullable SQL types for generated/hand-written models (see below)
 ```
 
@@ -214,6 +255,23 @@ see [`bun`'s `--array-nullable`](./pkg/writers/bun/README.md#nullablearrays)
 flag for nullable-array handling. The `SqlXxxArray` wrapper types remain
 available in `pkg/sqltypes` and are still used by the `gorm` writer.
 
+#### Custom type mapping
+
+Override the built-in SQL → Go mapping of the `bun` and `gorm` writers with the
+repeatable `--type-map sqltype=gotype` flag:
+
+```bash
+relspec convert --from pgsql --from-conn "$DSN" --to gorm --to-path models.go \
+  --type-map uuid=string --type-map jsonb=json.RawMessage
+```
+
+SQL type names are matched case-insensitively on the base type (modifiers such
+as `(10,2)` are ignored; aliases like `int4` resolve to `integer`). NOT NULL
+columns use the Go type verbatim, nullable columns get a `*` prefix (unless the
+type is already a pointer, slice, map or `any`), and arrays become `[]gotype`.
+Unmapped types keep their defaults. The flag does not add imports: use types
+that need none, or add the import afterwards (e.g. with `goimports`).
+
 ## Contributing
 
 1. Register or sign in with GitHub at [git.warky.dev](https://git.warky.dev)

SECURITY.md

@@ -0,0 +1,12 @@
+# Security Policy
+
+## Supported versions
+Security fixes are released for the latest minor version of RelSpec.
+
+## Reporting a vulnerability
+Please do not open a public issue for security problems.
+
+Report privately by email: warkydevs@gmail.com
+
+Reports are reviewed on a best-effort basis. No response time or fix timeline is
+guaranteed. Reporters may be credited in the release notes if they wish.

TODO.md

@@ -4,20 +4,25 @@
 
 - [✔️] **Database Inspector**
   - [✔️] PostgreSQL driver (reader + writer)
-  - [ ] MySQL driver
+  - [ ] MySQL driver (only MariaDB datatype conversion in pkg/mariadb, no reader/writer)
   - [✔️] SQLite driver (reader + writer with automatic schema flattening)
-  - [ ] MSSQL driver
+  - [✔️] MSSQL driver (reader + writer, generated and identity columns)
   - [✔️] Foreign key detection
   - [✔️] Index extraction
   - [✔️] .sql file generation (PostgreSQL, SQLite)
 - [✔️] .dbml: Database Markup Language (DBML) for textual schema representation.
 - [✔️] Prisma schema support (PSL format) .prisma
+- [ ] Sequelize (Typescript/Javascript)  (Use templates, 💲 Someone can do this, not me)
 - [✔️] Drizzle ORM support .ts (TypeScript / JavaScript) (Mr. Edd wanted to move from Prisma to Drizzle. If you are bugs, you are welcome to do pull requests or issues)
 - [☠️] Entity Framework (.NET) model .edmx (Fuck no, EDMX files were bloated, verbose XML nightmares—hard to merge, error-prone, and a pain in teams. Microsoft wisely ditched them in EF Core for code-first. Classic overkill from old MS era.)
 - [✔️] TypeORM support
-- [] .hbm.xml / schema.xml: Hibernate/Propel mappings (Java/PHP) (💲 Someone can do this, not me)
-- [ ] Django models.py (Python classes), Sequelize migrations (JS) (💲 Someone can do this, not me)
-- [] .avsc: Avro schema (JSON format for data serialization) (💲 Someone can do this, not me)
+- [ ] .hbm.xml / schema.xml: Hibernate/Propel mappings (Java/PHP) (Use templates, 💲 Someone can do this, not me)
+- [ ] Django models.py (Python classes), Sequelize migrations (JS) (Use templates, 💲 Someone can do this, not me)
+- [ ] SQLAlchemy, Python (Use templates, 💲 Someone can do this, not me)
+- [ ] .avsc: Avro schema (JSON format for data serialization) (Use templates, 💲 Someone can do this, not me)
+- [ ] Rails schema in db/schema.rb (Use templates, 💲 Someone can do this, not me)
+- [ ] Laravel migration (Use templates, 💲 Someone can do this, not me)
+- [ ] Hibernate / JPA (Use templates, 💲 Someone can do this, not me)
 - [✔️] GraphQL schema generation
 
 ## UI
@@ -39,13 +44,19 @@
 
 ## Advanced Features
 
-- [ ] Dry-run mode for validation
-- [x] Diff tool for comparing specifications
-- [ ] Migration script generation
+- [ ] Dry-run mode for validation (only `job run --dry-run`; not on convert/merge/split)
+- [✔️] Diff tool for comparing specifications
+- [✔️] Migration script generation (PostgreSQL diff, live database by default for direct output)
 - [ ] Custom type mapping configuration
-- [ ] Batch processing support
+- [ ] Batch processing support (partial: job files run named workflows via `relspec job run`)
 - [ ] Watch mode for auto-regeneration
 
+## Distribution
+
+- [ ] NSIS Windows installer
+  - [ ] Check https://git.warky.dev/wdevs/relspecgo/releases for new releases and prompt to update
+- [ ] CI pipeline to build and publish the Windows installer
+
 ## Future Considerations
 
 - [ ] Web UI for visual editing

VERSION

@@ -0,0 +1 @@
+1.0.86

cmd/relspec/batch.go

@@ -0,0 +1,214 @@
+package main
+
+import (
+	"fmt"
+	"os"
+	"path/filepath"
+	"sort"
+	"strings"
+
+	"github.com/spf13/cobra"
+)
+
+var (
+	batchSourceType      string
+	batchInputs          []string
+	batchTargetType      string
+	batchTargetDir       string
+	batchPackageName     string
+	batchSchemaFilter    string
+	batchFlattenSchema   bool
+	batchNullableTypes   string
+	batchNullableArrays  string
+	batchContinueOnError bool
+	batchKeepGoing       bool
+	batchDryRun          bool
+)
+
+var batchCmd = &cobra.Command{
+	Use:   "batch",
+	Short: "Convert many input files to a target format in one run",
+	Long: `Convert each input file independently to the target format.
+
+Unlike 'convert --from-list', which merges all inputs into one output, batch
+mode writes one output per input into --to-dir. The output is named after the
+input file (without its extension). Directory-style targets (gorm, bun,
+drizzle) get a sub-directory per input.
+
+Inputs are given with --input, which accepts file paths and glob patterns and
+may be repeated or comma-separated. Inputs are processed in sorted order and
+duplicates are removed. The command exits non-zero if any input fails.
+
+For named, multi-step workflows use 'relspec job run' instead.
+
+Examples:
+  # Convert every DBML file in a directory to JSON
+  relspec batch --from dbml --input "schemas/*.dbml" --to json --to-dir out/
+
+  # Convert specific files to GORM models, one package directory per input
+  relspec batch --from json --input a.json,b.json \
+                --to gorm --to-dir models/ --package models
+
+  # Validate everything first, writing nothing
+  relspec batch --from yaml --input "specs/*.yaml" --to pgsql --to-dir sql/ --dry-run
+
+  # Report all failures instead of stopping at the first
+  relspec batch --from json --input "*.json" --to yaml --to-dir out/ --keep-going`,
+	RunE: runBatch,
+}
+
+func init() {
+	batchCmd.Flags().StringVar(&batchSourceType, "from", "", "Source format for every input (dbml, dctx, drawdb, graphql, json, yaml, gorm, bun, drizzle, prisma, typeorm, sqlite)")
+	batchCmd.Flags().StringSliceVar(&batchInputs, "input", nil, "Input file path or glob pattern (repeatable, comma-separated)")
+	batchCmd.Flags().StringVar(&batchTargetType, "to", "", "Target format")
+	batchCmd.Flags().StringVar(&batchTargetDir, "to-dir", "", "Output directory; one output per input is written here")
+	batchCmd.Flags().StringVar(&batchPackageName, "package", "", "Package name (for code generation formats like gorm/bun)")
+	batchCmd.Flags().StringVar(&batchSchemaFilter, "schema", "", "Filter to a specific schema by name")
+	batchCmd.Flags().BoolVar(&batchFlattenSchema, "flatten-schema", false, "Flatten schema.table names to schema_table")
+	batchCmd.Flags().StringVar(&batchNullableTypes, "types", "", "Nullable type package for code-gen writers (bun/gorm)")
+	batchCmd.Flags().StringVar(&batchNullableArrays, "array-nullable", "", "Nullable array representation for the Bun writer")
+	batchCmd.Flags().BoolVar(&batchContinueOnError, "continue-on-error", false, "Prepend \\set ON_ERROR_STOP off to generated SQL (pgsql output only)")
+	batchCmd.Flags().BoolVar(&batchKeepGoing, "keep-going", false, "Process remaining inputs after a failure; still exits non-zero")
+	batchCmd.Flags().BoolVar(&batchDryRun, "dry-run", false, "Read and validate every input and print the plan without writing any output")
+
+	for _, f := range []string{"from", "input", "to", "to-dir"} {
+		if err := batchCmd.MarkFlagRequired(f); err != nil {
+			fmt.Fprintf(os.Stderr, "Error marking %s flag as required: %v\n", f, err)
+		}
+	}
+}
+
+// batchDirTargets are writers that emit a directory of files rather than one file.
+var batchDirTargets = map[string]bool{"gorm": true, "bun": true, "drizzle": true}
+
+// batchExtensions maps single-file target formats to their output extension.
+var batchExtensions = map[string]string{
+	"dbml": ".dbml", "dctx": ".dctx", "drawdb": ".ddb", "json": ".json",
+	"yaml": ".yaml", "yml": ".yaml", "pgsql": ".sql", "postgres": ".sql",
+	"postgresql": ".sql", "sql": ".sql", "mssql": ".sql", "sqlserver": ".sql",
+	"mssql2016": ".sql", "mssql2017": ".sql", "mssql2019": ".sql", "mssql2022": ".sql",
+	"sqlite": ".sql", "sqlite3": ".sql", "prisma": ".prisma", "typeorm": ".ts",
+	"graphql": ".graphql", "gql": ".graphql",
+}
+
+// expandBatchInputs resolves paths and glob patterns into a sorted,
+// de-duplicated file list. A pattern that matches nothing is an error.
+func expandBatchInputs(patterns []string) ([]string, error) {
+	seen := map[string]bool{}
+	var files []string
+	for _, p := range patterns {
+		p = strings.TrimSpace(p)
+		if p == "" {
+			continue
+		}
+		matches, err := filepath.Glob(p)
+		if err != nil {
+			return nil, fmt.Errorf("invalid pattern %q: %w", p, err)
+		}
+		if len(matches) == 0 {
+			return nil, fmt.Errorf("no files match %q", p)
+		}
+		for _, m := range matches {
+			if !seen[m] {
+				seen[m] = true
+				files = append(files, m)
+			}
+		}
+	}
+	if len(files) == 0 {
+		return nil, fmt.Errorf("no input files given")
+	}
+	sort.Strings(files)
+	return files, nil
+}
+
+// batchOutputPaths returns the output path for each input. It errors when two
+// inputs would collide on the same output name.
+func batchOutputPaths(files []string, targetType, dir string) ([]string, error) {
+	key := strings.ToLower(targetType)
+	ext := ""
+	if !batchDirTargets[key] {
+		var ok bool
+		if ext, ok = batchExtensions[key]; !ok {
+			return nil, fmt.Errorf("unsupported target format: %s", targetType)
+		}
+	}
+	outs := make([]string, len(files))
+	owner := map[string]string{}
+	for i, f := range files {
+		stem := strings.TrimSuffix(filepath.Base(f), filepath.Ext(f))
+		out := filepath.Join(dir, stem+ext)
+		if prev, dup := owner[out]; dup {
+			return nil, fmt.Errorf("inputs %s and %s would both write %s", prev, f, out)
+		}
+		owner[out] = f
+		outs[i] = out
+	}
+	return outs, nil
+}
+
+func runBatch(cmd *cobra.Command, args []string) error {
+	files, err := expandBatchInputs(batchInputs)
+	if err != nil {
+		return err
+	}
+	outs, err := batchOutputPaths(files, batchTargetType, batchTargetDir)
+	if err != nil {
+		return err
+	}
+
+	fmt.Fprintf(os.Stderr, "\n=== RelSpec Batch Converter ===\n")
+	fmt.Fprintf(os.Stderr, "Started at: %s\n", getCurrentTimestamp())
+	fmt.Fprintf(os.Stderr, "Inputs: %d file(s), %s -> %s\n\n", len(files), batchSourceType, batchTargetType)
+
+	out := outWriter(cmd)
+	if batchDryRun {
+		fmt.Fprintf(out, "RelSpec batch plan (dry run - nothing written):\n")
+	}
+
+	var failed []string
+	for i, f := range files {
+		fmt.Fprintf(os.Stderr, "[%d/%d] %s -> %s\n", i+1, len(files), f, outs[i])
+		if err := processBatchItem(cmd, f, outs[i]); err != nil {
+			fmt.Fprintf(os.Stderr, "      ✗ %v\n", err)
+			failed = append(failed, fmt.Sprintf("%s: %v", f, err))
+			if !batchKeepGoing {
+				break
+			}
+			continue
+		}
+		fmt.Fprintf(os.Stderr, "      ✓ done\n")
+	}
+
+	fmt.Fprintf(os.Stderr, "\n=== Batch Complete: %d ok, %d failed ===\n", len(files)-len(failed), len(failed))
+	if len(failed) > 0 {
+		return fmt.Errorf("batch finished with %d failure(s):\n  %s", len(failed), strings.Join(failed, "\n  "))
+	}
+	return nil
+}
+
+func processBatchItem(cmd *cobra.Command, in, outPath string) error {
+	db, err := readDatabaseForConvert(batchSourceType, in, "")
+	if err != nil {
+		return fmt.Errorf("failed to read source: %w", err)
+	}
+	finalizeCommentedRefs(db, stderrWarn)
+
+	if batchDryRun {
+		if err := validateWriteTarget(db, batchTargetType, batchPackageName, batchSchemaFilter, ""); err != nil {
+			return fmt.Errorf("dry run validation failed: %w", err)
+		}
+		w := outWriter(cmd)
+		fmt.Fprintf(w, "  %s -> %s (database '%s')\n", in, outPath, db.Name)
+		printDryRunPlan(w, db)
+		return nil
+	}
+
+	if err := os.MkdirAll(batchTargetDir, 0o755); err != nil {
+		return fmt.Errorf("failed to create output directory: %w", err)
+	}
+	if err := writeDatabase(db, batchTargetType, outPath, batchPackageName, batchSchemaFilter, batchFlattenSchema, batchNullableTypes, batchNullableArrays, batchContinueOnError, ""); err != nil {
+		return fmt.Errorf("failed to write target: %w", err)
+	}
+	return nil
+}

cmd/relspec/batch_test.go

@@ -0,0 +1,143 @@
+package main
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+)
+
+func saveBatchState(t *testing.T) {
+	t.Helper()
+	a, b, c, d, e, f, g := batchSourceType, batchInputs, batchTargetType, batchTargetDir, batchPackageName, batchKeepGoing, batchDryRun
+	t.Cleanup(func() {
+		batchSourceType, batchInputs, batchTargetType, batchTargetDir, batchPackageName, batchKeepGoing, batchDryRun = a, b, c, d, e, f, g
+	})
+}
+
+func TestRunBatch_ConvertsEachInput(t *testing.T) {
+	saveBatchState(t)
+	dir := t.TempDir()
+	writeTestJSON(t, filepath.Join(dir, "a.json"), []string{"users"})
+	writeTestJSON(t, filepath.Join(dir, "b.json"), []string{"posts"})
+	outDir := filepath.Join(dir, "out")
+
+	batchSourceType, batchTargetType, batchTargetDir = "json", "yaml", outDir
+	batchPackageName, batchKeepGoing, batchDryRun = "", false, false
+	batchInputs = []string{filepath.Join(dir, "*.json")}
+
+	cmd, _ := newDryRunCmd()
+	if err := runBatch(cmd, nil); err != nil {
+		t.Fatalf("batch: %v", err)
+	}
+	for _, name := range []string{"a.yaml", "b.yaml"} {
+		if _, err := os.Stat(filepath.Join(outDir, name)); err != nil {
+			t.Errorf("expected %s: %v", name, err)
+		}
+	}
+}
+
+func TestRunBatch_DryRunWritesNothing(t *testing.T) {
+	saveBatchState(t)
+	dir := t.TempDir()
+	writeTestJSON(t, filepath.Join(dir, "a.json"), []string{"users"})
+	outDir := filepath.Join(dir, "out")
+
+	batchSourceType, batchTargetType, batchTargetDir = "json", "yaml", outDir
+	batchPackageName, batchKeepGoing, batchDryRun = "", false, true
+	batchInputs = []string{filepath.Join(dir, "a.json")}
+
+	cmd, buf := newDryRunCmd()
+	if err := runBatch(cmd, nil); err != nil {
+		t.Fatalf("dry run: %v", err)
+	}
+	if _, err := os.Stat(outDir); !os.IsNotExist(err) {
+		t.Fatal("dry run must not create the output directory")
+	}
+	if !strings.Contains(buf.String(), "users") || !strings.Contains(buf.String(), "a.yaml") {
+		t.Errorf("plan incomplete:\n%s", buf.String())
+	}
+}
+
+func TestRunBatch_FailureHandling(t *testing.T) {
+	saveBatchState(t)
+	dir := t.TempDir()
+	writeTestJSON(t, filepath.Join(dir, "a.json"), []string{"users"})
+	if err := os.WriteFile(filepath.Join(dir, "b.json"), []byte("{not json"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	writeTestJSON(t, filepath.Join(dir, "c.json"), []string{"posts"})
+	outDir := filepath.Join(dir, "out")
+
+	batchSourceType, batchTargetType, batchTargetDir = "json", "yaml", outDir
+	batchPackageName, batchDryRun = "", false
+	batchInputs = []string{filepath.Join(dir, "*.json")}
+	cmd, _ := newDryRunCmd()
+
+	// Default: stop at first failure.
+	batchKeepGoing = false
+	err := runBatch(cmd, nil)
+	if err == nil || !strings.Contains(err.Error(), "b.json") {
+		t.Fatalf("expected failure naming b.json, got %v", err)
+	}
+	if _, statErr := os.Stat(filepath.Join(outDir, "c.yaml")); !os.IsNotExist(statErr) {
+		t.Error("c.json should not be processed without --keep-going")
+	}
+
+	// --keep-going: remaining inputs are processed, exit still fails.
+	batchKeepGoing = true
+	if err := runBatch(cmd, nil); err == nil {
+		t.Fatal("expected non-zero result with --keep-going")
+	}
+	if _, statErr := os.Stat(filepath.Join(outDir, "c.yaml")); statErr != nil {
+		t.Errorf("c.yaml should be written with --keep-going: %v", statErr)
+	}
+}
+
+func TestExpandBatchInputs(t *testing.T) {
+	dir := t.TempDir()
+	for _, n := range []string{"b.json", "a.json"} {
+		if err := os.WriteFile(filepath.Join(dir, n), []byte("{}"), 0o644); err != nil {
+			t.Fatal(err)
+		}
+	}
+	got, err := expandBatchInputs([]string{filepath.Join(dir, "*.json"), filepath.Join(dir, "a.json")})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(got) != 2 || filepath.Base(got[0]) != "a.json" || filepath.Base(got[1]) != "b.json" {
+		t.Errorf("want sorted deduped [a b], got %v", got)
+	}
+	if _, err := expandBatchInputs([]string{filepath.Join(dir, "*.nope")}); err == nil {
+		t.Error("unmatched pattern should error")
+	}
+}
+
+func TestBatchOutputPaths(t *testing.T) {
+	tests := []struct {
+		name    string
+		files   []string
+		target  string
+		want    []string
+		wantErr string
+	}{
+		{"file target", []string{"x/a.dbml"}, "json", []string{"out/a.json"}, ""},
+		{"dir target", []string{"x/a.json"}, "gorm", []string{"out/a"}, ""},
+		{"collision", []string{"x/a.json", "y/a.json"}, "yaml", nil, "both write"},
+		{"unsupported", []string{"a.json"}, "nope", nil, "unsupported target"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got, err := batchOutputPaths(tt.files, tt.target, "out")
+			if tt.wantErr != "" {
+				if err == nil || !strings.Contains(err.Error(), tt.wantErr) {
+					t.Fatalf("want error %q, got %v", tt.wantErr, err)
+				}
+				return
+			}
+			if err != nil || len(got) != len(tt.want) || got[0] != filepath.FromSlash(tt.want[0]) {
+				t.Fatalf("got %v, %v; want %v", got, err, tt.want)
+			}
+		})
+	}
+}

cmd/relspec/convert.go

@@ -3,6 +3,7 @@ package main
 import (
 	stdjson "encoding/json"
 	"fmt"
+	"io"
 	"os"
 	"strings"
 	"time"
@@ -21,6 +22,7 @@ import (
 	"git.warky.dev/wdevs/relspecgo/pkg/readers/graphql"
 	"git.warky.dev/wdevs/relspecgo/pkg/readers/json"
 	"git.warky.dev/wdevs/relspecgo/pkg/readers/mssql"
+	"git.warky.dev/wdevs/relspecgo/pkg/readers/mysql"
 	"git.warky.dev/wdevs/relspecgo/pkg/readers/pgsql"
 	"git.warky.dev/wdevs/relspecgo/pkg/readers/prisma"
 	"git.warky.dev/wdevs/relspecgo/pkg/readers/sqlite"
@@ -36,6 +38,7 @@ import (
 	wgraphql "git.warky.dev/wdevs/relspecgo/pkg/writers/graphql"
 	wjson "git.warky.dev/wdevs/relspecgo/pkg/writers/json"
 	wmssql "git.warky.dev/wdevs/relspecgo/pkg/writers/mssql"
+	wmysql "git.warky.dev/wdevs/relspecgo/pkg/writers/mysql"
 	wpgsql "git.warky.dev/wdevs/relspecgo/pkg/writers/pgsql"
 	wprisma "git.warky.dev/wdevs/relspecgo/pkg/writers/prisma"
 	wsqlite "git.warky.dev/wdevs/relspecgo/pkg/writers/sqlite"
@@ -57,6 +60,9 @@ var (
 	convertNullableArrays  string
 	convertContinueOnError bool
 	convertExtraFields     string
+	convertDryRun          bool
+	convertWatch           bool
+	convertWatchInterval   time.Duration
 )
 
 var convertCmd = &cobra.Command{
@@ -165,7 +171,11 @@ Examples:
 
   # Convert SQLite to PostgreSQL SQL
   relspec convert --from sqlite --from-path database.db \
-                  --to pgsql --to-path schema.sql`,
+                  --to pgsql --to-path schema.sql
+
+  # Regenerate GORM models every time the DBML file changes
+  relspec convert --from dbml --from-path schema.dbml \
+                  --to gorm --to-path models/ --package models --watch`,
 	RunE: runConvert,
 }
 
@@ -185,6 +195,11 @@ func init() {
 	convertCmd.Flags().BoolVar(&convertContinueOnError, "continue-on-error", false, "Prepend \\set ON_ERROR_STOP off to generated SQL so psql continues past errors (pgsql output only)")
 	convertCmd.Flags().StringVar(&convertExtraFields, "extra-fields", "", "Path to JSON file containing extra Bun model fields to inject (bun output only); fields support target_table, name, type, bun_tag, json_tag, comment")
 
+	convertCmd.Flags().BoolVar(&convertDryRun, "dry-run", false, "Read and validate the input and print the plan without writing any output")
+
+	convertCmd.Flags().BoolVar(&convertWatch, "watch", false, "Watch the source files (--from-path or --from-list) and regenerate the output whenever they change")
+	convertCmd.Flags().DurationVar(&convertWatchInterval, "watch-interval", 500*time.Millisecond, "Polling interval used by --watch")
+
 	err := convertCmd.MarkFlagRequired("from")
 	if err != nil {
 		fmt.Fprintf(os.Stderr, "Error marking from flag as required: %v\n", err)
@@ -200,6 +215,13 @@ func init() {
 }
 
 func runConvert(cmd *cobra.Command, args []string) error {
+	if convertWatch {
+		return runConvertWatch(cmd.Context(), os.Stderr, func() error { return runConvertOnce(cmd) })
+	}
+	return runConvertOnce(cmd)
+}
+
+func runConvertOnce(cmd *cobra.Command) error {
 	fmt.Fprintf(os.Stderr, "\n=== RelSpec Schema Converter ===\n")
 	fmt.Fprintf(os.Stderr, "Started at: %s\n\n", getCurrentTimestamp())
 
@@ -229,6 +251,7 @@ func runConvert(cmd *cobra.Command, args []string) error {
 	if err != nil {
 		return fmt.Errorf("failed to read source: %w", err)
 	}
+	finalizeCommentedRefs(db, stderrWarn)
 
 	fmt.Fprintf(os.Stderr, "      ✓ Successfully read database '%s'\n", db.Name)
 	fmt.Fprintf(os.Stderr, "      Found: %d schema(s)\n", len(db.Schemas))
@@ -239,6 +262,22 @@ func runConvert(cmd *cobra.Command, args []string) error {
 	}
 	fmt.Fprintf(os.Stderr, "      Found: %d table(s)\n\n", totalTables)
 
+	if convertDryRun {
+		if err := validateWriteTarget(db, convertTargetType, convertPackageName, convertSchemaFilter, convertExtraFields); err != nil {
+			return fmt.Errorf("dry run validation failed: %w", err)
+		}
+		out := outWriter(cmd)
+		fmt.Fprintf(out, "RelSpec convert plan (dry run - nothing written):\n")
+		fmt.Fprintf(out, "  Input:  %s database '%s'\n", convertSourceType, db.Name)
+		fmt.Fprintf(out, "  Output: %s -> %s\n", convertTargetType, convertTargetPath)
+		if convertSchemaFilter != "" {
+			fmt.Fprintf(out, "  Schema filter: %s\n", convertSchemaFilter)
+		}
+		printDryRunPlan(out, db)
+		fmt.Fprintf(os.Stderr, "=== Dry run complete: no output written ===\n\n")
+		return nil
+	}
+
 	// Write to target format
 	fmt.Fprintf(os.Stderr, "[2/2] Writing to target format...\n")
 	fmt.Fprintf(os.Stderr, "      Format: %s\n", convertTargetType)
@@ -261,6 +300,18 @@ func runConvert(cmd *cobra.Command, args []string) error {
 	return nil
 }
 
+// finalizeCommentedRefs resolves DBML `// Ref:` comments against the fully
+// loaded model and warns about refs whose target is not loaded.
+func finalizeCommentedRefs(db *models.Database, warn func(string)) {
+	for _, w := range dbml.ResolveCommentedRefs(db, true) {
+		warn(w)
+	}
+}
+
+func stderrWarn(msg string) {
+	fmt.Fprintf(os.Stderr, "      ⚠ %s\n", msg)
+}
+
 func readDatabaseListForConvert(dbType string, files []string) (*models.Database, error) {
 	if len(files) == 0 {
 		return nil, fmt.Errorf("file list is empty")
@@ -367,6 +418,12 @@ func readDatabaseForConvert(dbType, filePath, connString string) (*models.Databa
 		}
 		reader = mssql.NewReader(newReaderOptions("", connString))
 
+	case "mysql", "mariadb":
+		if connString == "" {
+			return nil, fmt.Errorf("connection string is required for MySQL format")
+		}
+		reader = mysql.NewReader(newReaderOptions("", connString))
+
 	case "sqlite", "sqlite3":
 		// SQLite can use either file path or connection string
 		dbPath := filePath
@@ -395,23 +452,12 @@ func writeDatabase(db *models.Database, dbType, outputPath, packageName, schemaF
 
 	writerOpts := newWriterOptions(outputPath, packageName, flattenSchema, nullableTypes, nullableArrays, continueOnError)
 	if extraFields != "" {
-		if !strings.EqualFold(dbType, "bun") {
-			return fmt.Errorf("--extra-fields is only supported for Bun output")
-		}
-		extraFieldsJSON, err := os.ReadFile(extraFields)
+		extraFieldsJSON, err := loadExtraFields(dbType, extraFields)
 		if err != nil {
-			return fmt.Errorf("failed to read --extra-fields file %q: %w", extraFields, err)
-		}
-
-		var parsed []wbun.ExtraFieldConfig
-		if err := stdjson.Unmarshal(extraFieldsJSON, &parsed); err != nil {
-			return fmt.Errorf("invalid --extra-fields JSON in %q: %w", extraFields, err)
-		}
-		if len(parsed) == 0 {
-			return fmt.Errorf("--extra-fields must contain at least one field")
+			return err
 		}
 		writerOpts.Metadata = map[string]interface{}{
-			"extra_fields": string(extraFieldsJSON),
+			"extra_fields": extraFieldsJSON,
 		}
 	}
 
@@ -452,6 +498,9 @@ func writeDatabase(db *models.Database, dbType, outputPath, packageName, schemaF
 	case "mssql", "sqlserver", "mssql2016", "mssql2017", "mssql2019", "mssql2022":
 		writer = wmssql.NewWriter(writerOpts)
 
+	case "mysql", "mariadb":
+		writer = wmysql.NewWriter(writerOpts)
+
 	case "sqlite", "sqlite3":
 		writer = wsqlite.NewWriter(writerOpts)
 
@@ -516,6 +565,95 @@ func writeDatabase(db *models.Database, dbType, outputPath, packageName, schemaF
 	return nil
 }
 
+// loadExtraFields reads and validates the --extra-fields JSON file, returning
+// its raw content.
+func loadExtraFields(dbType, path string) (string, error) {
+	if !strings.EqualFold(dbType, "bun") {
+		return "", fmt.Errorf("--extra-fields is only supported for Bun output")
+	}
+	extraFieldsJSON, err := os.ReadFile(path)
+	if err != nil {
+		return "", fmt.Errorf("failed to read --extra-fields file %q: %w", path, err)
+	}
+
+	var parsed []wbun.ExtraFieldConfig
+	if err := stdjson.Unmarshal(extraFieldsJSON, &parsed); err != nil {
+		return "", fmt.Errorf("invalid --extra-fields JSON in %q: %w", path, err)
+	}
+	if len(parsed) == 0 {
+		return "", fmt.Errorf("--extra-fields must contain at least one field")
+	}
+	return string(extraFieldsJSON), nil
+}
+
+// validateWriteTarget performs the checks writeDatabase makes before writing,
+// without constructing a writer or touching the output path. Used by --dry-run.
+func validateWriteTarget(db *models.Database, dbType, packageName, schemaFilter, extraFields string) error {
+	if extraFields != "" {
+		if _, err := loadExtraFields(dbType, extraFields); err != nil {
+			return err
+		}
+	}
+
+	switch strings.ToLower(dbType) {
+	case "dbml", "dctx", "drawdb", "json", "yaml", "yml", "drizzle",
+		"pgsql", "postgres", "postgresql", "sql",
+		"mssql", "sqlserver", "mssql2016", "mssql2017", "mssql2019", "mssql2022",
+		"sqlite", "sqlite3", "prisma", "typeorm", "graphql", "gql":
+	case "gorm", "bun":
+		if packageName == "" {
+			return fmt.Errorf("package name is required for %s format (use --package flag)", dbType)
+		}
+	default:
+		return fmt.Errorf("unsupported target format: %s", dbType)
+	}
+
+	if schemaFilter != "" {
+		for _, schema := range db.Schemas {
+			if schema.Name == schemaFilter {
+				return nil
+			}
+		}
+		return fmt.Errorf("schema '%s' not found in database. Available schemas: %v",
+			schemaFilter, getSchemaNames(db))
+	}
+
+	if strings.EqualFold(dbType, "dctx") {
+		if len(db.Schemas) == 0 {
+			return fmt.Errorf("no schemas found in database")
+		}
+		if len(db.Schemas) > 1 {
+			return fmt.Errorf("multiple schemas found, please specify which schema to export using --schema flag. Available schemas: %v",
+				getSchemaNames(db))
+		}
+	}
+	return nil
+}
+
+// outWriter returns the command's stdout, falling back to os.Stdout when the
+// command is nil (tests call the run functions directly).
+func outWriter(cmd *cobra.Command) io.Writer {
+	if cmd == nil {
+		return os.Stdout
+	}
+	return cmd.OutOrStdout()
+}
+
+// printDryRunPlan prints the schemas and tables that would be written.
+func printDryRunPlan(out io.Writer, db *models.Database) {
+	for _, schema := range db.Schemas {
+		names := make([]string, 0, len(schema.Tables))
+		for _, t := range schema.Tables {
+			names = append(names, t.Name)
+		}
+		fmt.Fprintf(out, "    schema %q: %d table(s)", schema.Name, len(schema.Tables))
+		if len(names) > 0 {
+			fmt.Fprintf(out, " [%s]", strings.Join(names, ", "))
+		}
+		fmt.Fprintln(out)
+	}
+}
+
 // getSchemaNames returns a slice of schema names from a database
 func getSchemaNames(db *models.Database) []string {
 	names := make([]string, len(db.Schemas))

cmd/relspec/convert_helpers_test.go

@@ -0,0 +1,312 @@
+package main
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+)
+
+const fixturesDir = "../../tests/assets"
+
+// readableFormats maps each file-based reader format to an existing fixture.
+var readableFormats = []struct {
+	format string
+	path   string
+}{
+	{"dbml", "dbml/simple.dbml"},
+	{"json", "json/database.json"},
+	{"yaml", "yaml/database.yaml"},
+	{"yml", "yaml/database.yaml"},
+	{"drawdb", "drawdb/simple.json"},
+	{"dctx", "dctx/p1.dctx"},
+	{"graphql", "graphql/simple.graphql"},
+	{"gql", "graphql/simple.graphql"},
+	{"prisma", "prisma/example.prisma"},
+	{"typeorm", "typeorm/example.ts"},
+	{"drizzle", "drizzle/schema.ts"},
+	{"gorm", "gorm/simple.go"},
+	{"bun", "bun/simple.go"},
+}
+
+func TestReadDatabaseForConvert_FileFormats(t *testing.T) {
+	for _, tt := range readableFormats {
+		t.Run(tt.format, func(t *testing.T) {
+			db, err := readDatabaseForConvert(tt.format, filepath.Join(fixturesDir, tt.path), "")
+			if err != nil {
+				t.Fatalf("read: %v", err)
+			}
+			if db == nil || len(db.Schemas) == 0 {
+				t.Fatalf("no schemas read: %+v", db)
+			}
+			// Uppercase format names are accepted.
+			if _, err := readDatabaseForConvert(strings.ToUpper(tt.format), filepath.Join(fixturesDir, tt.path), ""); err != nil {
+				t.Errorf("uppercase format: %v", err)
+			}
+		})
+	}
+}
+
+func TestReadDatabaseForConvert_Errors(t *testing.T) {
+	filePathFormats := []string{"dbml", "dctx", "drawdb", "json", "yaml", "gorm", "bun", "drizzle", "prisma", "typeorm", "graphql"}
+	for _, f := range filePathFormats {
+		t.Run("missing path "+f, func(t *testing.T) {
+			_, err := readDatabaseForConvert(f, "", "")
+			if err == nil || !strings.Contains(err.Error(), "file path is required") {
+				t.Errorf("got %v", err)
+			}
+		})
+	}
+	connFormats := []string{"pgsql", "postgres", "postgresql", "mssql", "sqlserver", "mysql", "mariadb"}
+	for _, f := range connFormats {
+		t.Run("missing conn "+f, func(t *testing.T) {
+			_, err := readDatabaseForConvert(f, "", "")
+			if err == nil || !strings.Contains(err.Error(), "connection string is required") {
+				t.Errorf("got %v", err)
+			}
+		})
+	}
+	if _, err := readDatabaseForConvert("sqlite", "", ""); err == nil || !strings.Contains(err.Error(), "required for SQLite") {
+		t.Errorf("sqlite: %v", err)
+	}
+	if _, err := readDatabaseForConvert("nope", "x", ""); err == nil || !strings.Contains(err.Error(), "unsupported source format") {
+		t.Errorf("unsupported: %v", err)
+	}
+	if _, err := readDatabaseForConvert("dbml", filepath.Join(t.TempDir(), "missing.dbml"), ""); err == nil || !strings.Contains(err.Error(), "failed to read database") {
+		t.Errorf("missing file: %v", err)
+	}
+}
+
+func TestReadDatabase_DiffReader(t *testing.T) {
+	for _, f := range []string{"dbml", "json", "yaml", "drawdb", "dctx"} {
+		for _, tt := range readableFormats {
+			if tt.format != f {
+				continue
+			}
+			t.Run(f, func(t *testing.T) {
+				db, err := readDatabase(f, filepath.Join(fixturesDir, tt.path), "", "source")
+				if err != nil || db == nil || len(db.Schemas) == 0 {
+					t.Fatalf("read: %v %+v", err, db)
+				}
+			})
+		}
+	}
+	for _, f := range []string{"dbml", "dctx", "drawdb", "json", "yaml", "sqldir"} {
+		if _, err := readDatabase(f, "", "", "src"); err == nil || !strings.Contains(err.Error(), "src: file path is required") {
+			t.Errorf("%s missing path: %v", f, err)
+		}
+	}
+	if _, err := readDatabase("pgsql", "", "", "src"); err == nil || !strings.Contains(err.Error(), "connection string is required") {
+		t.Errorf("pgsql: %v", err)
+	}
+	if _, err := readDatabase("sqlite", "", "", "src"); err == nil {
+		t.Error("sqlite without path must fail")
+	}
+	if _, err := readDatabase("nope", "x", "", "src"); err == nil || !strings.Contains(err.Error(), "unsupported database format") {
+		t.Errorf("unsupported: %v", err)
+	}
+	if _, err := readDatabase("json", filepath.Join(t.TempDir(), "missing.json"), "", "src"); err == nil || !strings.Contains(err.Error(), "src: failed to read database") {
+		t.Errorf("missing file: %v", err)
+	}
+}
+
+func TestMaskPassword(t *testing.T) {
+	tests := []struct {
+		in, want string
+	}{
+		{"", ""},
+		{"postgres://user:secret@host:5432/db", "postgres://user:***@host:5432/db"},
+		{"postgres://user@host:5432/db", "postgres://user@host:5432/db"},
+		{"host=h user=u password=secret dbname=d", "host=h user=u password=*** dbname=d"},
+		{"host=h user=u", "host=h user=u"},
+		{"/tmp/file.db", "/tmp/file.db"},
+	}
+	for _, tt := range tests {
+		if got := maskPassword(tt.in); got != tt.want {
+			t.Errorf("maskPassword(%q) = %q, want %q", tt.in, got, tt.want)
+		}
+		if got := maskPasswordInDiff(tt.in); got != tt.want {
+			t.Errorf("maskPasswordInDiff(%q) = %q, want %q", tt.in, got, tt.want)
+		}
+	}
+}
+
+func TestGetSchemaNames(t *testing.T) {
+	db := models.InitDatabase("d")
+	if got := getSchemaNames(db); len(got) != 0 {
+		t.Errorf("empty: %v", got)
+	}
+	db.Schemas = []*models.Schema{{Name: "a"}, {Name: "b"}}
+	if got := strings.Join(getSchemaNames(db), ","); got != "a,b" {
+		t.Errorf("got %s", got)
+	}
+}
+
+func TestLoadExtraFields(t *testing.T) {
+	dir := t.TempDir()
+	write := func(name, body string) string {
+		p := filepath.Join(dir, name)
+		if err := os.WriteFile(p, []byte(body), 0o644); err != nil {
+			t.Fatal(err)
+		}
+		return p
+	}
+	valid := write("valid.json", `[{"name":"extra"}]`)
+
+	if got, err := loadExtraFields("bun", valid); err != nil || !strings.Contains(got, "extra") {
+		t.Errorf("valid: %q %v", got, err)
+	}
+	if _, err := loadExtraFields("BUN", valid); err != nil {
+		t.Errorf("case-insensitive format: %v", err)
+	}
+	if _, err := loadExtraFields("gorm", valid); err == nil || !strings.Contains(err.Error(), "only supported for Bun") {
+		t.Errorf("non-bun: %v", err)
+	}
+	if _, err := loadExtraFields("bun", filepath.Join(dir, "missing.json")); err == nil || !strings.Contains(err.Error(), "failed to read") {
+		t.Errorf("missing: %v", err)
+	}
+	if _, err := loadExtraFields("bun", write("bad.json", `{not json`)); err == nil || !strings.Contains(err.Error(), "invalid --extra-fields JSON") {
+		t.Errorf("bad json: %v", err)
+	}
+	if _, err := loadExtraFields("bun", write("empty.json", `[]`)); err == nil || !strings.Contains(err.Error(), "at least one field") {
+		t.Errorf("empty: %v", err)
+	}
+}
+
+func multiSchemaDB() *models.Database {
+	db := models.InitDatabase("multi")
+	for _, n := range []string{"a", "b"} {
+		s := models.InitSchema(n)
+		tbl := models.InitTable("t_"+n, n)
+		c := models.InitColumn("id", tbl.Name, n)
+		c.Type = "integer"
+		c.IsPrimaryKey = true
+		tbl.Columns["id"] = c
+		s.Tables = append(s.Tables, tbl)
+		db.Schemas = append(db.Schemas, s)
+	}
+	return db
+}
+
+func TestValidateWriteTarget(t *testing.T) {
+	db := multiSchemaDB()
+	single := models.InitDatabase("single")
+	single.Schemas = []*models.Schema{models.InitSchema("only")}
+
+	tests := []struct {
+		name                                                  string
+		db                                                    *models.Database
+		dbType, pkg, schemaFilter, extraFields, wantErrSubstr string
+	}{
+		{"json ok", db, "json", "", "", "", ""},
+		{"pgsql alias ok", db, "sql", "", "", "", ""},
+		{"gorm needs package", db, "gorm", "", "", "", "package name is required"},
+		{"bun needs package", db, "bun", "", "", "", "package name is required"},
+		{"gorm with package", db, "gorm", "models", "", "", ""},
+		{"unsupported", db, "nope", "", "", "", "unsupported target format"},
+		{"schema filter found", db, "json", "", "a", "", ""},
+		{"schema filter missing", db, "json", "", "zzz", "", "not found in database"},
+		{"dctx multi schema", db, "dctx", "", "", "", "multiple schemas found"},
+		{"dctx multi schema with filter", db, "dctx", "", "a", "", ""},
+		{"dctx single schema", single, "dctx", "", "", "", ""},
+		{"dctx no schemas", models.InitDatabase("e"), "dctx", "", "", "", "no schemas found"},
+		{"extra fields non-bun", db, "json", "", "", "x.json", "only supported for Bun"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			err := validateWriteTarget(tt.db, tt.dbType, tt.pkg, tt.schemaFilter, tt.extraFields)
+			if tt.wantErrSubstr == "" {
+				if err != nil {
+					t.Errorf("unexpected error: %v", err)
+				}
+				return
+			}
+			if err == nil || !strings.Contains(err.Error(), tt.wantErrSubstr) {
+				t.Errorf("got %v, want substring %q", err, tt.wantErrSubstr)
+			}
+		})
+	}
+}
+
+func TestWriteDatabase_Formats(t *testing.T) {
+	db := multiSchemaDB()
+	formats := []struct{ format, file string }{
+		{"json", "out.json"},
+		{"yaml", "out.yaml"},
+		{"yml", "out.yml"},
+		{"dbml", "out.dbml"},
+		{"drawdb", "out.drawdb.json"},
+		{"pgsql", "out.sql"},
+		{"postgres", "out2.sql"},
+		{"sql", "out3.sql"},
+		{"mssql", "out_ms.sql"},
+		{"mysql", "out_my.sql"},
+		{"sqlite", "out_lite.sql"},
+		{"graphql", "out.graphql"},
+		{"gql", "out2.graphql"},
+		{"prisma", "out.prisma"},
+		{"typeorm", "out.ts"},
+		{"drizzle", "out_drizzle.ts"},
+	}
+	for _, tt := range formats {
+		t.Run(tt.format, func(t *testing.T) {
+			out := filepath.Join(t.TempDir(), tt.file)
+			if err := writeDatabase(db, tt.format, out, "", "", false, "", "", false, ""); err != nil {
+				t.Fatalf("write: %v", err)
+			}
+			info, err := os.Stat(out)
+			if err != nil || info.Size() == 0 {
+				t.Errorf("output missing or empty: %v", err)
+			}
+		})
+	}
+}
+
+func TestWriteDatabase_GoFormatsWriteIntoDir(t *testing.T) {
+	db := multiSchemaDB()
+	for _, f := range []string{"gorm", "bun"} {
+		t.Run(f, func(t *testing.T) {
+			out := filepath.Join(t.TempDir(), "models.go")
+			if err := writeDatabase(db, f, out, "models", "", false, "", "", false, ""); err != nil {
+				t.Fatal(err)
+			}
+			if _, err := os.Stat(out); err != nil {
+				t.Errorf("no output: %v", err)
+			}
+			if err := writeDatabase(db, f, out, "", "", false, "", "", false, ""); err == nil || !strings.Contains(err.Error(), "package name is required") {
+				t.Errorf("missing package: %v", err)
+			}
+		})
+	}
+}
+
+func TestWriteDatabase_SchemaFilterAndDCTX(t *testing.T) {
+	db := multiSchemaDB()
+	out := filepath.Join(t.TempDir(), "o.json")
+
+	if err := writeDatabase(db, "json", out, "", "a", false, "", "", false, ""); err != nil {
+		t.Errorf("schema filter: %v", err)
+	}
+	if err := writeDatabase(db, "json", out, "", "zzz", false, "", "", false, ""); err == nil || !strings.Contains(err.Error(), "not found in database") {
+		t.Errorf("missing schema: %v", err)
+	}
+	if err := writeDatabase(db, "dctx", out, "", "", false, "", "", false, ""); err == nil || !strings.Contains(err.Error(), "multiple schemas found") {
+		t.Errorf("dctx multi: %v", err)
+	}
+	single := models.InitDatabase("s")
+	single.Schemas = []*models.Schema{db.Schemas[0]}
+	if err := writeDatabase(single, "dctx", filepath.Join(t.TempDir(), "o.dctx"), "", "", false, "", "", false, ""); err != nil {
+		t.Errorf("dctx single: %v", err)
+	}
+	if err := writeDatabase(models.InitDatabase("e"), "dctx", out, "", "", false, "", "", false, ""); err == nil || !strings.Contains(err.Error(), "no schemas found") {
+		t.Errorf("dctx empty: %v", err)
+	}
+	if err := writeDatabase(db, "nope", out, "", "", false, "", "", false, ""); err == nil || !strings.Contains(err.Error(), "unsupported target format") {
+		t.Errorf("unsupported: %v", err)
+	}
+	if err := writeDatabase(db, "json", out, "", "", false, "", "", false, filepath.Join(t.TempDir(), "x.json")); err == nil || !strings.Contains(err.Error(), "only supported for Bun") {
+		t.Errorf("extra fields with json: %v", err)
+	}
+}

cmd/relspec/dry_run_test.go

@@ -0,0 +1,158 @@
+package main
+
+import (
+	"bytes"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"github.com/spf13/cobra"
+)
+
+func newDryRunCmd() (*cobra.Command, *bytes.Buffer) {
+	var buf bytes.Buffer
+	cmd := &cobra.Command{}
+	cmd.SetOut(&buf)
+	return cmd, &buf
+}
+
+func TestRunConvert_DryRunWritesNothing(t *testing.T) {
+	defer func(a, b, c, d string, e bool) {
+		convertSourceType, convertSourcePath, convertTargetType, convertTargetPath, convertDryRun = a, b, c, d, e
+	}(convertSourceType, convertSourcePath, convertTargetType, convertTargetPath, convertDryRun)
+
+	dir := t.TempDir()
+	in := filepath.Join(dir, "in.json")
+	out := filepath.Join(dir, "out.json")
+	writeTestJSON(t, in, []string{"users", "posts"})
+
+	convertSourceType, convertSourcePath = "json", in
+	convertTargetType, convertTargetPath = "json", out
+	convertDryRun = true
+
+	cmd, buf := newDryRunCmd()
+	if err := runConvert(cmd, nil); err != nil {
+		t.Fatalf("dry run: %v", err)
+	}
+	if _, err := os.Stat(out); !os.IsNotExist(err) {
+		t.Fatal("dry run must not create the output file")
+	}
+	for _, want := range []string{"dry run", "users", "posts", out} {
+		if !strings.Contains(buf.String(), want) {
+			t.Errorf("plan missing %q:\n%s", want, buf.String())
+		}
+	}
+
+	// Normal behavior is unchanged.
+	convertDryRun = false
+	if err := runConvert(cmd, nil); err != nil {
+		t.Fatalf("real run: %v", err)
+	}
+	if _, err := os.Stat(out); err != nil {
+		t.Fatalf("real run should write output: %v", err)
+	}
+}
+
+func TestRunConvert_DryRunValidatesTarget(t *testing.T) {
+	defer func(a, b, c, d string, e bool) {
+		convertSourceType, convertSourcePath, convertTargetType, convertTargetPath, convertDryRun = a, b, c, d, e
+	}(convertSourceType, convertSourcePath, convertTargetType, convertTargetPath, convertDryRun)
+
+	dir := t.TempDir()
+	in := filepath.Join(dir, "in.json")
+	out := filepath.Join(dir, "models")
+	writeTestJSON(t, in, []string{"users"})
+
+	convertSourceType, convertSourcePath = "json", in
+	convertDryRun = true
+
+	// gorm without --package must fail validation, as a real run would.
+	convertTargetType, convertTargetPath = "gorm", out
+	cmd, _ := newDryRunCmd()
+	err := runConvert(cmd, nil)
+	if err == nil || !strings.Contains(err.Error(), "package name is required") {
+		t.Fatalf("expected package validation error, got %v", err)
+	}
+
+	convertTargetType = "nope"
+	err = runConvert(cmd, nil)
+	if err == nil || !strings.Contains(err.Error(), "unsupported target format") {
+		t.Fatalf("expected unsupported format error, got %v", err)
+	}
+	if _, statErr := os.Stat(out); !os.IsNotExist(statErr) {
+		t.Fatal("dry run must not create the output path")
+	}
+}
+
+func TestRunSplit_DryRunWritesNothing(t *testing.T) {
+	defer func(a, b, c, d, e string, f bool) {
+		splitSourceType, splitSourcePath, splitTargetType, splitTargetPath, splitTables, splitDryRun = a, b, c, d, e, f
+	}(splitSourceType, splitSourcePath, splitTargetType, splitTargetPath, splitTables, splitDryRun)
+
+	dir := t.TempDir()
+	in := filepath.Join(dir, "in.json")
+	out := filepath.Join(dir, "subset.json")
+	writeTestJSON(t, in, []string{"users", "posts", "comments"})
+
+	splitSourceType, splitSourcePath = "json", in
+	splitTargetType, splitTargetPath = "json", out
+	splitTables = "users,posts"
+	splitDryRun = true
+
+	cmd, buf := newDryRunCmd()
+	if err := runSplit(cmd, nil); err != nil {
+		t.Fatalf("dry run: %v", err)
+	}
+	if _, err := os.Stat(out); !os.IsNotExist(err) {
+		t.Fatal("dry run must not create the output file")
+	}
+	got := buf.String()
+	if !strings.Contains(got, "2 table(s)") || strings.Contains(got, "comments") {
+		t.Errorf("plan should show only the 2 selected tables:\n%s", got)
+	}
+
+	// A selection that matches nothing fails validation in dry-run too.
+	splitTables = "does_not_exist"
+	if err := runSplit(cmd, nil); err == nil {
+		t.Fatal("expected error for empty selection")
+	}
+}
+
+func TestRunMerge_DryRunWritesNothing(t *testing.T) {
+	saved := saveMergeState()
+	defer restoreMergeState(saved)
+	defer func(v bool) { mergeDryRun = v }(mergeDryRun)
+
+	dir := t.TempDir()
+	target := filepath.Join(dir, "target.json")
+	source := filepath.Join(dir, "source.json")
+	out := filepath.Join(dir, "merged.json")
+	writeTestJSON(t, target, []string{"users"})
+	writeTestJSON(t, source, []string{"posts"})
+
+	mergeTargetType, mergeTargetPath, mergeTargetConn = "json", target, ""
+	mergeSourceType, mergeSourcePath, mergeSourceConn = "json", source, ""
+	mergeFromList = nil
+	mergeOutputType, mergeOutputPath, mergeOutputConn = "json", out, ""
+	mergeSkipTables, mergeReportPath = "", ""
+	mergeDryRun = true
+
+	cmd, buf := newDryRunCmd()
+	if err := runMerge(cmd, nil); err != nil {
+		t.Fatalf("dry run: %v", err)
+	}
+	if _, err := os.Stat(out); !os.IsNotExist(err) {
+		t.Fatal("dry run must not create the output file")
+	}
+	for _, want := range []string{"dry run", "users", "posts", out} {
+		if !strings.Contains(buf.String(), want) {
+			t.Errorf("plan missing %q:\n%s", want, buf.String())
+		}
+	}
+
+	mergeOutputType = "nope"
+	if err := runMerge(cmd, nil); err == nil || !strings.Contains(err.Error(), "unsupported format") {
+		t.Fatalf("expected unsupported output format error, got %v", err)
+	}
+}

cmd/relspec/job.go

@@ -60,8 +60,9 @@ inputs, output and options:
       logfile: .relspec/log/build-schema.log
 
 Rules and guarantees:
-  - command is a closed allow-list (convert, merge, scripts-list, templ). Arbitrary
-    shell strings are never executed.
+  - command is a closed allow-list (convert, merge, split, scripts-list,
+    scripts-exec, templ, inspect, diff). Arbitrary shell strings are never
+    executed.
   - Every path is relative to the directory holding the job file and may not
     escape it. Absolute and home-relative paths are rejected.
   - Remote database credentials are referenced by environment-variable name
@@ -246,22 +247,37 @@ type resolvedInput struct {
 func preflightJob(j *jobs.Job, resolvedByName map[string]*resolvedJob) (*resolvedJob, error) {
 	root := j.Dir()
 	rj := &resolvedJob{job: j, root: root, logPolicy: j.ResolvedLogPolicy()}
+	expandPath := func(label, value string) (string, error) {
+		expanded, err := jobs.ExpandEnv(value)
+		if err != nil {
+			return "", fmt.Errorf("%s: %w", label, err)
+		}
+		return expanded, nil
+	}
 
 	if j.Logfile != "" {
-		p, err := jobs.SafeJoin(root, j.Logfile)
+		logfile, err := expandPath("logfile", j.Logfile)
+		if err != nil {
+			return nil, err
+		}
+		p, err := jobs.SafeJoin(root, logfile)
 		if err != nil {
 			return nil, fmt.Errorf("logfile: %w", err)
 		}
 		rj.logPath = p
 	}
 	if j.Template != "" {
-		p, err := jobs.SafeJoin(root, j.Template)
+		templatePath, err := expandPath("template", j.Template)
+		if err != nil {
+			return nil, err
+		}
+		p, err := jobs.SafeJoin(root, templatePath)
 		if err != nil {
 			return nil, fmt.Errorf("template: %w", err)
 		}
 		info, err := os.Stat(p)
 		if err != nil || info.IsDir() {
-			return nil, fmt.Errorf("template %q: not found or is a directory", j.Template)
+			return nil, fmt.Errorf("template %q: not found or is a directory", templatePath)
 		}
 		rj.templatePath = p
 	}
@@ -291,16 +307,20 @@ func preflightJob(j *jobs.Job, resolvedByName map[string]*resolvedJob) (*resolve
 			ri.connEnv = in.ConnEnv
 			rj.secrets = append(rj.secrets, v)
 		} else {
-			p, err := jobs.SafeJoin(root, in.Path)
+			inputPath, err := expandPath(fmt.Sprintf("input[%d]", i), in.Path)
+			if err != nil {
+				return nil, err
+			}
+			p, err := jobs.SafeJoin(root, inputPath)
 			if err != nil {
 				return nil, fmt.Errorf("input[%d]: %w", i, err)
 			}
 			info, err := os.Stat(p)
 			if err != nil {
-				return nil, fmt.Errorf("input[%d]: %s: file not found", i, in.Path)
+				return nil, fmt.Errorf("input[%d]: %s: file not found", i, inputPath)
 			}
 			if info.IsDir() {
-				return nil, fmt.Errorf("input[%d]: %s: is a directory, not a file", i, in.Path)
+				return nil, fmt.Errorf("input[%d]: %s: is a directory, not a file", i, inputPath)
 			}
 			ri.path = p
 		}
@@ -308,13 +328,17 @@ func preflightJob(j *jobs.Job, resolvedByName map[string]*resolvedJob) (*resolve
 	}
 
 	for _, d := range j.ScriptDirs {
-		p, err := jobs.SafeJoin(root, d)
+		scriptDir, err := expandPath("script_dir", d)
+		if err != nil {
+			return nil, err
+		}
+		p, err := jobs.SafeJoin(root, scriptDir)
 		if err != nil {
 			return nil, fmt.Errorf("script_dir %q: %w", d, err)
 		}
 		info, err := os.Stat(p)
 		if err != nil {
-			return nil, fmt.Errorf("script_dir %q: not found", d)
+			return nil, fmt.Errorf("script_dir %q: not found", scriptDir)
 		}
 		if !info.IsDir() {
 			return nil, fmt.Errorf("script_dir %q: not a directory", d)
@@ -332,25 +356,33 @@ func preflightJob(j *jobs.Job, resolvedByName map[string]*resolvedJob) (*resolve
 			rj.outputConnEnv = j.Output.ConnEnv
 			rj.secrets = append(rj.secrets, v)
 		} else {
-			p, err := jobs.SafeJoin(root, j.Output.Path)
+			outputPath, err := expandPath("output", j.Output.Path)
+			if err != nil {
+				return nil, err
+			}
+			p, err := jobs.SafeJoin(root, outputPath)
 			if err != nil {
 				return nil, fmt.Errorf("output: %w", err)
 			}
 			if _, err := os.Stat(p); err == nil && !j.Output.Overwrite {
-				return nil, fmt.Errorf("output %s already exists (set output.overwrite: true to replace it)", j.Output.Path)
+				return nil, fmt.Errorf("output %s already exists (set output.overwrite: true to replace it)", outputPath)
 			}
 			rj.outputPath = p
 		}
 	}
 
 	if j.Rules != "" {
-		p, err := jobs.SafeJoin(root, j.Rules)
+		rulesPath, err := expandPath("rules", j.Rules)
+		if err != nil {
+			return nil, err
+		}
+		p, err := jobs.SafeJoin(root, rulesPath)
 		if err != nil {
 			return nil, fmt.Errorf("rules: %w", err)
 		}
 		info, err := os.Stat(p)
 		if err != nil || info.IsDir() {
-			return nil, fmt.Errorf("rules %q: not found or is a directory", j.Rules)
+			return nil, fmt.Errorf("rules %q: not found or is a directory", rulesPath)
 		}
 		rj.rulesPath = p
 	}
@@ -358,12 +390,16 @@ func preflightJob(j *jobs.Job, resolvedByName map[string]*resolvedJob) (*resolve
 	if j.Report != nil {
 		rj.reportFormat = strings.ToLower(j.Report.Format)
 		if j.Report.Path != "" {
-			p, err := jobs.SafeJoin(root, j.Report.Path)
+			reportPath, err := expandPath("report", j.Report.Path)
+			if err != nil {
+				return nil, err
+			}
+			p, err := jobs.SafeJoin(root, reportPath)
 			if err != nil {
 				return nil, fmt.Errorf("report: %w", err)
 			}
 			if _, err := os.Stat(p); err == nil && !j.Report.Overwrite {
-				return nil, fmt.Errorf("report %s already exists (set report.overwrite: true to replace it)", j.Report.Path)
+				return nil, fmt.Errorf("report %s already exists (set report.overwrite: true to replace it)", reportPath)
 			}
 			rj.reportPath = p
 		}
@@ -542,6 +578,7 @@ func runMergeJob(rj *resolvedJob, lg *jobLogger) error {
 		lg.logf("merging: %s", inputLabel(ri))
 		merge.MergeDatabases(base, db, opts)
 	}
+	finalizeCommentedRefs(base, func(w string) { lg.logf("warning: %s", w) })
 	base.UpdateDate()
 	return writeJobOutput(rj, base, lg)
 }
@@ -778,6 +815,7 @@ func readJobInputs(rj *resolvedJob, lg *jobLogger) (*models.Database, error) {
 	if base == nil {
 		return nil, fmt.Errorf("no inputs produced a database")
 	}
+	finalizeCommentedRefs(base, func(w string) { lg.logf("warning: %s", w) })
 	return base, nil
 }
 
@@ -805,8 +843,8 @@ func writeJobOutput(rj *resolvedJob, db *models.Database, lg *jobLogger) error {
 			return fmt.Errorf("database output is only supported for pgsql (got %q)", rj.job.Output.Format)
 		}
 		lg.logf("writing output to database env:%s", rj.outputConnEnv)
-		writerOpts := newWriterOptions("", o.Package, o.FlattenSchema, "", "", o.ContinueOnError)
-		writerOpts.Metadata = map[string]interface{}{"connection_string": rj.outputConn}
+		writerOpts := newWriterOptions("", o.Package, o.FlattenSchema, o.Types, o.ArrayNullable, o.ContinueOnError)
+		writerOpts.Metadata = map[string]interface{}{"connection_string": rj.outputConn, "full_ddl": o.FullDDL}
 		return wpgsql.NewWriter(writerOpts).WriteDatabase(db)
 	}
 
@@ -816,7 +854,7 @@ func writeJobOutput(rj *resolvedJob, db *models.Database, lg *jobLogger) error {
 	lg.logf("writing output: %s (%s)", rj.outputPath, format)
 
 	write := func(target string) error {
-		return writeDatabase(db, format, target, o.Package, o.Schema, o.FlattenSchema, "", "", o.ContinueOnError, "")
+		return writeDatabase(db, format, target, o.Package, o.Schema, o.FlattenSchema, o.Types, o.ArrayNullable, o.ContinueOnError, "")
 	}
 	// Single-file formats are written to a temp file and renamed into place so
 	// a failure never leaves a partial or truncated output. Directory-emitting

cmd/relspec/job_test.go

@@ -666,6 +666,7 @@ jobs:
         format: dbml
     template: templates/schema.tmpl
     output:
+      format: text
       path: build/schema.txt
       overwrite: true
 `)
@@ -682,3 +683,61 @@ jobs:
 		t.Fatalf("templ output missing users table: %s", out)
 	}
 }
+
+func TestJobRun_BunSqlTypesOption(t *testing.T) {
+	dir := jobFixture(t, `version: 1
+jobs:
+  models:
+    command: convert
+    inputs:
+      - path: schema/core.dbml
+        format: dbml
+    output:
+      format: bun
+      path: build/models
+    options:
+      package: models
+      types: sqltypes
+`)
+	set := mustLoadSet(t, filepath.Join(dir, "relspec.yml"))
+	if err := executeJobPlan(set, "models", false, false, &bytes.Buffer{}); err != nil {
+		t.Fatalf("execute Bun job: %v", err)
+	}
+	generated, err := os.ReadFile(filepath.Join(dir, "build", "models"))
+	if err != nil {
+		t.Fatalf("read Bun output: %v", err)
+	}
+	if !strings.Contains(string(generated), "git.warky.dev/wdevs/relspecgo/pkg/sqltypes") {
+		t.Fatalf("Bun output did not preserve options.types: %s", generated)
+	}
+}
+
+func TestJobValidation_TemplTextFormat(t *testing.T) {
+	validDir := jobFixture(t, `version: 1
+jobs:
+  docs:
+    command: templ
+    inputs: [{path: schema/core.dbml, format: dbml}]
+    template: schema.tmpl
+    output: {format: text, path: build/docs.txt}
+`)
+	valid := mustLoadSet(t, filepath.Join(validDir, "relspec.yml"))
+	if err := valid.Validate(); err != nil {
+		t.Fatalf("text output should be valid for templ: %v", err)
+	}
+	invalidDir := jobFixture(t, `version: 1
+jobs:
+  docs:
+    command: templ
+    inputs: [{path: schema/core.dbml, format: dbml}]
+    template: schema.tmpl
+    output: {format: json, path: build/docs.txt}
+`)
+	invalid, err := jobs.Load([]string{filepath.Join(invalidDir, "relspec.yml")})
+	if err != nil {
+		t.Fatalf("load invalid manifest: %v", err)
+	}
+	if err := invalid.Validate(); err == nil || !strings.Contains(err.Error(), "output.format: text") {
+		t.Fatalf("expected templ format rejection, got %v", err)
+	}
+}

cmd/relspec/main.go

@@ -5,6 +5,8 @@ import (
 	"os"
 )
 
+// asciiLogo (see version.go) is printed by the `version` command.
+
 func main() {
 	args := os.Args[1:]
 	isSilent := hasSilentFlag(args)

cmd/relspec/merge.go

@@ -59,7 +59,9 @@ var (
 	mergeSkipTables    string // Comma-separated table names to skip
 	mergeVerbose       bool
 	mergeReportPath    string // Path to write merge report
+	mergeFullDDL       bool   // Execute full DDL instead of diffing the live pgsql database
 	mergeFlattenSchema bool
+	mergeDryRun        bool
 )
 
 var mergeCmd = &cobra.Command{
@@ -127,7 +129,9 @@ func init() {
 	mergeCmd.Flags().BoolVar(&mergeSkipSequences, "skip-sequences", false, "Skip sequences during merge")
 	mergeCmd.Flags().StringVar(&mergeSkipTables, "skip-tables", "", "Comma-separated list of table names to skip during merge")
 	mergeCmd.Flags().BoolVar(&mergeVerbose, "verbose", false, "Show verbose output")
+	mergeCmd.Flags().BoolVar(&mergeFullDDL, "full-ddl", false, "pgsql database output: execute the full DDL instead of only the differences from the live database")
 	mergeCmd.Flags().StringVar(&mergeReportPath, "merge-report", "", "Path to write merge report (JSON format)")
+	mergeCmd.Flags().BoolVar(&mergeDryRun, "dry-run", false, "Read and merge in memory, then print the merge plan without writing any output")
 	mergeCmd.Flags().BoolVar(&mergeFlattenSchema, "flatten-schema", false, "Flatten schema.table names to schema_table (useful for databases like SQLite that do not support schemas)")
 }
 
@@ -248,6 +252,7 @@ func runMerge(cmd *cobra.Command, args []string) error {
 	}
 
 	result := merge.MergeDatabases(targetDB, sourceDB, opts)
+	finalizeCommentedRefs(targetDB, stderrWarn)
 
 	// Update timestamp
 	targetDB.UpdateDate()
@@ -261,6 +266,29 @@ func runMerge(cmd *cobra.Command, args []string) error {
 			merge.GetColumnTypeConflictSummary(result, 10))
 	}
 
+	if mergeDryRun {
+		if !isMergeOutputFormat(mergeOutputType) {
+			return fmt.Errorf("dry run validation failed: Output: unsupported format '%s'", mergeOutputType)
+		}
+		out := outWriter(cmd)
+		fmt.Fprintf(out, "RelSpec merge plan (dry run - nothing written):\n")
+		fmt.Fprintf(out, "  Target: %s database '%s'\n", mergeTargetType, targetDB.Name)
+		fmt.Fprintf(out, "  Source: %s database '%s'\n", mergeSourceType, sourceDB.Name)
+		switch {
+		case mergeOutputPath != "":
+			fmt.Fprintf(out, "  Output: %s -> %s\n", mergeOutputType, mergeOutputPath)
+		case mergeOutputConn != "":
+			fmt.Fprintf(out, "  Output: %s -> %s\n", mergeOutputType, maskPassword(mergeOutputConn))
+		default:
+			fmt.Fprintf(out, "  Output: %s\n", mergeOutputType)
+		}
+		fmt.Fprintf(out, "  Result:\n")
+		printDryRunPlan(out, targetDB)
+		fmt.Fprintf(out, "\n%s\n", merge.GetMergeSummary(result))
+		fmt.Fprintf(os.Stderr, "\n=== Dry run complete: no output written ===\n")
+		return nil
+	}
+
 	// Step 4: Write output
 	fmt.Fprintf(os.Stderr, "\n[4/4] Writing output...\n")
 	fmt.Fprintf(os.Stderr, "      Format: %s\n", mergeOutputType)
@@ -282,6 +310,16 @@ func runMerge(cmd *cobra.Command, args []string) error {
 	return nil
 }
 
+// isMergeOutputFormat reports whether writeDatabaseForMerge supports dbType.
+func isMergeOutputFormat(dbType string) bool {
+	switch strings.ToLower(dbType) {
+	case "dbml", "dctx", "drawdb", "graphql", "json", "yaml", "gorm", "bun",
+		"drizzle", "prisma", "typeorm", "sqlite", "sqlite3", "pgsql":
+		return true
+	}
+	return false
+}
+
 func readDatabaseForMerge(dbType, filePath, connString, label string) (*models.Database, error) {
 	var reader readers.Reader
 
@@ -442,6 +480,7 @@ func writeDatabaseForMerge(dbType, filePath, connString string, db *models.Datab
 		if connString != "" {
 			writerOpts.Metadata = map[string]interface{}{
 				"connection_string": connString,
+				"full_ddl":          mergeFullDDL,
 			}
 			// Add report path if merge report is enabled
 			if mergeReportPath != "" {

cmd/relspec/merge_inspect_report_helpers_test.go

@@ -0,0 +1,287 @@
+package main
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+	"time"
+)
+
+func TestReadDatabaseForMerge(t *testing.T) {
+	for _, tt := range readableFormats {
+		t.Run(tt.format, func(t *testing.T) {
+			db, err := readDatabaseForMerge(tt.format, filepath.Join(fixturesDir, tt.path), "", "Target")
+			if err != nil {
+				t.Skipf("format %s not supported by merge reader: %v", tt.format, err)
+			}
+			if db == nil || len(db.Schemas) == 0 {
+				t.Errorf("no schemas: %+v", db)
+			}
+		})
+	}
+	for _, f := range []string{"dbml", "dctx", "drawdb", "graphql", "json", "yaml", "gorm", "bun", "drizzle", "prisma", "typeorm"} {
+		if _, err := readDatabaseForMerge(f, "", "", "Src"); err == nil || !strings.Contains(err.Error(), "Src: file path is required") {
+			t.Errorf("%s missing path: %v", f, err)
+		}
+	}
+	if _, err := readDatabaseForMerge("pgsql", "", "", "Src"); err == nil || !strings.Contains(err.Error(), "Src:") {
+		t.Errorf("pgsql: %v", err)
+	}
+	if _, err := readDatabaseForMerge("sqlite", "", "", "Src"); err == nil || !strings.Contains(err.Error(), "Src:") {
+		t.Errorf("sqlite: %v", err)
+	}
+	if _, err := readDatabaseForMerge("nope", "x", "", "Src"); err == nil || !strings.Contains(err.Error(), "unsupported format 'nope'") {
+		t.Errorf("unsupported: %v", err)
+	}
+}
+
+func TestWriteDatabaseForMerge(t *testing.T) {
+	db := multiSchemaDB()
+	single := multiSchemaDB()
+	single.Schemas = single.Schemas[:1]
+
+	files := map[string]string{
+		"dbml": "o.dbml", "dctx": "o.dctx", "drawdb": "o.drawdb.json", "graphql": "o.graphql",
+		"json": "o.json", "yaml": "o.yaml", "gorm": "gorm.go", "bun": "bun.go",
+		"drizzle": "o.ts", "prisma": "o.prisma", "typeorm": "te.ts",
+	}
+	for f, name := range files {
+		t.Run(f, func(t *testing.T) {
+			out := filepath.Join(t.TempDir(), name)
+			if f == "dctx" {
+				// DCTX cannot write a full database.
+				if err := writeDatabaseForMerge(f, out, "", single, "Output", false); err == nil || !strings.Contains(err.Error(), "not supported for DCTX") {
+					t.Errorf("dctx: %v", err)
+				}
+				if err := writeDatabaseForMerge(f, "", "", single, "Output", false); err == nil || !strings.Contains(err.Error(), "file path is required") {
+					t.Errorf("dctx missing path: %v", err)
+				}
+				return
+			}
+			src := db
+			if err := writeDatabaseForMerge(f, out, "", src, "Output", false); err != nil {
+				t.Fatalf("write: %v", err)
+			}
+			if _, err := os.Stat(out); err != nil {
+				t.Errorf("no output: %v", err)
+			}
+			if err := writeDatabaseForMerge(f, "", "", src, "Output", false); err == nil || !strings.Contains(err.Error(), "Output: file path is required") {
+				t.Errorf("missing path: %v", err)
+			}
+		})
+	}
+	for _, f := range []string{"pgsql", "sqlite"} {
+		out := filepath.Join(t.TempDir(), "o.sql")
+		if err := writeDatabaseForMerge(f, out, "", db, "Output", false); err != nil {
+			t.Errorf("%s script write: %v", f, err)
+		}
+	}
+	if err := writeDatabaseForMerge("pgsql", "", "postgres://u:p@127.0.0.1:1/none?connect_timeout=1", db, "Output", false); err == nil {
+		t.Error("pgsql with unreachable conn must fail")
+	}
+	if err := writeDatabaseForMerge("nope", "x", "", db, "Output", false); err == nil || !strings.Contains(err.Error(), "unsupported") {
+		t.Errorf("unsupported: %v", err)
+	}
+}
+
+func TestIsMergeOutputFormat(t *testing.T) {
+	for _, f := range []string{"dbml", "JSON", "pgsql", "sqlite3", "prisma"} {
+		if !isMergeOutputFormat(f) {
+			t.Errorf("%s should be supported", f)
+		}
+	}
+	for _, f := range []string{"", "nope", "mssql"} {
+		if isMergeOutputFormat(f) {
+			t.Errorf("%s should not be supported", f)
+		}
+	}
+}
+
+func TestExpandPath(t *testing.T) {
+	home, err := os.UserHomeDir()
+	if err != nil {
+		t.Skip("no home dir")
+	}
+	tests := []struct{ in, want string }{
+		{"", ""},
+		{"/abs/path", "/abs/path"},
+		{"rel/path", "rel/path"},
+		{"~/x/y", filepath.Join(home, "/x/y")},
+		{"~", home},
+	}
+	for _, tt := range tests {
+		if got := expandPath(tt.in); got != tt.want {
+			t.Errorf("expandPath(%q) = %q, want %q", tt.in, got, tt.want)
+		}
+	}
+}
+
+func TestParseSkipTables(t *testing.T) {
+	tests := []struct {
+		in   string
+		want []string
+	}{
+		{"", nil},
+		{" , ,", nil},
+		{"Users", []string{"users"}},
+		{" Users , ORDERS,,items ", []string{"users", "orders", "items"}},
+	}
+	for _, tt := range tests {
+		got := parseSkipTables(tt.in)
+		if len(got) != len(tt.want) {
+			t.Errorf("parseSkipTables(%q) = %v", tt.in, got)
+		}
+		for _, w := range tt.want {
+			if !got[w] {
+				t.Errorf("parseSkipTables(%q) missing %q", tt.in, w)
+			}
+		}
+	}
+}
+
+func TestReadDatabaseForInspect(t *testing.T) {
+	for _, tt := range readableFormats {
+		t.Run(tt.format, func(t *testing.T) {
+			db, err := readDatabaseForInspect(tt.format, filepath.Join(fixturesDir, tt.path), "")
+			if err != nil {
+				t.Skipf("format %s not supported by inspect reader: %v", tt.format, err)
+			}
+			if db == nil || len(db.Schemas) == 0 {
+				t.Errorf("no schemas: %+v", db)
+			}
+		})
+	}
+	for _, f := range []string{"dbml", "dctx", "drawdb", "graphql", "json", "yaml", "gorm", "bun", "drizzle", "prisma", "typeorm"} {
+		if _, err := readDatabaseForInspect(f, "", ""); err == nil || !strings.Contains(err.Error(), "file path is required") {
+			t.Errorf("%s missing path: %v", f, err)
+		}
+	}
+	if _, err := readDatabaseForInspect("pgsql", "", ""); err == nil {
+		t.Error("pgsql without conn must fail")
+	}
+	if _, err := readDatabaseForInspect("nope", "x", ""); err == nil || !strings.Contains(err.Error(), "unsupported database type") {
+		t.Errorf("unsupported: %v", err)
+	}
+}
+
+func TestFilterDatabaseBySchema(t *testing.T) {
+	db := multiSchemaDB()
+	db.Description = "desc"
+	got := filterDatabaseBySchema(db, "b")
+	if len(got.Schemas) != 1 || got.Schemas[0].Name != "b" || got.Name != db.Name || got.Description != "desc" {
+		t.Errorf("filtered: %+v", got)
+	}
+	if got := filterDatabaseBySchema(db, "zzz"); len(got.Schemas) != 0 {
+		t.Errorf("missing schema should yield no schemas: %+v", got.Schemas)
+	}
+	if len(db.Schemas) != 2 {
+		t.Error("input mutated")
+	}
+}
+
+func TestHasSilentFlag(t *testing.T) {
+	tests := []struct {
+		args []string
+		want bool
+	}{
+		{nil, false},
+		{[]string{"convert"}, false},
+		{[]string{"convert", "--silent"}, true},
+		{[]string{"--silent=true"}, true},
+		{[]string{"--silent=false"}, false},
+	}
+	for _, tt := range tests {
+		if got := hasSilentFlag(tt.args); got != tt.want {
+			t.Errorf("hasSilentFlag(%v) = %v", tt.args, got)
+		}
+	}
+}
+
+func TestPrintVersionHeader(t *testing.T) {
+	capture := func(args []string) string {
+		old := os.Stdout
+		r, w, _ := os.Pipe()
+		os.Stdout = w
+		printVersionHeader(args)
+		w.Close()
+		os.Stdout = old
+		b := make([]byte, 4096)
+		n, _ := r.Read(b)
+		return string(b[:n])
+	}
+	if out := capture([]string{"convert"}); !strings.HasPrefix(out, "RelSpec ") {
+		t.Errorf("header: %q", out)
+	}
+	if out := capture([]string{"convert", "--no-version"}); out != "" {
+		t.Errorf("--no-version: %q", out)
+	}
+	if out := capture([]string{"version"}); out != "" {
+		t.Errorf("version cmd: %q", out)
+	}
+	if out := capture(nil); !strings.HasPrefix(out, "RelSpec ") {
+		t.Errorf("no args: %q", out)
+	}
+}
+
+func TestReportState(t *testing.T) {
+	cfg := t.TempDir()
+	t.Setenv("XDG_CONFIG_HOME", cfg)
+	t.Setenv("HOME", cfg)
+
+	dir, err := reportStateDir()
+	if err != nil || !strings.HasPrefix(dir, cfg) {
+		t.Fatalf("dir: %q %v", dir, err)
+	}
+
+	state, path, err := loadReportState()
+	if err != nil || !state.LastReport.IsZero() || state.MachineID != "" {
+		t.Fatalf("fresh state: %+v %v", state, err)
+	}
+
+	want := reportState{LastReport: time.Now().UTC().Truncate(time.Second), MachineID: "abc"}
+	if err := saveReportState(path, want); err != nil {
+		t.Fatal(err)
+	}
+	got, _, err := loadReportState()
+	if err != nil || !got.LastReport.Equal(want.LastReport) || got.MachineID != "abc" {
+		t.Errorf("round trip: %+v %v", got, err)
+	}
+
+	// Corrupt state is ignored.
+	if err := os.WriteFile(path, []byte("{bad"), 0o600); err != nil {
+		t.Fatal(err)
+	}
+	if got, _, err := loadReportState(); err != nil || got.MachineID != "" {
+		t.Errorf("corrupt: %+v %v", got, err)
+	}
+}
+
+func TestSystemUniqueID_NonEmpty(t *testing.T) {
+	cfg := t.TempDir()
+	t.Setenv("XDG_CONFIG_HOME", cfg)
+	state, path, _ := loadReportState()
+	id, err := systemUniqueID(state, path)
+	if err != nil || id == "" {
+		t.Errorf("id: %q %v", id, err)
+	}
+}
+
+func TestReportToken_Decodes(t *testing.T) {
+	if _, err := reportToken(); err != nil {
+		t.Errorf("token must decode: %v", err)
+	}
+}
+
+func TestSubmitReport_RateLimited(t *testing.T) {
+	cfg := t.TempDir()
+	t.Setenv("XDG_CONFIG_HOME", cfg)
+	_, path, _ := loadReportState()
+	if err := saveReportState(path, reportState{LastReport: time.Now()}); err != nil {
+		t.Fatal(err)
+	}
+	// Rate limit rejects before any network call is made.
+	if err := submitReport("bug", "t", "b", "", ""); err == nil || !strings.Contains(err.Error(), "please wait") {
+		t.Errorf("got %v", err)
+	}
+}

cmd/relspec/prisma_options.go

@@ -27,6 +27,7 @@ func newWriterOptions(outputPath, packageName string, flattenSchema bool, nullab
 		FlattenSchema:    flattenSchema,
 		NullableTypes:    nullableTypes,
 		NullableArrays:   nullableArrays,
+		TypeMappings:     typeMappings,
 		Prisma7:          prisma7,
 		ContinueOnError:  continueOnError,
 		StrictDirectives: strictDirectives,

cmd/relspec/root.go

@@ -2,52 +2,26 @@ package main
 
 import (
 	"fmt"
-	"runtime/debug"
-	"time"
 
 	"github.com/spf13/cobra"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/buildinfo"
+	"git.warky.dev/wdevs/relspecgo/pkg/writers"
 )
 
+// version/buildDate mirror pkg/buildinfo so existing call sites keep working.
+// The actual values are set there via ldflags (see Makefile).
 var (
-	// Version information, set via ldflags during build
-	version          = "dev"
-	buildDate        = "unknown"
+	version          = buildinfo.Version
+	buildDate        = buildinfo.BuildDate
 	prisma7          bool
 	noVersion        bool
 	silent           bool
 	strictDirectives bool
+	typeMapFlags     []string
+	typeMappings     map[string]string
 )
 
-func init() {
-	// If version wasn't set via ldflags, try to get it from build info
-	if version == "dev" {
-		if info, ok := debug.ReadBuildInfo(); ok {
-			// Try to get version from VCS
-			var vcsRevision, vcsTime string
-			for _, setting := range info.Settings {
-				switch setting.Key {
-				case "vcs.revision":
-					if len(setting.Value) >= 7 {
-						vcsRevision = setting.Value[:7]
-					}
-				case "vcs.time":
-					vcsTime = setting.Value
-				}
-			}
-
-			if vcsRevision != "" {
-				version = vcsRevision
-			}
-
-			if vcsTime != "" {
-				if t, err := time.Parse(time.RFC3339, vcsTime); err == nil {
-					buildDate = t.UTC().Format("2006-01-02 15:04:05 UTC")
-				}
-			}
-		}
-	}
-}
-
 var rootCmd = &cobra.Command{
 	Use:   "relspec",
 	Short: "RelSpec - Database schema conversion and analysis tool",
@@ -57,10 +31,16 @@ bidirectional conversion between various database schema formats.
 It reads database schemas from multiple sources (live databases, DBML,
 DCTX, DrawDB, etc.) and writes them to various formats (GORM, Bun,
 JSON, YAML, SQL, etc.).`,
+	PersistentPreRunE: func(cmd *cobra.Command, args []string) error {
+		var err error
+		typeMappings, err = writers.ParseTypeMappings(typeMapFlags)
+		return err
+	},
 }
 
 func init() {
 	rootCmd.AddCommand(convertCmd)
+	rootCmd.AddCommand(batchCmd)
 	rootCmd.AddCommand(diffCmd)
 	rootCmd.AddCommand(inspectCmd)
 	rootCmd.AddCommand(scriptsCmd)
@@ -71,8 +51,10 @@ func init() {
 	rootCmd.AddCommand(mergeCmd)
 	rootCmd.AddCommand(splitCmd)
 	rootCmd.AddCommand(versionCmd)
+	rootCmd.AddCommand(updateCmd)
 	rootCmd.AddCommand(reportCmd)
 	rootCmd.PersistentFlags().BoolVar(&prisma7, "prisma7", false, "Use Prisma 7 generator conventions when reading/writing Prisma schemas")
+	rootCmd.PersistentFlags().StringArrayVar(&typeMapFlags, "type-map", nil, "Override a SQL-to-Go type mapping for bun/gorm output as sqltype=gotype (repeatable), e.g. --type-map uuid=uuid.UUID --type-map numeric=decimal.Decimal")
 	rootCmd.PersistentFlags().BoolVar(&noVersion, "no-version", false, "Suppress the RelSpec version header")
 	rootCmd.PersistentFlags().BoolVar(&silent, "silent", false, "Suppress progress and status messages (errors are still shown)")
 	rootCmd.PersistentFlags().BoolVar(&strictDirectives, "strict-directives", false, "Fail on unknown or untranslatable DBML dialect directives (@postgres:, @sqlite:, …)")

cmd/relspec/run_diff_inspect_test.go

@@ -0,0 +1,86 @@
+package main
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+)
+
+func TestRunDiff(t *testing.T) {
+	oldS, oldSP, oldSC, oldT, oldTP, oldTC, oldF, oldO := sourceType, sourcePath, sourceConn, targetType, targetPath, targetConn, outputFormat, outputPath
+	t.Cleanup(func() {
+		sourceType, sourcePath, sourceConn, targetType, targetPath, targetConn, outputFormat, outputPath = oldS, oldSP, oldSC, oldT, oldTP, oldTC, oldF, oldO
+	})
+
+	src := filepath.Join(fixturesDir, "dbml/simple.dbml")
+	cmplx := filepath.Join(fixturesDir, "dbml/complex.dbml")
+
+	for _, format := range []string{"summary", "json", "html"} {
+		t.Run(format, func(t *testing.T) {
+			sourceType, sourcePath, sourceConn = "dbml", src, ""
+			targetType, targetPath, targetConn = "dbml", cmplx, ""
+			outputFormat = format
+			outputPath = filepath.Join(t.TempDir(), "diff.out")
+			if format == "summary" {
+				outputPath = ""
+			}
+			if err := runDiff(nil, nil); err != nil {
+				t.Fatalf("runDiff: %v", err)
+			}
+			if outputPath != "" {
+				if b, err := os.ReadFile(outputPath); err != nil || len(b) == 0 {
+					t.Errorf("empty output: %v", err)
+				}
+			}
+		})
+	}
+
+	t.Run("bad source", func(t *testing.T) {
+		sourceType, sourcePath = "dbml", filepath.Join(t.TempDir(), "missing.dbml")
+		targetType, targetPath = "dbml", src
+		outputFormat, outputPath = "summary", ""
+		if err := runDiff(nil, nil); err == nil || !strings.Contains(err.Error(), "failed to read source database") {
+			t.Errorf("got %v", err)
+		}
+	})
+	t.Run("bad target", func(t *testing.T) {
+		sourceType, sourcePath = "dbml", src
+		targetType, targetPath = "dbml", filepath.Join(t.TempDir(), "missing.dbml")
+		outputFormat, outputPath = "summary", ""
+		if err := runDiff(nil, nil); err == nil || !strings.Contains(err.Error(), "failed to read target database") {
+			t.Errorf("got %v", err)
+		}
+	})
+}
+
+func TestRunInspect(t *testing.T) {
+	oldT, oldP, oldC, oldR, oldF, oldO, oldS := inspectSourceType, inspectSourcePath, inspectSourceConn, inspectRulesPath, inspectOutputFormat, inspectOutputPath, inspectSchemaFilter
+	t.Cleanup(func() {
+		inspectSourceType, inspectSourcePath, inspectSourceConn, inspectRulesPath, inspectOutputFormat, inspectOutputPath, inspectSchemaFilter = oldT, oldP, oldC, oldR, oldF, oldO, oldS
+	})
+
+	inspectSourceType = "dbml"
+	inspectSourcePath = filepath.Join(fixturesDir, "dbml/simple.dbml")
+	inspectSourceConn = ""
+	inspectRulesPath = filepath.Join(t.TempDir(), "no-rules.yaml") // missing: defaults used or error
+	inspectSchemaFilter = ""
+
+	// Whatever the rules outcome, the run must not panic; formats are exercised.
+	for _, format := range []string{"markdown", "json"} {
+		inspectOutputFormat = format
+		inspectOutputPath = filepath.Join(t.TempDir(), "report."+format)
+		_ = runInspect(nil, nil)
+	}
+
+	inspectOutputFormat = "bogus"
+	inspectOutputPath = ""
+	if err := runInspect(nil, nil); err == nil {
+		t.Error("bogus output format must fail")
+	}
+
+	inspectSourcePath = filepath.Join(t.TempDir(), "missing.dbml")
+	if err := runInspect(nil, nil); err == nil || !strings.Contains(err.Error(), "failed to read source") {
+		t.Errorf("missing source: %v", err)
+	}
+}

cmd/relspec/split.go

@@ -24,6 +24,7 @@ var (
 	splitExcludeTables  string
 	splitNullableTypes  string
 	splitNullableArrays string
+	splitDryRun         bool
 )
 
 var splitCmd = &cobra.Command{
@@ -115,6 +116,8 @@ func init() {
 	splitCmd.Flags().StringVar(&splitNullableTypes, "types", "", "Nullable type package for code-gen writers (bun/gorm): 'baselib' (default, Go pointer types), 'stdlib' (database/sql), or 'sqltypes'")
 	splitCmd.Flags().StringVar(&splitNullableArrays, "array-nullable", "", "Nullable PostgreSQL array representation for the Bun writer in stdlib/baselib --types mode: 'slice' (default, plain slice) or 'pointer_slice' (*[]T, distinguishes NULL from '{}')")
 
+	splitCmd.Flags().BoolVar(&splitDryRun, "dry-run", false, "Read, filter and validate the selection and print the plan without writing any output")
+
 	err := splitCmd.MarkFlagRequired("from")
 	if err != nil {
 		fmt.Fprintf(os.Stderr, "Error marking from flag as required: %v\n", err)
@@ -174,6 +177,26 @@ func runSplit(cmd *cobra.Command, args []string) error {
 	}
 	fmt.Fprintf(os.Stderr, "      ✓ Filtered to: %d schema(s), %d table(s)\n\n", len(filteredDB.Schemas), filteredTables)
 
+	if splitDryRun {
+		if err := validateWriteTarget(filteredDB, splitTargetType, splitPackageName, "", ""); err != nil {
+			return fmt.Errorf("dry run validation failed: %w", err)
+		}
+		out := outWriter(cmd)
+		fmt.Fprintf(out, "RelSpec split plan (dry run - nothing written):\n")
+		fmt.Fprintf(out, "  Input:  %s database '%s'\n", splitSourceType, db.Name)
+		fmt.Fprintf(out, "  Output: %s -> %s\n", splitTargetType, splitTargetPath)
+		fmt.Fprintf(out, "  Selection: %s\n", splitSelection{
+			Schemas:        parseCommaSeparated(splitSchemas),
+			Tables:         parseCommaSeparated(splitTables),
+			ExcludeSchemas: parseCommaSeparated(splitExcludeSchema),
+			ExcludeTables:  parseCommaSeparated(splitExcludeTables),
+			DatabaseName:   splitDatabaseName,
+		}.summary())
+		printDryRunPlan(out, filteredDB)
+		fmt.Fprintf(os.Stderr, "=== Dry run complete: no output written ===\n\n")
+		return nil
+	}
+
 	// Write to target format
 	fmt.Fprintf(os.Stderr, "[3/3] Writing to target format...\n")
 	fmt.Fprintf(os.Stderr, "      Format: %s\n", splitTargetType)

cmd/relspec/templ.go

@@ -114,6 +114,7 @@ func runTempl(cmd *cobra.Command, args []string) error {
 	if err != nil {
 		return fmt.Errorf("failed to read source: %w", err)
 	}
+	finalizeCommentedRefs(db, stderrWarn)
 
 	// Print database stats
 	schemaCount := len(db.Schemas)

cmd/relspec/update.go

@@ -0,0 +1,148 @@
+package main
+
+import (
+	"bufio"
+	"context"
+	"fmt"
+	"io"
+	"net/http"
+	"os"
+	"os/exec"
+	"path/filepath"
+	"runtime"
+	"strings"
+	"time"
+
+	"github.com/spf13/cobra"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/buildinfo"
+	"git.warky.dev/wdevs/relspecgo/pkg/updatecheck"
+)
+
+// windowsInstallerAsset is the release asset built by the NSIS installer step.
+const windowsInstallerAsset = "relspec-setup-windows-amd64.exe"
+
+var (
+	updateCheckOnly bool
+	updateAssumeYes bool
+	updateAPIURL    = updatecheck.DefaultAPIURL
+)
+
+var updateCmd = &cobra.Command{
+	Use:   "update",
+	Short: "Check for a newer release and offer to install it",
+	Long: `Check the project releases for a version newer than this binary.
+
+If one exists you are prompted to update. On Windows the installer is
+downloaded and started; on other platforms the release page URL is shown.
+Use --check to only report, and --yes to skip the prompt.`,
+	RunE: func(cmd *cobra.Command, args []string) error {
+		ctx, cancel := context.WithTimeout(cmd.Context(), 10*time.Minute)
+		defer cancel()
+		return runUpdate(ctx, updateOptions{
+			current:   buildinfo.Version,
+			apiURL:    updateAPIURL,
+			goos:      runtime.GOOS,
+			checkOnly: updateCheckOnly,
+			assumeYes: updateAssumeYes,
+			in:        cmd.InOrStdin(),
+			out:       cmd.OutOrStdout(),
+			install:   downloadAndRunInstaller,
+		})
+	},
+}
+
+func init() {
+	updateCmd.Flags().BoolVar(&updateCheckOnly, "check", false, "Only report whether an update is available")
+	updateCmd.Flags().BoolVarP(&updateAssumeYes, "yes", "y", false, "Update without prompting")
+}
+
+type updateOptions struct {
+	current   string
+	apiURL    string
+	goos      string
+	checkOnly bool
+	assumeYes bool
+	in        io.Reader
+	out       io.Writer
+	// install downloads and starts the installer found at url.
+	install func(ctx context.Context, url string, out io.Writer) error
+}
+
+func runUpdate(ctx context.Context, o updateOptions) error {
+	rel, err := updatecheck.Latest(ctx, nil, o.apiURL)
+	if err != nil {
+		return err
+	}
+
+	if !updatecheck.IsNewer(o.current, rel.Tag) {
+		_, _ = fmt.Fprintf(o.out, "RelSpec %s is up to date (latest release: %s)\n", o.current, rel.Tag)
+		return nil
+	}
+
+	_, _ = fmt.Fprintf(o.out, "A newer RelSpec is available: %s (installed: %s)\n", rel.Tag, o.current)
+	if o.checkOnly {
+		_, _ = fmt.Fprintf(o.out, "Release: %s\n", rel.URL)
+		return nil
+	}
+
+	installer, canInstall := rel.FindAsset(windowsInstallerAsset)
+	if o.goos != "windows" || !canInstall {
+		_, _ = fmt.Fprintf(o.out, "Download it from: %s\n", rel.URL)
+		return nil
+	}
+
+	if !o.assumeYes && !confirm(o.in, o.out, "Download and run the installer now?") {
+		_, _ = fmt.Fprintf(o.out, "Skipped. Release: %s\n", rel.URL)
+		return nil
+	}
+	return o.install(ctx, installer.URL, o.out)
+}
+
+// confirm asks a yes/no question, defaulting to no.
+func confirm(in io.Reader, out io.Writer, question string) bool {
+	_, _ = fmt.Fprintf(out, "%s [y/N]: ", question)
+	line, _ := bufio.NewReader(in).ReadString('\n')
+	switch strings.ToLower(strings.TrimSpace(line)) {
+	case "y", "yes":
+		return true
+	}
+	return false
+}
+
+// downloadAndRunInstaller saves the installer to a temp directory and starts
+// it detached so this process can exit and release relspec.exe.
+func downloadAndRunInstaller(ctx context.Context, url string, out io.Writer) error {
+	req, err := http.NewRequestWithContext(ctx, http.MethodGet, url, nil)
+	if err != nil {
+		return err
+	}
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		return fmt.Errorf("downloading installer: %w", err)
+	}
+	defer resp.Body.Close()
+	if resp.StatusCode != http.StatusOK {
+		return fmt.Errorf("downloading installer: unexpected status %s", resp.Status)
+	}
+
+	dir, err := os.MkdirTemp("", "relspec-update-")
+	if err != nil {
+		return err
+	}
+	path := filepath.Join(dir, windowsInstallerAsset)
+	f, err := os.Create(path)
+	if err != nil {
+		return err
+	}
+	if _, err := io.Copy(f, resp.Body); err != nil {
+		_ = f.Close()
+		return fmt.Errorf("downloading installer: %w", err)
+	}
+	if err := f.Close(); err != nil {
+		return err
+	}
+
+	_, _ = fmt.Fprintf(out, "Starting installer: %s\n", path)
+	return exec.Command(path).Start()
+}

cmd/relspec/update_test.go

@@ -0,0 +1,110 @@
+package main
+
+import (
+	"bytes"
+	"context"
+	"errors"
+	"fmt"
+	"io"
+	"net/http"
+	"net/http/httptest"
+	"strings"
+	"testing"
+)
+
+func releaseServer(t *testing.T, tag string, withInstaller bool) *httptest.Server {
+	t.Helper()
+	assets := ""
+	if withInstaller {
+		assets = fmt.Sprintf(`{"name":%q,"browser_download_url":"https://x/setup.exe"}`, windowsInstallerAsset)
+	}
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		_, _ = fmt.Fprintf(w, `{"tag_name":%q,"html_url":"https://x/release","assets":[%s]}`, tag, assets)
+	}))
+	t.Cleanup(srv.Close)
+	return srv
+}
+
+func TestRunUpdate(t *testing.T) {
+	tests := []struct {
+		name          string
+		current       string
+		latest        string
+		withInstaller bool
+		goos          string
+		checkOnly     bool
+		assumeYes     bool
+		stdin         string
+		wantOut       []string
+		wantInstall   bool
+	}{
+		{name: "up to date", current: "v1.0.5", latest: "v1.0.5", goos: "windows", withInstaller: true, wantOut: []string{"up to date"}},
+		{name: "dev build never prompts", current: "dev", latest: "v9.0.0", goos: "windows", withInstaller: true, wantOut: []string{"up to date"}},
+		{name: "check only", current: "v1.0.5", latest: "v1.0.6", goos: "windows", withInstaller: true, checkOnly: true, wantOut: []string{"v1.0.6", "https://x/release"}},
+		{name: "non-windows shows url", current: "v1.0.5", latest: "v1.0.6", goos: "linux", withInstaller: true, wantOut: []string{"Download it from: https://x/release"}},
+		{name: "windows without installer asset", current: "v1.0.5", latest: "v1.0.6", goos: "windows", wantOut: []string{"Download it from"}},
+		{name: "windows prompt yes", current: "v1.0.5", latest: "v1.0.6", goos: "windows", withInstaller: true, stdin: "y\n", wantOut: []string{"[y/N]"}, wantInstall: true},
+		{name: "windows prompt no", current: "v1.0.5", latest: "v1.0.6", goos: "windows", withInstaller: true, stdin: "n\n", wantOut: []string{"Skipped"}},
+		{name: "windows prompt empty", current: "v1.0.5", latest: "v1.0.6", goos: "windows", withInstaller: true, wantOut: []string{"Skipped"}},
+		{name: "windows --yes", current: "v1.0.5", latest: "v1.0.6", goos: "windows", withInstaller: true, assumeYes: true, wantInstall: true},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			srv := releaseServer(t, tt.latest, tt.withInstaller)
+			var out bytes.Buffer
+			var installedURL string
+			err := runUpdate(context.Background(), updateOptions{
+				current: tt.current, apiURL: srv.URL, goos: tt.goos,
+				checkOnly: tt.checkOnly, assumeYes: tt.assumeYes,
+				in: strings.NewReader(tt.stdin), out: &out,
+				install: func(_ context.Context, url string, _ io.Writer) error {
+					installedURL = url
+					return nil
+				},
+			})
+			if err != nil {
+				t.Fatal(err)
+			}
+			for _, want := range tt.wantOut {
+				if !strings.Contains(out.String(), want) {
+					t.Errorf("output missing %q:\n%s", want, out.String())
+				}
+			}
+			if (installedURL != "") != tt.wantInstall {
+				t.Errorf("installer called = %v, want %v", installedURL != "", tt.wantInstall)
+			}
+			if tt.wantInstall && installedURL != "https://x/setup.exe" {
+				t.Errorf("installer url = %q", installedURL)
+			}
+		})
+	}
+}
+
+func TestRunUpdate_Errors(t *testing.T) {
+	bad := httptest.NewServer(http.NotFoundHandler())
+	defer bad.Close()
+	if err := runUpdate(context.Background(), updateOptions{current: "v1.0.0", apiURL: bad.URL, out: io.Discard}); err == nil {
+		t.Error("expected lookup error")
+	}
+
+	srv := releaseServer(t, "v1.0.6", true)
+	want := errors.New("boom")
+	err := runUpdate(context.Background(), updateOptions{
+		current: "v1.0.5", apiURL: srv.URL, goos: "windows", assumeYes: true, out: io.Discard,
+		install: func(context.Context, string, io.Writer) error { return want },
+	})
+	if !errors.Is(err, want) {
+		t.Errorf("err = %v", err)
+	}
+}
+
+func TestDownloadAndRunInstaller_BadStatus(t *testing.T) {
+	srv := httptest.NewServer(http.NotFoundHandler())
+	defer srv.Close()
+	if err := downloadAndRunInstaller(context.Background(), srv.URL, io.Discard); err == nil {
+		t.Error("expected error for non-200")
+	}
+	if err := downloadAndRunInstaller(context.Background(), "://bad", io.Discard); err == nil {
+		t.Error("expected error for bad url")
+	}
+}

cmd/relspec/version.go

@@ -4,13 +4,16 @@ import (
 	"fmt"
 
 	"github.com/spf13/cobra"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/buildinfo"
 )
 
 var versionCmd = &cobra.Command{
 	Use:   "version",
 	Short: "Print version information",
 	Run: func(cmd *cobra.Command, args []string) {
-		fmt.Printf("RelSpec %s\n", version)
-		fmt.Printf("Built: %s\n", buildDate)
+		fmt.Print(buildinfo.AsciiLogo)
+		fmt.Printf("RelSpec %s\n", buildinfo.Version)
+		fmt.Printf("Built: %s\n", buildinfo.BuildDate)
 	},
 }

cmd/relspec/watch.go

@@ -0,0 +1,152 @@
+package main
+
+import (
+	"context"
+	"fmt"
+	"io"
+	"io/fs"
+	"os"
+	"os/signal"
+	"path/filepath"
+	"strings"
+	"syscall"
+	"time"
+)
+
+// watchSnapshot maps a file path to its modification time and size.
+type watchSnapshot map[string]string
+
+// takeWatchSnapshot records the state of every file under the given paths.
+// Directories are walked recursively. Anything at or below the excluded path
+// (typically the output path) is skipped so regenerating output does not
+// retrigger the watcher. Missing paths are simply absent from the snapshot, so
+// creating them later counts as a change.
+func takeWatchSnapshot(paths []string, exclude string) watchSnapshot {
+	snap := watchSnapshot{}
+	exclude = absPathOrSelf(exclude)
+	for _, root := range paths {
+		_ = filepath.WalkDir(root, func(p string, d fs.DirEntry, err error) error {
+			if err != nil {
+				return nil
+			}
+			if exclude != "" && isWithin(absPathOrSelf(p), exclude) {
+				if d.IsDir() {
+					return filepath.SkipDir
+				}
+				return nil
+			}
+			if d.IsDir() {
+				return nil
+			}
+			info, err := d.Info()
+			if err != nil {
+				return nil
+			}
+			snap[p] = fmt.Sprintf("%d-%d", info.ModTime().UnixNano(), info.Size())
+			return nil
+		})
+	}
+	return snap
+}
+
+func (s watchSnapshot) equal(o watchSnapshot) bool {
+	if len(s) != len(o) {
+		return false
+	}
+	for k, v := range s {
+		if o[k] != v {
+			return false
+		}
+	}
+	return true
+}
+
+func absPathOrSelf(p string) string {
+	if p == "" {
+		return ""
+	}
+	if abs, err := filepath.Abs(p); err == nil {
+		return abs
+	}
+	return p
+}
+
+// isWithin reports whether path equals dir or is located below it.
+func isWithin(path, dir string) bool {
+	if path == dir {
+		return true
+	}
+	return strings.HasPrefix(path, dir+string(filepath.Separator))
+}
+
+// watchLoop runs fn once immediately and again whenever the watched paths
+// change, until ctx is cancelled. Changes are debounced: fn runs only after
+// the snapshot has stayed unchanged for one poll interval. Errors from fn are
+// reported to w and do not stop the loop.
+func watchLoop(ctx context.Context, w io.Writer, paths []string, exclude string, interval time.Duration, fn func() error) {
+	run := func() {
+		if err := fn(); err != nil {
+			fmt.Fprintf(w, "Error: %v\n", err)
+		}
+		fmt.Fprintf(w, "Watching for changes (Ctrl-C to stop)...\n")
+	}
+
+	last := takeWatchSnapshot(paths, exclude)
+	run()
+
+	ticker := time.NewTicker(interval)
+	defer ticker.Stop()
+	for {
+		select {
+		case <-ctx.Done():
+			return
+		case <-ticker.C:
+		}
+		cur := takeWatchSnapshot(paths, exclude)
+		if cur.equal(last) {
+			continue
+		}
+		// Debounce: wait until writes settle.
+		for settled := false; !settled; {
+			select {
+			case <-ctx.Done():
+				return
+			case <-time.After(interval):
+			}
+			next := takeWatchSnapshot(paths, exclude)
+			settled = next.equal(cur)
+			cur = next
+		}
+		fmt.Fprintf(w, "\nChange detected, regenerating...\n")
+		last = cur
+		run()
+	}
+}
+
+// runConvertWatch runs the conversion once and then again whenever the source
+// files change, until interrupted.
+func runConvertWatch(parent context.Context, w io.Writer, run func() error) error {
+	var paths []string
+	switch {
+	case len(convertFromList) > 0:
+		paths = convertFromList
+	case convertSourcePath != "":
+		paths = []string{convertSourcePath}
+	default:
+		return fmt.Errorf("--watch requires --from-path or --from-list (live database connections cannot be watched)")
+	}
+	if convertDryRun {
+		return fmt.Errorf("--watch cannot be combined with --dry-run")
+	}
+	if convertWatchInterval <= 0 {
+		return fmt.Errorf("--watch-interval must be positive")
+	}
+	if parent == nil {
+		parent = context.Background()
+	}
+
+	ctx, stop := signal.NotifyContext(parent, os.Interrupt, syscall.SIGTERM)
+	defer stop()
+	watchLoop(ctx, w, paths, convertTargetPath, convertWatchInterval, run)
+	return nil
+}

cmd/relspec/watch_test.go

@@ -0,0 +1,92 @@
+package main
+
+import (
+	"bytes"
+	"context"
+	"io"
+	"os"
+	"path/filepath"
+	"sync/atomic"
+	"testing"
+	"time"
+)
+
+func TestWatchSnapshotExcludesOutput(t *testing.T) {
+	dir := t.TempDir()
+	out := filepath.Join(dir, "out")
+	if err := os.MkdirAll(out, 0o755); err != nil {
+		t.Fatal(err)
+	}
+	src := filepath.Join(dir, "schema.dbml")
+	if err := os.WriteFile(src, []byte("a"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	before := takeWatchSnapshot([]string{dir}, out)
+	if err := os.WriteFile(filepath.Join(out, "gen.go"), []byte("x"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	if after := takeWatchSnapshot([]string{dir}, out); !before.equal(after) {
+		t.Errorf("writing into the excluded output path changed the snapshot")
+	}
+	if err := os.WriteFile(src, []byte("changed"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	if after := takeWatchSnapshot([]string{dir}, out); before.equal(after) {
+		t.Errorf("modifying a source file did not change the snapshot")
+	}
+}
+
+func TestWatchLoopRerunsOnChange(t *testing.T) {
+	dir := t.TempDir()
+	src := filepath.Join(dir, "schema.dbml")
+	if err := os.WriteFile(src, []byte("a"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	var runs atomic.Int32
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+	done := make(chan struct{})
+	go func() {
+		defer close(done)
+		watchLoop(ctx, io.Discard, []string{src}, "", 10*time.Millisecond, func() error {
+			runs.Add(1)
+			return nil
+		})
+	}()
+
+	waitFor(t, func() bool { return runs.Load() == 1 })
+	if err := os.WriteFile(src, []byte("changed content"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	waitFor(t, func() bool { return runs.Load() == 2 })
+	cancel()
+	<-done
+}
+
+func TestRunConvertWatchValidation(t *testing.T) {
+	oldPath, oldList, oldDry, oldInt := convertSourcePath, convertFromList, convertDryRun, convertWatchInterval
+	defer func() {
+		convertSourcePath, convertFromList, convertDryRun, convertWatchInterval = oldPath, oldList, oldDry, oldInt
+	}()
+	convertSourcePath, convertFromList, convertDryRun, convertWatchInterval = "", nil, false, time.Second
+	if err := runConvertWatch(context.Background(), &bytes.Buffer{}, nil); err == nil {
+		t.Error("expected error without --from-path/--from-list")
+	}
+	convertSourcePath, convertDryRun = "x.dbml", true
+	if err := runConvertWatch(context.Background(), &bytes.Buffer{}, nil); err == nil {
+		t.Error("expected error combining --watch with --dry-run")
+	}
+}
+
+func waitFor(t *testing.T, cond func() bool) {
+	t.Helper()
+	deadline := time.Now().Add(5 * time.Second)
+	for time.Now().Before(deadline) {
+		if cond() {
+			return
+		}
+		time.Sleep(5 * time.Millisecond)
+	}
+	t.Fatal("condition not met in time")
+}

docs/FORMAT_EXAMPLES.md

@@ -0,0 +1,103 @@
+# Format Usage Examples
+
+Examples for `relspec convert` covering the file-based reader and writer
+formats. The "Writers" and "Readers" sections below were run against
+`examples/test_schema.dbml`. The cross-format and live-database examples were
+not run; they follow the flags shown in `relspec convert --help` and require
+matching input files or reachable databases.
+
+Any reader can be combined with any writer: pick `--from`/`--from-path` for the
+source and `--to`/`--to-path` for the target. Add `--silent` to suppress progress
+output.
+
+## Writers: DBML to every format
+
+```bash
+S="--from dbml --from-path examples/test_schema.dbml"
+
+relspec convert $S --to json    --to-path schema.json
+relspec convert $S --to yaml    --to-path schema.yaml
+relspec convert $S --to dctx    --to-path schema.dctx
+relspec convert $S --to drawdb  --to-path schema.drawdb.json
+relspec convert $S --to graphql --to-path schema.graphql
+relspec convert $S --to prisma  --to-path schema.prisma
+relspec convert $S --to pgsql   --to-path schema.pg.sql
+relspec convert $S --to mssql   --to-path schema.mssql.sql
+relspec convert $S --to sqlite  --to-path schema.sqlite.sql
+relspec convert $S --to drizzle --to-path schema.ts
+relspec convert $S --to typeorm --to-path entities.ts
+relspec convert $S --to gorm    --to-path models.go --package models
+relspec convert $S --to bun     --to-path models.go --package models
+```
+
+Notes:
+
+- Code-generation writers (`gorm`, `bun`) take `--package`. They also accept
+  `--types baselib|stdlib|sqltypes` to choose the nullable type package.
+- When `--to-path` is a directory it must already exist.
+- `sqlite` output automatically flattens `schema.table` names. Use
+  `--flatten-schema` for other formats if the target has no schema support.
+- `dctx` supports a single schema only; use `--schema <name>` to select one.
+
+## Readers: file-based formats into DBML (or JSON where noted)
+
+```bash
+relspec convert --from json     --from-path schema.json         --to dbml --to-path out.dbml
+relspec convert --from yaml     --from-path schema.yaml         --to dbml --to-path out.dbml
+relspec convert --from dctx     --from-path schema.dctx         --to dbml --to-path out.dbml
+relspec convert --from drawdb   --from-path schema.drawdb.json  --to dbml --to-path out.dbml
+relspec convert --from graphql  --from-path schema.graphql      --to dbml --to-path out.dbml
+relspec convert --from prisma   --from-path schema.prisma       --to dbml --to-path out.dbml
+relspec convert --from drizzle  --from-path schema.ts           --to dbml --to-path out.dbml
+relspec convert --from typeorm  --from-path entities.ts         --to dbml --to-path out.dbml
+relspec convert --from bun      --from-path models.go           --to dbml --to-path out.dbml
+relspec convert --from gorm     --from-path models.go           --to json --to-path out.json
+```
+
+Code-first readers (`gorm`, `bun`, `drizzle`, `typeorm`) accept a single file or a
+directory of model files.
+
+> Known issue: reading GORM models and writing DBML currently panics in the DBML
+> writer (`pkg/writers/dbml/writer.go`, `constraintToDBML`). Use another target
+> such as JSON until this is fixed.
+
+## Cross-format combinations
+
+```bash
+# ORM models to SQL DDL
+relspec convert --from gorm --from-path models.go --to pgsql --to-path schema.sql
+
+# Prisma to Drizzle
+relspec convert --from prisma --from-path schema.prisma --to drizzle --to-path schema.ts
+
+# DrawDB diagram to GraphQL
+relspec convert --from drawdb --from-path diagram.json --to graphql --to-path schema.graphql
+
+# Merge several files while converting
+relspec convert --from json --from-list "a.json,b.json" --to yaml --to-path merged.yaml
+```
+
+## Live databases
+
+These need a reachable database:
+
+```bash
+# PostgreSQL
+relspec convert --from pgsql --from-conn "postgres://user:pass@localhost:5432/mydb" \
+                --to dbml --to-path schema.dbml
+
+# SQL Server
+relspec convert --from mssql --from-conn "<mssql connection string>" \
+                --to json --to-path schema.json
+
+# SQLite database file (--from-conn takes the file path)
+relspec convert --from sqlite --from-conn ./app.db --to dbml --to-path schema.dbml
+```
+
+## Formats outside `convert`
+
+- `sqldir` (SQL script directory reader) and `sqlexec` (SQL execution writer) are
+  used by `relspec scripts` and `relspec job`, and `sqldir` by `relspec diff`.
+  See [SCRIPTS_COMMAND.md](SCRIPTS_COMMAND.md) and [JOB_FILES.md](JOB_FILES.md).
+- The `template` writer is exposed through `relspec templ`. See
+  [TEMPLATE_MODE.md](TEMPLATE_MODE.md).

docs/JOB_FILES.md

@@ -55,6 +55,10 @@ already forbid duplicate keys within a single file.
   `script_dirs[]`, `template`, `logfile`) is **relative to the directory
   containing the job file that declared the job**, not the process working
   directory.
+* Those path fields support `${NAME}` environment-variable references. They
+  are expanded immediately before execution; a missing variable is an error.
+  The expanded value is still subject to all relative-path and symlink safety
+  checks.
 * Absolute paths, `~`-relative paths and any path that resolves outside the job
   file directory (`../`, `a/../../b`, …) are **rejected during validation** —
   before anything runs.
@@ -183,12 +187,17 @@ jobs:
       format: pgsql
       path: build/schema.sql                   # file output, OR:
       conn_env: TARGET_DB_URL                  # execute against DB (pgsql only)
+                                               # default: reads live DB, executes only the diff
+                                               # (file output = full DDL); set options.full_ddl to force full DDL
       overwrite: false                         # default false
     options:
       flatten_schema: false
       schema: public
       package: models                          # for gorm/bun output
+      types: sqltypes                           # Bun/GORM nullable types: baselib|stdlib|sqltypes
+      array_nullable: pointer_slice             # Bun: slice|pointer_slice
       continue_on_error: false                 # pgsql / scripts-exec output
+      full_ddl: false                          # pgsql database output: true = run full DDL; default diffs against live DB
       skip_relations: false                    # merge only
       skip_enums: false
       skip_views: false
@@ -201,8 +210,10 @@ jobs:
 
 For `templ`, `inputs` use the same file or `pgsql`/`conn_env` source forms as
 schema conversion. `output` is optional (empty means stdout); when present it
-contains only `path` and `overwrite`, because templates do not select a schema
-writer format.
+contains `path` and `overwrite`, plus the explicit `format: text` marker. The
+marker is templ-only: it describes arbitrary text produced by a Go template,
+not a database schema format. It is optional for compatibility with older
+manifests and has no effect on template execution.
 
 A `from_job` input takes no `path`, `format` or `conn_env`: it resolves to the
 named job's `output.path` and inherits its format, and implies a dependency on

docs/TUI_MOUSE_SUPPORT_PLAN.md

@@ -0,0 +1,212 @@
+# TUI mouse support plan
+
+Issue: #46
+Status: design only; this document does not implement mouse input.
+
+## 1. Current implementation and scope
+
+The editor is created in `cmd/relspec/edit.go` by
+`ui.NewSchemaEditorWithConfigs(...).Run()`. `pkg/ui/editor.go` owns the
+`tview.Application`, `tview.Pages`, and application lifecycle. The current
+code never calls `Application.EnableMouse`, so tcell mouse reporting is off.
+The module uses tview v0.42.0 and tcell/v2 v2.13.9.
+
+The first implementation should add `--no-mouse` to the `edit` Cobra command
+only. The flag is a local boolean, defaulting to false, and should be passed
+explicitly into the editor (prefer an options/config field rather than a
+package-global or environment variable). It must not affect convert, inspect,
+merge, or other commands. There is no environment-variable or persistent
+configuration setting in this issue: a command-line opt-out is predictable,
+visible in `edit --help`, and avoids adding configuration precedence rules.
+
+At startup, the editor should call `app.EnableMouse(!noMouse)` before
+`Run()`. `--no-mouse` must mean that the application does not enable terminal
+mouse reporting and that no custom mouse handlers are relied upon. Keyboard
+behavior must remain identical in both modes.
+
+Likely implementation files are `cmd/relspec/edit.go`,
+`pkg/ui/editor.go`, focused TUI mouse helpers/tests under `pkg/ui`, and a
+short user-facing note in the command help or TUI documentation. Do not
+refactor unrelated screens or data operations.
+
+## 2. Widget and screen coverage
+
+The application composes `Pages`, `Flex`, `TextView`, `List`, `Table`, `Form`,
+`Button`, `InputField`, `DropDown`, `TextArea`, `CheckBox`, and `Modal`.
+Vendored tview confirms mouse handlers exist for all of those relevant
+primitives, including focus on left-down, list/table selection, button clicks,
+form child dispatch, dropdown opening/drag selection, text-area cursor and
+scrolling, and modal button dispatch. `Pages`, `Flex`, and `Form` forward events
+to their children.
+
+The coverage plan is:
+
+* Main menu (`pkg/ui/main_menu.go`): left click focuses/selects a list entry;
+  second activation opens it; buttons and exit confirmation remain reachable.
+* Schema, table, domain, object, relation, and database screens: click a row
+  to select it; double-click the row to perform the same action as the
+  keyboard Enter/selected callback where opening is meaningful; scroll lists
+  and tables; click each action button.
+* Tables (`schema_screens.go`, `table_screens.go`, and object/relation tables):
+  tview's table handler provides selection and scrolling, but it does not
+  provide application-specific double-click activation. Add a small reusable
+  wrapper/helper for the table instances that need it. It must preserve the
+  existing selected row/column behavior and invoke the same callback as Enter,
+  not duplicate mutation logic.
+* Forms (`load_save_screens.go` and the form-building screen files): click an
+  input to focus it, click buttons to activate them, click a dropdown to open
+  it and choose an option, scroll multiline help/text areas, and retain all
+  existing keyboard Tab/Shift-Tab, shortcut, Enter, and Escape behavior.
+* Dialogs (`pkg/ui/dialogs.go` plus confirmation/error/success modals): modal
+  buttons are clickable and the modal keeps focus above the underlying page.
+  Clicking outside a modal must not activate the hidden page or dismiss a
+  destructive confirmation. Escape and the existing button-key behavior stay
+  authoritative.
+* The planned file browser and connection-string builder from issue #44 must
+  use the same contracts: clickable entries/buttons and scrolling, with
+  keyboard navigation and explicit cancel/accept paths. #46 should not
+  implement #44's widgets; it should define the integration point and test
+  them when #44 lands.
+
+Do not promise drag semantics for every widget. Drag is appropriate for text
+selection/cursor movement and dropdown selection where tview already supports
+it. For ordinary list/table navigation, a click selects and the wheel scrolls;
+row dragging should not mutate data.
+
+## 3. Exact mouse action contract
+
+| Widget/type | Left down/click | Double click | Wheel/drag | Keyboard fallback |
+| --- | --- | --- | --- | --- |
+| Main/list menu | focus and select row | invoke row selected callback | scroll list | arrows, Enter, shortcuts |
+| Data table | focus and select cell/row | invoke the screen's existing open/edit action for the selected row | vertical/horizontal scroll as supported by tview | arrows, PageUp/PageDown, Enter, existing shortcuts |
+| Button | focus | same as one activation, never duplicate the callback | none | Tab/Shift-Tab, Enter/Space and existing shortcut |
+| Input field | focus; place cursor if supported | no destructive action | text-area behavior if provided by tview | typing, arrows, Home/End, Tab, Escape |
+| Text area/help | focus and position cursor | select word only where tview supports it; no application action | scroll; drag text selection if supported | arrows, PageUp/PageDown, standard editing keys |
+| Dropdown | focus/open and choose the hit option | same as click; no duplicate selection | drag through options only while open | arrows, Enter, Escape, Tab |
+| Checkbox | toggle on click | no second toggle | none | Space and existing form navigation |
+| Modal | focus/click visible button | same button action once | no underlying-page scrolling | Tab/Shift-Tab, Enter, Escape, existing button keys |
+| Blank/border/title area | focus containing primitive where useful | none | no mutation | current screen shortcuts |
+
+Right and middle clicks should have no application action in the first
+release. Wheel events should be consumed only by the scrollable primitive
+under the pointer. Double-click timing/translation should come from tview/
+tcell; custom code must not fire the action once for both the click and the
+double-click. Any custom table wrapper needs a small state machine or tview
+mouse action handling that is tested for this property.
+
+## 4. tview gaps and implementation boundaries
+
+Enabling mouse support is not sufficient for the desired behavior. tview's
+built-in Table handler selects cells and scrolls but has no repository-specific
+row-open callback on double click. Existing screen code also wires keyboard
+input captures directly on individual widgets, so mouse actions must call the
+same screen callbacks rather than route through synthetic key events.
+
+Use tview's `MouseHandler`/`WrapMouseHandler` contracts and `setFocus` rather
+than reading terminal coordinates in each screen. A reusable table adapter
+may embed `*tview.Table`, delegate ordinary actions to the original table
+handler, and add the screen's double-click callback. Keep the adapter in
+`pkg/ui` and use it only where a row-opening action exists. Do not modify the
+vendored tview copy.
+
+The `Pages`/`Modal` dispatch order must be verified: a visible modal consumes
+its click before the page below it. Page transitions should happen only in the
+existing callbacks, so a stale hidden page cannot receive a click.
+
+## 5. Keyboard, terminal, and copy/paste behavior
+
+Mouse is an enhancement, never a requirement. Every acceptance path must be
+reachable with the existing keyboard controls, including load/save, navigation,
+editing, confirmations, cancel, and exit. `--no-mouse` is the regression mode
+for proving this contract.
+
+Mouse reporting is terminal capability dependent. On local terminals it is
+negotiated by tcell; tmux and SSH can suppress, translate, or fail to pass
+mouse reporting depending on their configuration. The application must still
+start and remain keyboard usable if mouse reporting is unavailable or broken.
+Documentation should state that terminal/tmux configuration may be required,
+and that SSH behavior depends on the remote terminal path. Windows Terminal and
+other Windows console hosts should be treated as supported only insofar as the
+selected tcell backend reports mouse events; the CLI must not assume POSIX
+escape sequences or add platform-specific code in this issue.
+
+Enabling mouse capture normally prevents terminal-native selection/copy from
+seeing ordinary button-drag events. Document the standard workaround: hold the
+terminal's bypass modifier (commonly Shift, terminal-dependent) for selection,
+or use `--no-mouse` when native copy/paste is the priority. Do not implement a
+second clipboard protocol. Input-field/text-area copy/paste must continue to
+use tview/tcell paste handling and keyboard shortcuts; verify that enabling
+mouse does not intercept paste events.
+
+## 6. Test strategy using tcell simulation
+
+Add focused tests rather than attempting a full interactive end-to-end test.
+Use `tcell.NewSimulationScreen("")`, `screen.Init()`, construct the editor or
+an isolated primitive tree, and inject events with the actual API:
+`SimulationScreen.InjectMouse(x, y, buttons, mod)` and `InjectKey(...)`.
+Coordinates must be derived from the primitive's drawn rectangle or fixed by a
+small deterministic test layout; do not use arbitrary coordinates without
+checking the rendered screen.
+
+Minimum cases:
+
+1. Default editor configuration enables mouse; the explicit disabled option
+   leaves it disabled. If the Application API is not observable directly,
+   test through the simulation screen's event path plus a constructor-level
+   option assertion.
+2. A list click changes focus/selection, and double-click invokes the existing
+   selected action exactly once.
+3. A table click selects the expected row/cell, wheel events change the visible
+   offset, and double-click invokes the row action exactly once.
+4. Form button, input field, checkbox, and dropdown clicks match their
+   keyboard callbacks.
+5. A modal button click acts on the modal and cannot activate the underlying
+   page; Escape still cancels.
+6. `--no-mouse` leaves keyboard selection/activation unchanged and mouse
+   injection has no application effect.
+7. Existing dialogs and screen transitions do not leave a stale mouse capture
+   after a page is removed.
+
+Prefer callback counters and selected-index assertions over screen-text-only
+assertions. Run the relevant `pkg/ui` tests with `go test -race ./pkg/ui` and
+run the full repository test suite if time/resources permit.
+
+## 7. Rollout and acceptance criteria
+
+Implementation is ready for review when:
+
+* `relspec edit --help` documents `--no-mouse` and mouse is enabled by default.
+* Only the edit TUI is affected; non-TUI commands have no changed behavior.
+* Main screens, tables, lists, forms, dropdowns, buttons, text areas, and
+  visible dialogs support the action contract above.
+* Keyboard-only operation is complete and verified with `--no-mouse`.
+* Modal clicks cannot fall through to an underlying page.
+* Table double-click behavior is explicit, tested, and does not duplicate
+  activation.
+* tcell simulation tests cover default-on, opt-out, selection, scrolling,
+  activation, dialog focus, and keyboard fallback.
+* `go test -race ./pkg/ui`, appropriate command tests, `go test ./...`,
+  formatting, and `git diff --check` pass (or any limitation is recorded).
+* User-facing docs explain tmux/SSH/Windows variability and the terminal
+  modifier workaround for native copy/paste.
+
+Roll out in two implementation slices if needed: first application option,
+standard tview handlers, tests, and documentation; second only the reusable
+table double-click adapter and screen wiring. Do not block the first slice on
+issue #44, but do not claim #44's future widgets are covered until they use the
+same contract.
+
+## 8. Open decisions and dependencies
+
+* Confirm whether the project wants a public editor options type or a small
+  `SetMouseEnabled`/constructor parameter; avoid a global flag.
+* Confirm the preferred terminal copy modifier in project documentation, since
+  tmux, SSH clients, and Windows Terminal differ.
+* Decide whether horizontal wheel events should be supported where tview/table
+  exposes them; vertical scrolling is mandatory, horizontal is optional.
+* Decide whether double-click opens every data table or only tables with an
+  unambiguous row action. The plan recommends the latter.
+* Confirm #44's file-browser and connection-builder primitive choices before
+  wiring their mouse tests.
+* Confirm CI has a stable non-terminal environment for simulation-screen tests;
+  no real terminal, tmux session, database, or network should be required.

examples/jobs/README.md

@@ -0,0 +1,80 @@
+# RelSpec job examples
+
+Run these commands from this directory:
+
+```bash
+cd examples/jobs
+
+# List every job from relspec.yml and the additional relspec.*.yml files.
+relspec job list
+
+# Validate a job and print its resolved plan without reading or writing data.
+relspec job run build-schema --plan
+
+# Build PostgreSQL DDL from the two DBML inputs.
+relspec job run build-schema
+
+# Run a job's dependency chain. This runs build-schema first, then build-json.
+relspec job run build-json
+
+# Inspect build-json's output. The from_job input adds the dependency automatically.
+relspec job run lint-schema
+
+# Extract only the posts table.
+relspec job run posts-only
+
+# List migration scripts in deterministic order without connecting to a database.
+relspec job run migration-order
+
+# Render one TypeScript file per table using the example template.
+relspec job run generate-typescript
+
+# Compare the two example schemas and write the summary to the logfile.
+relspec job run compare-schemas
+
+# Use an environment variable in an output path.
+export VST=build/typescript
+relspec job run generate-typescript --plan
+relspec job run generate-typescript
+```
+
+Database jobs use environment variables for connection strings. Set them in
+the shell before running the jobs; do not put connection strings in YAML:
+
+```bash
+export SOURCE_DB_URL='postgres://user:password@localhost/source_db'
+export TARGET_DB_URL='postgres://user:password@localhost/target_db'
+
+# Read the source database and write a local DBML snapshot.
+relspec job run snapshot-database
+
+# Preview migration execution without connecting to PostgreSQL.
+relspec job run apply-migrations --plan
+
+# Execute the migrations against TARGET_DB_URL.
+relspec job run apply-migrations
+```
+
+The `apply-migrations` job demonstrates execution from multiple directories:
+
+```yaml
+script_dirs:
+  - migrations/core
+  - migrations/tenant
+output:
+  format: pgsql
+  conn_env: TARGET_DB_URL
+```
+
+RelSpec combines the SQL files from both folders in deterministic order before
+executing them against the configured PostgreSQL database.
+
+To use a different directory or manifest, pass `--dir` or `--file`:
+
+```bash
+relspec job list --dir /path/to/project
+relspec job run build-schema --file /path/to/project/relspec.yml --plan
+```
+
+See [../../docs/JOB_FILES.md](../../docs/JOB_FILES.md) for the complete job
+file reference.

examples/jobs/relspec.database.yml

@@ -0,0 +1,29 @@
+# Database-oriented examples. Set the referenced environment variables before
+# running these jobs; connection strings never belong in a job file.
+version: 1
+
+jobs:
+  snapshot-database:
+    command: convert
+    description: Read a PostgreSQL database and write a DBML snapshot
+    inputs:
+      - format: pgsql
+        conn_env: SOURCE_DB_URL
+    output:
+      format: dbml
+      path: build/database-snapshot.dbml
+      overwrite: true
+
+  apply-migrations:
+    command: scripts-exec
+    description: Execute migrations from core and tenant folders against PostgreSQL
+    # All SQL files from these directories are discovered in deterministic order.
+    script_dirs:
+      - migrations/core
+      - migrations/tenant
+    output:
+      format: pgsql
+      conn_env: TARGET_DB_URL
+    options:
+      continue_on_error: false
+    logfile: .relspec/log/apply-migrations.log

examples/jobs/relspec.reports.yml

@@ -0,0 +1,32 @@
+# Reporting and templating examples. These jobs are discovered together with
+# relspec.yml when running `relspec job list` from this directory.
+version: 1
+
+jobs:
+  generate-typescript:
+    command: templ
+    description: Render one TypeScript file per table from the merged schema
+    inputs:
+      - path: schema/core.dbml
+        format: dbml
+      - path: schema/tenant.dbml
+        format: dbml
+    template: templates/schema.tmpl
+    mode: table
+    filename_pattern: "{{.Name}}.ts"
+    output:
+      format: text
+      path: "${VST}"
+      overwrite: true
+
+  compare-schemas:
+    command: diff
+    description: Compare the two example schemas and print a summary
+    inputs:
+      - path: schema/core.dbml
+        format: dbml
+      - path: schema/tenant.dbml
+        format: dbml
+    report:
+      format: summary
+    logfile: .relspec/log/compare-schemas.log

examples/jobs/templates/schema.tmpl

@@ -0,0 +1,6 @@
+// Code generated by RelSpec. DO NOT EDIT.
+export interface {{.Name}} {
+{{- range values .Table.Columns}}
+  {{.Name}}: {{.Type}};
+{{- end}}
+}

go.mod

@@ -4,6 +4,7 @@ go 1.25.13
 
 require (
 	github.com/gdamore/tcell/v2 v2.13.9
+	github.com/go-sql-driver/mysql v1.9.3
 	github.com/google/uuid v1.6.0
 	github.com/jackc/pgx/v5 v5.9.2
 	github.com/microsoft/go-mssqldb v1.10.0
@@ -18,6 +19,7 @@ require (
 )
 
 require (
+	filippo.io/edwards25519 v1.1.0 // indirect
 	github.com/davecgh/go-spew v1.1.1 // indirect
 	github.com/dustin/go-humanize v1.0.1 // indirect
 	github.com/gdamore/encoding v1.0.1 // indirect

go.sum

@@ -1,3 +1,5 @@
+filippo.io/edwards25519 v1.1.0 h1:FNf4tywRC1HmFuKW5xopWpigGjJKiJSV0Cqo0cJWDaA=
+filippo.io/edwards25519 v1.1.0/go.mod h1:BxyFTGdWcka3PhytdK4V28tE5sGfRvvvRV7EaN4VDT4=
 github.com/Azure/azure-sdk-for-go/sdk/azcore v1.21.1 h1:jHb/wfvRikGdxMXYV3QG/SzUOPYN9KEUUuC0Yd0/vC0=
 github.com/Azure/azure-sdk-for-go/sdk/azcore v1.21.1/go.mod h1:pzBXCYn05zvYIrwLgtK8Ap8QcjRg+0i76tMQdWN6wOk=
 github.com/Azure/azure-sdk-for-go/sdk/azidentity v1.13.1 h1:Hk5QBxZQC1jb2Fwj6mpzme37xbCDdNTxU7O9eb5+LB4=
@@ -21,6 +23,8 @@ github.com/gdamore/encoding v1.0.1 h1:YzKZckdBL6jVt2Gc+5p82qhrGiqMdG/eNs6Wy0u3Uh
 github.com/gdamore/encoding v1.0.1/go.mod h1:0Z0cMFinngz9kS1QfMjCP8TY7em3bZYeeklsSDPivEo=
 github.com/gdamore/tcell/v2 v2.13.9 h1:uI5l3DYPcFvHINKlGft+en23evOKL+dwtD21QR8ejVA=
 github.com/gdamore/tcell/v2 v2.13.9/go.mod h1:+Wfe208WDdB7INEtCsNrAN6O2m+wsTPk1RAovjaILlo=
+github.com/go-sql-driver/mysql v1.9.3 h1:U/N249h2WzJ3Ukj8SowVFjdtZKfu9vlLZxjPXV1aweo=
+github.com/go-sql-driver/mysql v1.9.3/go.mod h1:qn46aNg1333BRMNU69Lq93t8du/dwxI64Gl8i5p1WMU=
 github.com/golang-jwt/jwt/v5 v5.3.1 h1:kYf81DTWFe7t+1VvL7eS+jKFVWaUnK9cB1qbwn63YCY=
 github.com/golang-jwt/jwt/v5 v5.3.1/go.mod h1:fxCRLWMO43lRc8nhHWY6LGqRcf+1gQWArsqaEUEa5bE=
 github.com/golang-sql/civil v0.0.0-20220223132316-b832511892a9 h1:au07oEsX2xN0ktxqI+Sida1w446QrXBRJ0nee3SNZlA=

linux/arch/PKGBUILD

@@ -1,6 +1,6 @@
 # Maintainer: Hein (Warky Devs) <hein@warky.dev>
 pkgname=relspec
-pkgver=1.0.74
+pkgver=1.0.86
 pkgrel=1
 pkgdesc="RelSpec is a comprehensive database relations management tool that reads, transforms, and writes database table specifications across multiple formats and ORMs."
 arch=('x86_64' 'aarch64')
@@ -15,7 +15,7 @@ build() {
     export CGO_ENABLED=0
     go build \
         -trimpath \
-        -ldflags "-X git.warky.dev/wdevs/relspecgo/cmd/relspec.version=$pkgver" \
+        -ldflags "-X 'git.warky.dev/wdevs/relspecgo/pkg/buildinfo.Version=v$pkgver' -X 'git.warky.dev/wdevs/relspecgo/pkg/buildinfo.BuildDate=$(date -u +"%Y-%m-%d %H:%M:%S UTC")'" \
         -o "$pkgname" ./cmd/relspec
 }
 

linux/centos/relspec.spec

@@ -1,5 +1,5 @@
 Name:           relspec
-Version:        1.0.74
+Version:        1.0.86
 Release:        1%{?dist}
 Summary:        RelSpec is a comprehensive database relations management tool that reads, transforms, and writes database table specifications across multiple formats and ORMs.
 
@@ -25,7 +25,7 @@ DBML, GraphQL, and more.
 export CGO_ENABLED=0
 go build \
     -trimpath \
-    -ldflags "-X git.warky.dev/wdevs/relspecgo/cmd/relspec.version=%{version}" \
+    -ldflags "-X 'git.warky.dev/wdevs/relspecgo/pkg/buildinfo.Version=v%{version}' -X 'git.warky.dev/wdevs/relspecgo/pkg/buildinfo.BuildDate=$(date -u +"%%Y-%%m-%%d %%H:%%M:%%S UTC")'" \
     -o %{name} ./cmd/relspec
 
 %install

pkg/buildinfo/buildinfo.go

@@ -0,0 +1,65 @@
+// Package buildinfo exposes the RelSpec version and build date so that both the
+// CLI and the schema writers can stamp generated output with the same values.
+package buildinfo
+
+import (
+	"fmt"
+	"runtime/debug"
+	"time"
+)
+
+// Version and BuildDate are set via -ldflags at build time (see Makefile). When
+// built without ldflags they are backfilled from the Go module build info.
+var (
+	Version   = "dev"
+	BuildDate = "unknown"
+)
+
+func init() {
+	if Version != "dev" {
+		return
+	}
+	info, ok := debug.ReadBuildInfo()
+	if !ok {
+		return
+	}
+	var rev, vcsTime string
+	for _, s := range info.Settings {
+		switch s.Key {
+		case "vcs.revision":
+			if len(s.Value) >= 7 {
+				rev = s.Value[:7]
+			}
+		case "vcs.time":
+			vcsTime = s.Value
+		}
+	}
+	if rev != "" {
+		Version = rev
+	}
+	if t, err := time.Parse(time.RFC3339, vcsTime); err == nil {
+		BuildDate = t.UTC().Format("2006-01-02 15:04:05 UTC")
+	}
+}
+
+// GeneratedComment returns the one-line provenance string embedded in generated
+// files, e.g. "RelSpec dev (built: unknown)".
+func GeneratedComment() string {
+	return fmt.Sprintf("RelSpec %s (built: %s)", Version, BuildDate)
+}
+
+const AsciiLogo = `
+██████╗ ███████╗██╗     ███████╗██████╗ ███████╗ ██████╗
+██╔══██╗██╔════╝██║     ██╔════╝██╔══██╗██╔════╝██╔════╝
+██████╔╝█████╗  ██║     ███████╗██████╔╝█████╗  ██║
+██╔══██╗██╔══╝  ██║     ╚════██║██╔═══╝ ██╔══╝  ██║
+██║  ██║███████╗███████╗███████║██║     ███████╗╚██████╗
+╚═╝  ╚═╝╚══════╝╚══════╝╚══════╝╚═╝     ╚══════╝ ╚═════╝
+[ IN ] ──▶ [ RELSPEC ] ──▶ [ OUT ]
+╔══════════════════════════════════════╗
+║                                      ║
+║   © WARKY DEVS                       ║
+║   Author: Hein (hein@warky.dev)      ║
+║                                      ║
+╚══════════════════════════════════════╝
+`

pkg/diff/diff.go

@@ -71,6 +71,13 @@ func compareSchemas(source, target []*models.Schema) *SchemaDiff {
 	return diff
 }
 
+func (c *SchemaChange) addChange(field string, source, target any) {
+	if c.Changes == nil {
+		c.Changes = make(map[string]any)
+	}
+	c.Changes[field] = map[string]any{"source": source, "target": target}
+}
+
 func compareSchemaDetails(source, target *models.Schema) *SchemaChange {
 	change := &SchemaChange{
 		Name: source.Name,
@@ -78,6 +85,16 @@ func compareSchemaDetails(source, target *models.Schema) *SchemaChange {
 
 	hasChanges := false
 
+	// Compare schema attributes
+	if source.Description != target.Description {
+		change.addChange("description", source.Description, target.Description)
+		hasChanges = true
+	}
+	if source.Owner != target.Owner {
+		change.addChange("owner", source.Owner, target.Owner)
+		hasChanges = true
+	}
+
 	// Compare tables
 	tableDiff := compareTables(source.Tables, target.Tables)
 	if !isEmpty(tableDiff) {

pkg/diff/diff_objects_test.go

@@ -0,0 +1,345 @@
+package diff
+
+import (
+	"reflect"
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+)
+
+func TestCompareSchemaDetails(t *testing.T) {
+	mk := func() *models.Schema {
+		s := models.InitSchema("public")
+		s.Tables = []*models.Table{models.InitTable("t", "public")}
+		return s
+	}
+
+	if got := compareSchemaDetails(mk(), mk()); got != nil {
+		t.Errorf("identical schemas must yield nil, got %+v", got)
+	}
+
+	tests := []struct {
+		name   string
+		mutate func(*models.Schema)
+		check  func(*SchemaChange) bool
+	}{
+		{
+			"table added", func(s *models.Schema) { s.Tables = append(s.Tables, models.InitTable("u", "public")) },
+			func(c *SchemaChange) bool { return c.Tables != nil && len(c.Tables.Extra) == 1 },
+		},
+		{
+			"view added", func(s *models.Schema) { s.Views = []*models.View{models.InitView("v", "public")} },
+			func(c *SchemaChange) bool { return c.Views != nil && len(c.Views.Extra) == 1 },
+		},
+		{
+			"sequence added", func(s *models.Schema) { s.Sequences = []*models.Sequence{models.InitSequence("sq", "public")} },
+			func(c *SchemaChange) bool { return c.Sequences != nil && len(c.Sequences.Extra) == 1 },
+		},
+		{
+			"script added", func(s *models.Schema) { s.Scripts = []*models.Script{models.InitScript("sc")} },
+			func(c *SchemaChange) bool { return c.Scripts != nil && len(c.Scripts.Extra) == 1 },
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			target := mk()
+			tt.mutate(target)
+			got := compareSchemaDetails(mk(), target)
+			if got == nil || got.Name != "public" || !tt.check(got) {
+				t.Errorf("unexpected change: %+v", got)
+			}
+		})
+	}
+}
+
+func TestCompareConstraintDetails(t *testing.T) {
+	base := func() *models.Constraint {
+		c := models.InitConstraint("fk", models.ForeignKeyConstraint)
+		c.Columns = []string{"a"}
+		c.ReferencedTable = "users"
+		c.ReferencedColumns = []string{"id"}
+		c.OnDelete = "CASCADE"
+		c.OnUpdate = "NO ACTION"
+		return c
+	}
+	if got := compareConstraintDetails(base(), base()); len(got) != 0 {
+		t.Errorf("identical: %v", got)
+	}
+
+	tests := []struct {
+		name    string
+		mutate  func(*models.Constraint)
+		wantKey string
+	}{
+		{"type", func(c *models.Constraint) { c.Type = models.UniqueConstraint }, "type"},
+		{"columns", func(c *models.Constraint) { c.Columns = []string{"b"} }, "columns"},
+		{"referenced table", func(c *models.Constraint) { c.ReferencedTable = "other" }, "referenced_table"},
+		{"referenced columns", func(c *models.Constraint) { c.ReferencedColumns = []string{"x"} }, "referenced_columns"},
+		{"on delete", func(c *models.Constraint) { c.OnDelete = "SET NULL" }, "on_delete"},
+		{"on update", func(c *models.Constraint) { c.OnUpdate = "CASCADE" }, "on_update"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			target := base()
+			tt.mutate(target)
+			got := compareConstraintDetails(base(), target)
+			if _, ok := got[tt.wantKey]; !ok || len(got) != 1 {
+				t.Errorf("got %v, want only %q", got, tt.wantKey)
+			}
+		})
+	}
+
+	// Action spelling variants that mean the same thing are not changes.
+	a, b := base(), base()
+	a.OnDelete, b.OnDelete = "cascade", " CASCADE "
+	a.OnUpdate, b.OnUpdate = "", "no action"
+	if got := compareConstraintDetails(a, b); len(got) != 0 {
+		t.Errorf("equivalent actions reported as changes: %v", got)
+	}
+}
+
+func TestNormalizeConstraintAction(t *testing.T) {
+	tests := []struct{ in, want string }{
+		{"", ""},
+		{"NO ACTION", ""},
+		{"no action", ""},
+		{"  No Action ", ""},
+		{"cascade", "CASCADE"},
+		{" set null ", "SET NULL"},
+		{"RESTRICT", "RESTRICT"},
+	}
+	for _, tt := range tests {
+		if got := normalizeConstraintAction(tt.in); got != tt.want {
+			t.Errorf("normalizeConstraintAction(%q) = %q, want %q", tt.in, got, tt.want)
+		}
+	}
+}
+
+func TestConstraintCompareKey(t *testing.T) {
+	uq := &models.Constraint{Name: "UQ_Name", Type: models.UniqueConstraint}
+	if got := constraintCompareKey(uq); got != "uq_name" {
+		t.Errorf("non-FK key: %q", got)
+	}
+	fk := func(name string) *models.Constraint {
+		return &models.Constraint{
+			Name: name, Type: models.ForeignKeyConstraint, Schema: "Public", Table: "Orders",
+			Columns: []string{"user_id"}, ReferencedSchema: "Public", ReferencedTable: "Users", ReferencedColumns: []string{"id"},
+		}
+	}
+	if constraintCompareKey(fk("a")) != constraintCompareKey(fk("b")) {
+		t.Error("FK key must ignore the constraint name")
+	}
+	other := fk("a")
+	other.ReferencedColumns = []string{"uid"}
+	if constraintCompareKey(fk("a")) == constraintCompareKey(other) {
+		t.Error("FK key must include referenced columns")
+	}
+}
+
+func TestFilterPrimaryKeyConstraints(t *testing.T) {
+	in := map[string]*models.Constraint{
+		"pk": {Name: "pk", Type: models.PrimaryKeyConstraint},
+		"uq": {Name: "uq", Type: models.UniqueConstraint},
+		"fk": {Name: "fk", Type: models.ForeignKeyConstraint},
+	}
+	got := filterPrimaryKeyConstraints(in)
+	if len(got) != 2 || got["pk"] != nil || got["uq"] == nil || got["fk"] == nil {
+		t.Errorf("got %v", got)
+	}
+	if len(in) != 3 {
+		t.Error("input must not be modified")
+	}
+	if got := filterPrimaryKeyConstraints(nil); got == nil || len(got) != 0 {
+		t.Errorf("nil: %v", got)
+	}
+}
+
+func TestCompareRelationshipDetails(t *testing.T) {
+	base := func() *models.Relationship {
+		r := models.InitRelationship("r", models.RelationType("one_to_many"))
+		r.FromTable, r.ToTable = "orders", "users"
+		r.FromColumns, r.ToColumns = []string{"user_id"}, []string{"id"}
+		return r
+	}
+	if got := compareRelationshipDetails(base(), base()); len(got) != 0 {
+		t.Errorf("identical: %v", got)
+	}
+	tests := []struct {
+		name    string
+		mutate  func(*models.Relationship)
+		wantKey string
+	}{
+		{"type", func(r *models.Relationship) { r.Type = "one_to_one" }, "type"},
+		{"from table", func(r *models.Relationship) { r.FromTable = "x" }, "from_table"},
+		{"to table", func(r *models.Relationship) { r.ToTable = "x" }, "to_table"},
+		{"from columns", func(r *models.Relationship) { r.FromColumns = []string{"x"} }, "from_columns"},
+		{"to columns", func(r *models.Relationship) { r.ToColumns = []string{"x"} }, "to_columns"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			target := base()
+			tt.mutate(target)
+			got := compareRelationshipDetails(base(), target)
+			if _, ok := got[tt.wantKey]; !ok || len(got) != 1 {
+				t.Errorf("got %v", got)
+			}
+		})
+	}
+}
+
+func TestCompareRelationshipsModified(t *testing.T) {
+	src := map[string]*models.Relationship{
+		"same":    {Name: "same", Type: "one_to_many"},
+		"changed": {Name: "changed", Type: "one_to_many"},
+		"missing": {Name: "missing"},
+	}
+	tgt := map[string]*models.Relationship{
+		"same":    {Name: "same", Type: "one_to_many"},
+		"changed": {Name: "changed", Type: "many_to_many"},
+		"extra":   {Name: "extra"},
+	}
+	d := compareRelationships(src, tgt)
+	if len(d.Missing) != 1 || d.Missing[0].Name != "missing" || len(d.Extra) != 1 || d.Extra[0].Name != "extra" ||
+		len(d.Modified) != 1 || d.Modified[0].Name != "changed" {
+		t.Errorf("got %+v", d)
+	}
+	if _, ok := d.Modified[0].Changes["type"]; !ok {
+		t.Errorf("changes: %v", d.Modified[0].Changes)
+	}
+}
+
+func TestCompareViews(t *testing.T) {
+	v := func(name, def string) *models.View { return &models.View{Name: name, Definition: def} }
+	src := []*models.View{v("Keep", "select 1"), v("Changed", "select 1"), v("Gone", "select 1")}
+	tgt := []*models.View{v("keep", "select 1"), v("changed", "select 2"), v("New", "select 1")}
+
+	d := compareViews(src, tgt)
+	if len(d.Missing) != 1 || d.Missing[0].Name != "Gone" {
+		t.Errorf("missing: %+v", d.Missing)
+	}
+	if len(d.Extra) != 1 || d.Extra[0].Name != "New" {
+		t.Errorf("extra: %+v", d.Extra)
+	}
+	if len(d.Modified) != 1 || d.Modified[0].Name != "changed" || d.Modified[0].Source.Definition != "select 1" || d.Modified[0].Target.Definition != "select 2" {
+		t.Errorf("modified: %+v", d.Modified)
+	}
+	want := map[string]any{"definition": map[string]string{"source": "select 1", "target": "select 2"}}
+	if !reflect.DeepEqual(d.Modified[0].Changes, want) {
+		t.Errorf("changes: %v", d.Modified[0].Changes)
+	}
+	if !isEmpty(compareViews(nil, nil)) {
+		t.Error("nil views must be empty")
+	}
+	if got := compareViewDetails(v("a", "x"), v("a", "x")); len(got) != 0 {
+		t.Errorf("same definition: %v", got)
+	}
+}
+
+func TestCompareSequences(t *testing.T) {
+	seq := func(name string, start, inc, min, max int64, cycle bool) *models.Sequence {
+		return &models.Sequence{Name: name, StartValue: start, IncrementBy: inc, MinValue: min, MaxValue: max, Cycle: cycle}
+	}
+	src := []*models.Sequence{seq("Same", 1, 1, 1, 100, false), seq("Diff", 1, 1, 1, 100, false), seq("Gone", 1, 1, 1, 1, false)}
+	tgt := []*models.Sequence{seq("same", 1, 1, 1, 100, false), seq("diff", 5, 2, 3, 200, true), seq("New", 1, 1, 1, 1, false)}
+
+	d := compareSequences(src, tgt)
+	if len(d.Missing) != 1 || d.Missing[0].Name != "Gone" || len(d.Extra) != 1 || d.Extra[0].Name != "New" || len(d.Modified) != 1 {
+		t.Fatalf("got %+v", d)
+	}
+	ch := d.Modified[0].Changes
+	for _, key := range []string{"start_value", "increment_by", "min_value", "max_value", "cycle"} {
+		if _, ok := ch[key]; !ok {
+			t.Errorf("missing change key %q in %v", key, ch)
+		}
+	}
+	if got := ch["increment_by"].(map[string]int64); got["source"] != 1 || got["target"] != 2 {
+		t.Errorf("increment_by: %v", got)
+	}
+	if got := ch["cycle"].(map[string]bool); got["source"] || !got["target"] {
+		t.Errorf("cycle: %v", got)
+	}
+	if got := compareSequenceDetails(seq("a", 1, 1, 1, 1, false), seq("a", 1, 1, 1, 1, false)); len(got) != 0 {
+		t.Errorf("identical: %v", got)
+	}
+}
+
+func TestCompareScriptDetailsAllFields(t *testing.T) {
+	a := &models.Script{Name: "s", SQL: "a", Rollback: "ra", RunAfter: []string{"x"}, Schema: "p", Version: "1", Priority: 1, Sequence: 1}
+	b := &models.Script{Name: "s", SQL: "b", Rollback: "rb", RunAfter: []string{"y"}, Schema: "q", Version: "2", Priority: 2, Sequence: 2}
+	got := compareScriptDetails(a, b)
+	for _, key := range []string{"sql", "rollback", "run_after", "schema", "version", "priority", "sequence"} {
+		if _, ok := got[key]; !ok {
+			t.Errorf("missing %q in %v", key, got)
+		}
+	}
+	if got := compareScriptDetails(a, a); len(got) != 0 {
+		t.Errorf("identical: %v", got)
+	}
+}
+
+func TestIsEmptyAllTypes(t *testing.T) {
+	if !isEmpty(&ViewDiff{}) || !isEmpty(&SequenceDiff{}) {
+		t.Error("empty view/sequence diffs must be empty")
+	}
+	if isEmpty(&ViewDiff{Extra: []*models.View{{Name: "v"}}}) || isEmpty(&SequenceDiff{Modified: []*SequenceChange{{Name: "s"}}}) {
+		t.Error("non-empty diffs reported as empty")
+	}
+	if isEmpty(&ConstraintDiff{Modified: []*ConstraintChange{{Name: "c"}}}) || isEmpty(&RelationshipDiff{Missing: []*models.Relationship{{Name: "r"}}}) {
+		t.Error("non-empty diffs reported as empty")
+	}
+	if isEmpty(&IndexDiff{Modified: []*IndexChange{{Name: "i"}}}) || isEmpty(&TableDiff{Modified: []*TableChange{{Name: "t"}}}) {
+		t.Error("non-empty diffs reported as empty")
+	}
+	if isEmpty("something else") || isEmpty(nil) {
+		t.Error("unknown types must not be treated as empty")
+	}
+}
+
+func TestComputeSummaryFullTree(t *testing.T) {
+	res := &DiffResult{Schemas: &SchemaDiff{
+		Missing: []*models.Schema{{Name: "m"}},
+		Extra:   []*models.Schema{{Name: "e"}},
+		Modified: []*SchemaChange{{
+			Name: "public",
+			Tables: &TableDiff{
+				Missing: []*models.Table{{Name: "a"}},
+				Extra:   []*models.Table{{Name: "b"}, {Name: "c"}},
+				Modified: []*TableChange{{
+					Name:          "t",
+					Columns:       &ColumnDiff{Missing: []*models.Column{{}}, Extra: []*models.Column{{}, {}}, Modified: []*ColumnChange{{}}},
+					Indexes:       &IndexDiff{Missing: []*models.Index{{}}, Extra: []*models.Index{{}}, Modified: []*IndexChange{{}, {}}},
+					Constraints:   &ConstraintDiff{Missing: []*models.Constraint{{}}, Modified: []*ConstraintChange{{}}},
+					Relationships: &RelationshipDiff{Extra: []*models.Relationship{{}}},
+				}},
+			},
+			Views:     &ViewDiff{Missing: []*models.View{{}}, Extra: []*models.View{{}}, Modified: []*ViewChange{{}}},
+			Sequences: &SequenceDiff{Missing: []*models.Sequence{{}}, Extra: []*models.Sequence{{}, {}}},
+			Scripts:   &ScriptDiff{Modified: []*ScriptChange{{}}},
+		}},
+	}}
+	s := ComputeSummary(res)
+	checks := []struct {
+		name string
+		got  [3]int
+		want [3]int
+	}{
+		{"schemas", [3]int{s.Schemas.Missing, s.Schemas.Extra, s.Schemas.Modified}, [3]int{1, 1, 1}},
+		{"tables", [3]int{s.Tables.Missing, s.Tables.Extra, s.Tables.Modified}, [3]int{1, 2, 1}},
+		{"columns", [3]int{s.Columns.Missing, s.Columns.Extra, s.Columns.Modified}, [3]int{1, 2, 1}},
+		{"indexes", [3]int{s.Indexes.Missing, s.Indexes.Extra, s.Indexes.Modified}, [3]int{1, 1, 2}},
+		{"constraints", [3]int{s.Constraints.Missing, s.Constraints.Extra, s.Constraints.Modified}, [3]int{1, 0, 1}},
+		{"relationships", [3]int{s.Relationships.Missing, s.Relationships.Extra, s.Relationships.Modified}, [3]int{0, 1, 0}},
+		{"views", [3]int{s.Views.Missing, s.Views.Extra, s.Views.Modified}, [3]int{1, 1, 1}},
+		{"sequences", [3]int{s.Sequences.Missing, s.Sequences.Extra, s.Sequences.Modified}, [3]int{1, 2, 0}},
+		{"scripts", [3]int{s.Scripts.Missing, s.Scripts.Extra, s.Scripts.Modified}, [3]int{0, 0, 1}},
+	}
+	for _, c := range checks {
+		if c.got != c.want {
+			t.Errorf("%s: got %v, want %v", c.name, c.got, c.want)
+		}
+	}
+
+	if got := ComputeSummary(&DiffResult{}); got == nil || got.Schemas != (SchemaSummary{}) {
+		t.Errorf("nil Schemas: %+v", got)
+	}
+}

pkg/diff/schema_attrs_test.go

@@ -0,0 +1,63 @@
+package diff
+
+import (
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+)
+
+func TestCompareSchemaDetails_DescriptionAndOwner(t *testing.T) {
+	mk := func(desc, owner string) *models.Schema {
+		s := models.InitSchema("public")
+		s.Description, s.Owner = desc, owner
+		return s
+	}
+	tests := []struct {
+		name       string
+		src, tgt   *models.Schema
+		wantFields []string
+	}{
+		{"identical", mk("d", "o"), mk("d", "o"), nil},
+		{"description", mk("a", "o"), mk("b", "o"), []string{"description"}},
+		{"owner", mk("d", "x"), mk("d", "y"), []string{"owner"}},
+		{"both", mk("a", "x"), mk("b", "y"), []string{"description", "owner"}},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := compareSchemaDetails(tt.src, tt.tgt)
+			if len(tt.wantFields) == 0 {
+				if got != nil {
+					t.Fatalf("expected no change, got %+v", got)
+				}
+				return
+			}
+			if got == nil || len(got.Changes) != len(tt.wantFields) {
+				t.Fatalf("changes: %+v", got)
+			}
+			for _, f := range tt.wantFields {
+				c, ok := got.Changes[f].(map[string]any)
+				if !ok {
+					t.Fatalf("missing %s: %+v", f, got.Changes)
+				}
+				if c["source"] == c["target"] {
+					t.Errorf("%s source and target equal: %v", f, c)
+				}
+			}
+		})
+	}
+}
+
+func TestCompareDatabases_SchemaAttrsCounted(t *testing.T) {
+	src, tgt := models.InitDatabase("a"), models.InitDatabase("b")
+	s1, s2 := models.InitSchema("public"), models.InitSchema("public")
+	s1.Owner, s2.Owner = "alice", "bob"
+	src.Schemas, tgt.Schemas = append(src.Schemas, s1), append(tgt.Schemas, s2)
+
+	res := CompareDatabases(src, tgt)
+	if res.Schemas == nil || len(res.Schemas.Modified) != 1 {
+		t.Fatalf("schema owner change not reported: %+v", res.Schemas)
+	}
+	if ComputeSummary(res).Schemas.Modified != 1 {
+		t.Error("summary must count the modified schema")
+	}
+}

pkg/diff/types.go

@@ -18,11 +18,12 @@ type SchemaDiff struct {
 
 // SchemaChange represents changes within a schema
 type SchemaChange struct {
-	Name      string        `json:"name"`
-	Tables    *TableDiff    `json:"tables,omitempty"`
-	Views     *ViewDiff     `json:"views,omitempty"`
-	Sequences *SequenceDiff `json:"sequences,omitempty"`
-	Scripts   *ScriptDiff   `json:"scripts,omitempty"`
+	Name      string         `json:"name"`
+	Changes   map[string]any `json:"changes,omitempty"` // Schema attributes that differ (description, owner), keyed by field name
+	Tables    *TableDiff     `json:"tables,omitempty"`
+	Views     *ViewDiff      `json:"views,omitempty"`
+	Sequences *SequenceDiff  `json:"sequences,omitempty"`
+	Scripts   *ScriptDiff    `json:"scripts,omitempty"`
 }
 
 // TableDiff represents differences in tables

pkg/inspector/README.md

@@ -123,6 +123,7 @@ rules:
 | `missing_primary_key` | `have_primary_key` | Ensure tables have primary keys |
 | `orphaned_foreign_key` | `orphaned_foreign_key` | Detect FKs referencing non-existent tables |
 | `circular_dependency` | `circular_dependency` | Detect circular FK dependencies |
+| `duplicate_index_name` | `duplicate_index_name` | Index / PK / unique names must be unique per schema (default: `enforce`) |
 
 ## Rule Configuration
 

pkg/inspector/inspector.go

@@ -161,6 +161,7 @@ func getValidator(functionName string) (validatorFunc, bool) {
 		"have_primary_key":              validateMissingPrimaryKey,
 		"orphaned_foreign_key":          validateOrphanedForeignKey,
 		"circular_dependency":           validateCircularDependency,
+		"duplicate_index_name":          validateDuplicateIndexName,
 	}
 
 	fn, exists := validators[functionName]

pkg/inspector/rules.go

@@ -154,6 +154,11 @@ func GetDefaultConfig() *Config {
 				Function: "circular_dependency",
 				Message:  "Circular foreign key dependency detected",
 			},
+			"duplicate_index_name": {
+				Enabled:  "enforce",
+				Function: "duplicate_index_name",
+				Message:  "Index name is reused within the schema; PostgreSQL skips the duplicate CREATE INDEX IF NOT EXISTS",
+			},
 		},
 	}
 }

pkg/inspector/rules_test.go

@@ -37,6 +37,7 @@ func TestGetDefaultConfig(t *testing.T) {
 		"missing_primary_key",
 		"orphaned_foreign_key",
 		"circular_dependency",
+		"duplicate_index_name",
 	}
 
 	for _, ruleName := range expectedRules {

pkg/inspector/validators.go

@@ -643,3 +643,64 @@ func contains(slice []string, value string) bool {
 	}
 	return false
 }
+
+// validateDuplicateIndexName checks that index names are unique per schema.
+// PostgreSQL keeps indexes in the schema-wide relation namespace, so a name
+// reused on another table makes CREATE INDEX IF NOT EXISTS silently skip it.
+// Primary key and unique constraints create backing indexes and share that
+// namespace too. An index and a constraint with the same name on the same
+// table describe one object and are not reported.
+func validateDuplicateIndexName(db *models.Database, rule Rule, ruleName string) []ValidationResult {
+	results := []ValidationResult{}
+
+	for _, schema := range db.Schemas {
+		// lowercased name -> "table" entries, one per distinct object
+		owners := make(map[string][]string)
+		display := make(map[string]string)
+		order := []string{}
+
+		add := func(name, table string, sameTableMerges bool) {
+			key := strings.ToLower(name)
+			if _, seen := owners[key]; !seen {
+				order = append(order, key)
+				display[key] = name
+			}
+			if sameTableMerges && contains(owners[key], table) {
+				return
+			}
+			owners[key] = append(owners[key], table)
+		}
+
+		for _, table := range schema.Tables {
+			for _, key := range sortedKeys(table.Indexes) {
+				if name := table.Indexes[key].Name; name != "" {
+					add(name, table.Name, false)
+				}
+			}
+			for _, c := range sortConstraints(table.Constraints) {
+				if c.Name == "" || (c.Type != models.PrimaryKeyConstraint && c.Type != models.UniqueConstraint) {
+					continue
+				}
+				add(c.Name, table.Name, true)
+			}
+		}
+
+		for _, key := range order {
+			tables := owners[key]
+			results = append(results, createResult(
+				ruleName,
+				len(tables) == 1,
+				rule.Message,
+				formatLocation(schema.Name, display[key], "")+" on "+strings.Join(tables, ", "),
+				map[string]interface{}{
+					"schema":      schema.Name,
+					"index":       display[key],
+					"tables":      tables,
+					"occurrences": len(tables),
+				},
+			))
+		}
+	}
+
+	return results
+}

pkg/inspector/validators_test.go

@@ -835,3 +835,84 @@ func TestFormatLocation(t *testing.T) {
 		})
 	}
 }
+
+func TestValidateDuplicateIndexName(t *testing.T) {
+	db := &models.Database{
+		Name: "testdb",
+		Schemas: []*models.Schema{
+			{
+				Name: "entity",
+				Tables: []*models.Table{
+					{
+						Name: "actor_phone",
+						Indexes: map[string]*models.Index{
+							"idx_actor": {Name: "idx_actor", Columns: []string{"rid_actor"}},
+							"uk_phone":  {Name: "uk_phone", Columns: []string{"phone"}, Unique: true},
+						},
+						Constraints: map[string]*models.Constraint{
+							// Same name as the index on the same table: one object.
+							"uk_phone": {Name: "uk_phone", Type: models.UniqueConstraint, Columns: []string{"phone"}},
+						},
+					},
+					{
+						Name: "actor_email",
+						Indexes: map[string]*models.Index{
+							"idx_actor": {Name: "idx_actor", Columns: []string{"rid_actor"}},
+						},
+						Constraints: map[string]*models.Constraint{
+							"UK_Phone": {Name: "UK_Phone", Type: models.UniqueConstraint, Columns: []string{"email"}},
+						},
+					},
+					{
+						Name: "actor_address",
+						Indexes: map[string]*models.Index{
+							"idx_actor":   {Name: "idx_actor", Columns: []string{"rid_actor"}},
+							"idx_actor#2": {Name: "idx_actor", Columns: []string{"rid_actor", "kind"}},
+							"idx_address": {Name: "idx_address", Columns: []string{"line1"}},
+						},
+					},
+				},
+			},
+			{
+				// Same names in another schema do not collide.
+				Name: "org",
+				Tables: []*models.Table{
+					{Name: "api_provider", Indexes: map[string]*models.Index{"idx_actor": {Name: "idx_actor"}}},
+				},
+			},
+		},
+	}
+
+	results := validateDuplicateIndexName(db, Rule{Message: "dup"}, "duplicate_index_name")
+
+	got := map[string]bool{}
+	occ := map[string]int{}
+	for _, r := range results {
+		key := r.Context["schema"].(string) + "." + r.Context["index"].(string)
+		got[key] = r.Passed
+		occ[key] = r.Context["occurrences"].(int)
+	}
+
+	want := map[string]struct {
+		passed bool
+		occ    int
+	}{
+		"entity.idx_actor":   {false, 4},
+		"entity.uk_phone":    {false, 2},
+		"entity.idx_address": {true, 1},
+		"org.idx_actor":      {true, 1},
+	}
+	if len(got) != len(want) {
+		t.Fatalf("got %d results %v, want %d", len(got), got, len(want))
+	}
+	for k, w := range want {
+		p, ok := got[k]
+		if !ok {
+			t.Errorf("missing result for %s", k)
+			continue
+		}
+		if p != w.passed || occ[k] != w.occ {
+			t.Errorf("%s: passed=%v occurrences=%d, want passed=%v occurrences=%d", k, p, occ[k], w.passed, w.occ)
+		}
+	}
+}

pkg/jobs/jobs.go

@@ -19,6 +19,7 @@ import (
 	"fmt"
 	"os"
 	"path/filepath"
+	"regexp"
 	"sort"
 	"strconv"
 	"strings"
@@ -276,15 +277,22 @@ func parseHumanSize(s string) (int64, error) {
 
 // Options carries the subset of command flags a job file may set.
 type Options struct {
-	FlattenSchema   bool   `yaml:"flatten_schema"`
-	Schema          string `yaml:"schema"`
-	Package         string `yaml:"package"`
+	FlattenSchema bool   `yaml:"flatten_schema"`
+	Schema        string `yaml:"schema"`
+	Package       string `yaml:"package"`
+	// Types and ArrayNullable mirror the existing convert/split writer flags.
+	// They are passed through only to writers that already support them.
+	Types           string `yaml:"types"`
+	ArrayNullable   string `yaml:"array_nullable"`
 	ContinueOnError bool   `yaml:"continue_on_error"`
-	SkipRelations   bool   `yaml:"skip_relations"`
-	SkipEnums       bool   `yaml:"skip_enums"`
-	SkipViews       bool   `yaml:"skip_views"`
-	SkipDomains     bool   `yaml:"skip_domains"`
-	SkipSequences   bool   `yaml:"skip_sequences"`
+	// FullDDL makes direct pgsql output execute the full idempotent DDL instead of
+	// diffing against the live database first.
+	FullDDL       bool `yaml:"full_ddl"`
+	SkipRelations bool `yaml:"skip_relations"`
+	SkipEnums     bool `yaml:"skip_enums"`
+	SkipViews     bool `yaml:"skip_views"`
+	SkipDomains   bool `yaml:"skip_domains"`
+	SkipSequences bool `yaml:"skip_sequences"`
 }
 
 // Dir returns the directory that a job's relative paths resolve against:
@@ -568,7 +576,9 @@ func (j *Job) validate() []string {
 			e = append(e, "command \"templ\" does not support database output")
 		}
 		if j.Output != nil && j.Output.Format != "" {
-			e = append(e, "output.format is not valid for command \"templ\"")
+			if !strings.EqualFold(j.Output.Format, "text") {
+				e = append(e, "command \"templ\" accepts only output.format: text")
+			}
 		}
 	case CommandSplit:
 		if len(j.Inputs) < 1 {
@@ -850,6 +860,29 @@ func SafeJoin(root, rel string) (string, error) {
 	return joined, nil
 }
 
+var envReferencePattern = regexp.MustCompile(`\$\{([A-Za-z_][A-Za-z0-9_]*)\}`)
+
+// ExpandEnv expands ${NAME} references using the current process environment.
+// It deliberately supports only shell-independent variable references; job
+// files are never passed through a shell. An unset variable is an error so a
+// typo cannot silently turn into a relative path.
+func ExpandEnv(value string) (string, error) {
+	var missing string
+	expanded := envReferencePattern.ReplaceAllStringFunc(value, func(reference string) string {
+		name := reference[2 : len(reference)-1]
+		resolved, ok := os.LookupEnv(name)
+		if !ok {
+			missing = name
+			return reference
+		}
+		return resolved
+	})
+	if missing != "" {
+		return "", fmt.Errorf("environment variable %q referenced by ${%s} is not set", missing, missing)
+	}
+	return expanded, nil
+}
+
 // deepestExistingAncestor returns p itself if it exists, otherwise the nearest
 // existing parent directory (falling back to the filesystem root).
 func deepestExistingAncestor(p string) string {

pkg/jobs/jobs_test.go

@@ -135,6 +135,21 @@ func TestParseHumanSize(t *testing.T) {
 	}
 }
 
+func TestExpandEnv(t *testing.T) {
+	t.Setenv("RELSPEC_TEST_ROOT", "build/output")
+	got, err := ExpandEnv("${RELSPEC_TEST_ROOT}/schema-${RELSPEC_TEST_ROOT}")
+	if err != nil {
+		t.Fatal(err)
+	}
+	if want := "build/output/schema-build/output"; got != want {
+		t.Fatalf("ExpandEnv = %q, want %q", got, want)
+	}
+
+	if _, err := ExpandEnv("build/${RELSPEC_TEST_MISSING}"); err == nil {
+		t.Fatal("expected missing environment variable error")
+	}
+}
+
 func TestLoadRejectsDuplicateJobAcrossFiles(t *testing.T) {
 	dir := t.TempDir()
 	a := filepath.Join(dir, "relspec.yml")

pkg/jobs/jobs_validate_test.go

@@ -0,0 +1,262 @@
+package jobs
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+)
+
+// validateYAML loads one job file and returns the Validate error text ("" when valid).
+func validateYAML(t *testing.T, body string) string {
+	t.Helper()
+	set := loadOne(t, "version: 1\njobs:\n"+body)
+	if err := set.Validate(); err != nil {
+		return err.Error()
+	}
+	return ""
+}
+
+func TestValidateJobTable(t *testing.T) {
+	in := "    inputs:\n      - path: a.dbml\n        format: dbml\n"
+	out := "    output:\n      format: json\n      path: out.json\n"
+	tests := []struct {
+		name string
+		job  string
+		want string // substring of the error, "" for valid
+	}{
+		{"missing command", "  x:\n    description: d\n", "missing command"},
+		{"convert valid", "  x:\n    command: convert\n" + in + out, ""},
+		{"convert script dirs", "  x:\n    command: convert\n    script_dirs: [s]\n" + in + out, "script_dirs is not valid"},
+		{"convert missing output", "  x:\n    command: convert\n" + in, "missing output"},
+		{"convert output missing format", "  x:\n    command: convert\n" + in + "    output:\n      path: o\n", "output: missing format"},
+		{"convert output unsupported format", "  x:\n    command: convert\n" + in + "    output:\n      format: nope\n      path: o\n", "unsupported output format"},
+		{"convert output missing path", "  x:\n    command: convert\n" + in + "    output:\n      format: json\n", "output: missing path"},
+		{"convert output conn_env on non-exec format", "  x:\n    command: convert\n" + in + "    output:\n      format: json\n      conn_env: DB\n", "not supported for format"},
+		{"convert output path and conn_env", "  x:\n    command: convert\n" + in + "    output:\n      format: pgsql\n      conn_env: DB\n      path: o.sql\n", "either path or conn_env"},
+		{"convert output conn_env ok", "  x:\n    command: convert\n" + in + "    output:\n      format: pgsql\n      conn_env: DB\n", ""},
+		{"output secret conn_env", "  x:\n    command: convert\n" + in + "    output:\n      format: pgsql\n      conn_env: postgres://u:p@h/db\n", "environment variable name"},
+		{"merge needs two inputs", "  x:\n    command: merge\n" + in + out, "at least 2 input"},
+		{"input missing format", "  x:\n    command: convert\n    inputs:\n      - path: a\n" + out, "missing format"},
+		{"input unsupported format", "  x:\n    command: convert\n    inputs:\n      - path: a\n        format: nope\n" + out, "unsupported input format"},
+		{"input file missing path", "  x:\n    command: convert\n    inputs:\n      - format: dbml\n" + out, "missing path"},
+		{"input file with conn_env", "  x:\n    command: convert\n    inputs:\n      - path: a\n        format: dbml\n        conn_env: DB\n" + out, "does not use conn_env"},
+		{"input db missing conn_env", "  x:\n    command: convert\n    inputs:\n      - format: pgsql\n" + out, "requires conn_env"},
+		{"input db with path", "  x:\n    command: convert\n    inputs:\n      - format: pgsql\n        conn_env: DB\n        path: a\n" + out, "takes conn_env, not path"},
+		{"input db ok", "  x:\n    command: convert\n    inputs:\n      - format: pgsql\n        conn_env: DB\n" + out, ""},
+		{"input secret conn_env", "  x:\n    command: convert\n    inputs:\n      - format: pgsql\n        conn_env: \"host=h password=p\"\n" + out, "environment variable name"},
+		{"bad log size", "  x:\n    command: convert\n    log_max_size: lots\n" + in + out, "log_max_size"},
+		{"absolute logfile", "  x:\n    command: convert\n    logfile: /var/log/x.log\n" + in + out, "absolute paths"},
+		{"home path", "  x:\n    command: convert\n    template: ~/t\n" + in + out, "home-relative"},
+		{"report path traversal", "  x:\n    command: inspect\n" + in + "    report:\n      format: json\n      path: ../r.json\n", "escapes"},
+		{"script_dir traversal", "  x:\n    command: scripts-list\n    script_dirs: [../x]\n", "escapes"},
+
+		{"templ valid", "  x:\n    command: templ\n" + in + "    template: t.tmpl\n    mode: table\n    output:\n      format: text\n      path: o\n", ""},
+		{"templ pgsql input valid", "  x:\n    command: templ\n    inputs:\n      - format: pgsql\n        conn_env: DB\n    template: t.tmpl\n", ""},
+		{"templ no inputs", "  x:\n    command: templ\n    template: t.tmpl\n", "at least 1 input"},
+		{"templ no template", "  x:\n    command: templ\n" + in, "requires template"},
+		{"templ bad mode", "  x:\n    command: templ\n" + in + "    template: t\n    mode: weird\n", "unsupported mode"},
+		{"templ script dirs", "  x:\n    command: templ\n" + in + "    template: t\n    script_dirs: [s]\n", "script_dirs is not valid"},
+		{"templ db output", "  x:\n    command: templ\n" + in + "    template: t\n    output:\n      conn_env: DB\n", "does not support database output"},
+		{"templ non-text output", "  x:\n    command: templ\n" + in + "    template: t\n    output:\n      format: json\n      path: o\n", "only output.format: text"},
+		{"templ input missing format", "  x:\n    command: templ\n    inputs:\n      - path: a\n    template: t\n", "missing format"},
+		{"templ pgsql input without conn_env", "  x:\n    command: templ\n    inputs:\n      - format: pgsql\n    template: t\n", "requires conn_env"},
+		{"templ pgsql input with path", "  x:\n    command: templ\n    inputs:\n      - format: pgsql\n        conn_env: DB\n        path: a\n    template: t\n", "takes conn_env, not path"},
+		{"templ file input without path", "  x:\n    command: templ\n    inputs:\n      - format: dbml\n    template: t\n", "missing path"},
+		{"templ file input with conn_env", "  x:\n    command: templ\n    inputs:\n      - path: a\n        format: dbml\n        conn_env: DB\n    template: t\n", "does not use conn_env"},
+		{"templ unsupported input format", "  x:\n    command: templ\n    inputs:\n      - path: a\n        format: nope\n    template: t\n", "unsupported templ input format"},
+		{"templ secret conn_env", "  x:\n    command: templ\n    inputs:\n      - format: pgsql\n        conn_env: a/b\n    template: t\n", "environment variable name"},
+
+		{"split needs input", "  x:\n    command: split\n" + out, "at least 1 input"},
+		{"split script dirs", "  x:\n    command: split\n" + in + "    script_dirs: [s]\n" + out, "script_dirs is not valid"},
+		{"split report", "  x:\n    command: split\n" + in + "    report:\n      format: json\n      path: r\n" + out, "report is not valid"},
+		{"split db output", "  x:\n    command: split\n" + in + "    output:\n      format: pgsql\n      conn_env: DB\n", "writes a file"},
+
+		{"inspect script dirs", "  x:\n    command: inspect\n" + in + "    script_dirs: [s]\n    report:\n      path: r\n", "script_dirs is not valid"},
+		{"inspect output", "  x:\n    command: inspect\n" + in + out + "    report:\n      path: r\n", "output is not valid"},
+		{"inspect bad report format", "  x:\n    command: inspect\n" + in + "    report:\n      format: html\n      path: r\n", "not supported"},
+		{"inspect report without path", "  x:\n    command: inspect\n" + in + "    report:\n      format: json\n", "requires report.path"},
+		{"inspect default format ok", "  x:\n    command: inspect\n" + in + "    report:\n      path: r.md\n", ""},
+		{"diff summary without path ok", "  x:\n    command: diff\n" + in + "      - path: b.dbml\n        format: dbml\n    report:\n      format: summary\n", ""},
+		{"diff json needs path", "  x:\n    command: diff\n" + in + "      - path: b.dbml\n        format: dbml\n    report:\n      format: json\n", "requires report.path"},
+		{"diff output", "  x:\n    command: diff\n" + in + "      - path: b.dbml\n        format: dbml\n" + out + "    report:\n      format: summary\n", "output is not valid"},
+		{"diff script dirs", "  x:\n    command: diff\n" + in + "      - path: b.dbml\n        format: dbml\n    script_dirs: [s]\n    report:\n      format: summary\n", "script_dirs is not valid"},
+		{"diff no report", "  x:\n    command: diff\n" + in + "      - path: b.dbml\n        format: dbml\n", "requires a report block"},
+
+		{"scripts-list inputs", "  x:\n    command: scripts-list\n    script_dirs: [s]\n" + in, "inputs is not valid"},
+		{"scripts-list output", "  x:\n    command: scripts-list\n    script_dirs: [s]\n" + out, "output is not valid"},
+		{"scripts-exec inputs", "  x:\n    command: scripts-exec\n    script_dirs: [s]\n" + in + "    output:\n      conn_env: DB\n", "inputs is not valid"},
+		{"scripts-exec report", "  x:\n    command: scripts-exec\n    script_dirs: [s]\n    report:\n      path: r\n    output:\n      conn_env: DB\n", "report is not valid"},
+		{"scripts-exec output path", "  x:\n    command: scripts-exec\n    script_dirs: [s]\n    output:\n      conn_env: DB\n      path: p\n", "output.path is not supported"},
+		{"scripts-exec non-pgsql", "  x:\n    command: scripts-exec\n    script_dirs: [s]\n    output:\n      conn_env: DB\n      format: mssql\n", "only supports pgsql"},
+		{"scripts-exec secret conn_env", "  x:\n    command: scripts-exec\n    script_dirs: [s]\n    output:\n      conn_env: \"postgres://u@h/d\"\n", "environment variable name"},
+		{"scripts-exec no script dirs", "  x:\n    command: scripts-exec\n    output:\n      conn_env: DB\n", "requires at least one script_dir"},
+		{"scripts-exec pgsql format ok", "  x:\n    command: scripts-exec\n    script_dirs: [s]\n    output:\n      conn_env: DB\n      format: pgsql\n", ""},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := validateYAML(t, tt.job)
+			if tt.want == "" {
+				if got != "" {
+					t.Errorf("expected valid, got: %s", got)
+				}
+				return
+			}
+			if !strings.Contains(got, tt.want) {
+				t.Errorf("error %q does not contain %q", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestFromJobInputShape(t *testing.T) {
+	producer := "  p:\n    command: convert\n    inputs:\n      - path: a.dbml\n        format: dbml\n    output:\n      format: json\n      path: out.json\n"
+	tests := []struct {
+		name  string
+		input string
+		want  string
+	}{
+		{"path", "      - from_job: p\n        path: x\n", "takes no path"},
+		{"format", "      - from_job: p\n        format: json\n", "drop format"},
+		{"conn_env", "      - from_job: p\n        conn_env: DB\n", "takes no conn_env"},
+	}
+	for _, tt := range tests {
+		for _, cmd := range []string{"convert", "templ"} {
+			t.Run(cmd+"/"+tt.name, func(t *testing.T) {
+				extra := "    output:\n      format: json\n      path: o.json\n"
+				if cmd == "templ" {
+					extra = "    template: t.tmpl\n"
+				}
+				got := validateYAML(t, producer+"  c:\n    command: "+cmd+"\n    inputs:\n"+tt.input+extra)
+				if !strings.Contains(got, tt.want) {
+					t.Errorf("error %q does not contain %q", got, tt.want)
+				}
+			})
+		}
+	}
+}
+
+func TestResolvedLogPolicy(t *testing.T) {
+	keep2 := 2
+	keep0 := 0
+	tests := []struct {
+		name string
+		job  Job
+		want LogPolicy
+	}{
+		{"built-in defaults", Job{}, LogPolicy{MaxSizeBytes: defaultLogMaxSizeBytes, Keep: defaultLogKeep}},
+		{"file defaults", Job{fileDefaults: &Defaults{LogMaxSize: "1MB", LogKeep: 7}}, LogPolicy{MaxSizeBytes: 1 << 20, Keep: 7}},
+		{"file defaults invalid size falls back", Job{fileDefaults: &Defaults{LogMaxSize: "junk", LogKeep: 0}}, LogPolicy{MaxSizeBytes: defaultLogMaxSizeBytes, Keep: defaultLogKeep}},
+		{"job overrides file", Job{fileDefaults: &Defaults{LogMaxSize: "1MB", LogKeep: 7}, LogMaxSize: "2kb", LogKeep: &keep2}, LogPolicy{MaxSizeBytes: 2 << 10, Keep: 2}},
+		{"job keep zero is honoured", Job{LogKeep: &keep0}, LogPolicy{MaxSizeBytes: defaultLogMaxSizeBytes, Keep: 0}},
+		{"job invalid size ignored", Job{LogMaxSize: "junk"}, LogPolicy{MaxSizeBytes: defaultLogMaxSizeBytes, Keep: defaultLogKeep}},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := tt.job.ResolvedLogPolicy(); got != tt.want {
+				t.Errorf("got %+v, want %+v", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestLoadAppliesFileDefaultsAndDir(t *testing.T) {
+	dir := t.TempDir()
+	p := filepath.Join(dir, "relspec.yml")
+	write(t, p, "version: 1\ndefaults:\n  log_max_size: 1MB\n  log_keep: 9\n"+"jobs:\n  a:\n    command: convert\n    inputs:\n      - path: a.dbml\n        format: dbml\n    output:\n      format: json\n      path: o.json\n")
+	set, err := Load([]string{p})
+	if err != nil {
+		t.Fatal(err)
+	}
+	job := set.Jobs["a"]
+	if job.Dir() != dir {
+		t.Errorf("Dir = %q, want %q", job.Dir(), dir)
+	}
+	if pol := job.ResolvedLogPolicy(); pol.MaxSizeBytes != 1<<20 || pol.Keep != 9 {
+		t.Errorf("policy %+v", pol)
+	}
+}
+
+func TestSetNamesSorted(t *testing.T) {
+	set := &Set{Jobs: map[string]*Job{"b": {}, "a": {}, "c": {}}}
+	if got := strings.Join(set.Names(), ","); got != "a,b,c" {
+		t.Errorf("got %s", got)
+	}
+}
+
+func TestPlanErrors(t *testing.T) {
+	set := loadOne(t, "version: 1\njobs:\n"+
+		"  a:\n    command: convert\n    depends_on: [ghost]\n    inputs:\n      - path: a.dbml\n        format: dbml\n    output:\n      format: json\n      path: o.json\n"+
+		"  b:\n    command: convert\n    inputs:\n      - path: a.dbml\n        format: dbml\n    output:\n      format: json\n      path: o2.json\n")
+
+	if _, err := set.Plan("nope", true); err == nil || !strings.Contains(err.Error(), "unknown job") || !strings.Contains(err.Error(), "a, b") {
+		t.Errorf("unknown job: %v", err)
+	}
+	if _, err := set.Plan("a", true); err == nil || !strings.Contains(err.Error(), "unknown job \"ghost\"") {
+		t.Errorf("unknown dependency: %v", err)
+	}
+	// Without dependencies the declared dependency is not walked.
+	if got, err := set.Plan("a", false); err != nil || len(got) != 1 || got[0].Name != "a" {
+		t.Errorf("no-deps plan: %v %v", got, err)
+	}
+}
+
+func TestPlanCycleAtRuntime(t *testing.T) {
+	set := &Set{Jobs: map[string]*Job{
+		"a": {Name: "a", DependsOn: []string{"b"}},
+		"b": {Name: "b", DependsOn: []string{"a"}},
+	}}
+	if _, err := set.Plan("a", true); err == nil || !strings.Contains(err.Error(), "cycle") {
+		t.Errorf("want cycle error, got %v", err)
+	}
+}
+
+func TestDiscoverErrorsAndFiltering(t *testing.T) {
+	if _, err := Discover(filepath.Join(t.TempDir(), "missing")); err == nil {
+		t.Error("missing dir must fail")
+	}
+	dir := t.TempDir()
+	for _, f := range []string{"relspec.yaml", "relspec.b.yml", "relspec.a.yaml", "relspec.txt", "other.yml", "relspec"} {
+		write(t, filepath.Join(dir, f), "")
+	}
+	if err := os.Mkdir(filepath.Join(dir, "relspec.dir.yml"), 0o755); err != nil {
+		t.Fatal(err)
+	}
+	got, err := Discover(dir)
+	if err != nil {
+		t.Fatal(err)
+	}
+	var names []string
+	for _, p := range got {
+		names = append(names, filepath.Base(p))
+	}
+	if strings.Join(names, ",") != "relspec.yaml,relspec.a.yaml,relspec.b.yml" {
+		t.Errorf("got %v", names)
+	}
+}
+
+func TestSafeJoinCases(t *testing.T) {
+	root := t.TempDir()
+	if got, err := SafeJoin(root, "sub/file.sql"); err != nil || !strings.HasSuffix(got, filepath.Join("sub", "file.sql")) {
+		t.Errorf("nested: %q %v", got, err)
+	}
+	for _, bad := range []string{"", "/etc/passwd", "~/x", "..", "../x", "a/../../x"} {
+		if _, err := SafeJoin(root, bad); err == nil {
+			t.Errorf("SafeJoin(%q) must fail", bad)
+		}
+	}
+	if _, err := SafeJoin(filepath.Join(root, "does", "not", "exist"), "x"); err == nil {
+		t.Error("unresolvable root must fail")
+	}
+}
+
+func TestLooksLikeSecret(t *testing.T) {
+	for in, want := range map[string]bool{
+		"": false, "DB_URL": false, "MY_DB": false,
+		"postgres://u:p@h/db": true, "host=h": true, "a b": true, "a/b": true, "u@h": true, "k:v": true,
+	} {
+		if got := looksLikeSecret(in); got != want {
+			t.Errorf("looksLikeSecret(%q) = %v, want %v", in, got, want)
+		}
+	}
+}

pkg/merge/merge.go

@@ -82,6 +82,15 @@ func (r *MergeResult) merge(target, source *models.Database, opts *MergeOptions)
 		} else {
 			// Schema doesn't exist, add it
 			newSchema := cloneSchema(srcSchema)
+			if len(opts.SkipTableNames) > 0 {
+				kept := newSchema.Tables[:0]
+				for _, t := range newSchema.Tables {
+					if !opts.SkipTableNames[strings.ToLower(t.SQLName())] {
+						kept = append(kept, t)
+					}
+				}
+				newSchema.Tables = kept
+			}
 			target.Schemas = append(target.Schemas, newSchema)
 			r.SchemasAdded++
 		}
@@ -91,6 +100,45 @@ func (r *MergeResult) merge(target, source *models.Database, opts *MergeOptions)
 	if !opts.SkipDomains {
 		r.mergeDomains(target, source)
 	}
+
+	mergeDatabaseMetadata(target, source)
+}
+
+// mergeDatabaseMetadata adds missing metadata keys and unions []string values,
+// so per-file reader state (e.g. pending DBML commented refs) survives a merge.
+func mergeDatabaseMetadata(target, source *models.Database) {
+	if len(source.Metadata) == 0 {
+		return
+	}
+	if target.Metadata == nil {
+		target.Metadata = make(map[string]any, len(source.Metadata))
+	}
+	for key, srcVal := range source.Metadata {
+		tgtVal, exists := target.Metadata[key]
+		if !exists {
+			if list, ok := srcVal.([]string); ok {
+				srcVal = append([]string(nil), list...)
+			}
+			target.Metadata[key] = srcVal
+			continue
+		}
+		tgtList, tgtOK := tgtVal.([]string)
+		srcList, srcOK := srcVal.([]string)
+		if !tgtOK || !srcOK {
+			continue
+		}
+		seen := make(map[string]bool, len(tgtList))
+		for _, v := range tgtList {
+			seen[v] = true
+		}
+		for _, v := range srcList {
+			if !seen[v] {
+				tgtList = append(tgtList, v)
+				seen[v] = true
+			}
+		}
+		target.Metadata[key] = tgtList
+	}
 }
 
 func (r *MergeResult) mergeSchemaContents(target, source *models.Schema, opts *MergeOptions) {
@@ -401,6 +449,8 @@ func cloneTable(table *models.Table) *models.Table {
 		Description: table.Description,
 		Schema:      table.Schema,
 		Comment:     table.Comment,
+		Tablespace:  table.Tablespace,
+		GUID:        table.GUID,
 		Sequence:    table.Sequence,
 		UpdatedAt:   table.UpdatedAt,
 		Columns:     make(map[string]*models.Column),
@@ -430,6 +480,14 @@ func cloneTable(table *models.Table) *models.Table {
 		newTable.Indexes[idxName] = cloneIndex(index)
 	}
 
+	// Clone relationships
+	if table.Relationships != nil {
+		newTable.Relationships = make(map[string]*models.Relationship, len(table.Relationships))
+		for relName, rel := range table.Relationships {
+			newTable.Relationships[relName] = cloneRelation(rel)
+		}
+	}
+
 	return newTable
 }
 

pkg/merge/merge_clone_test.go

@@ -0,0 +1,277 @@
+package merge
+
+import (
+	"strings"
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+)
+
+func TestMergeSequences(t *testing.T) {
+	target := models.InitSchema("public")
+	target.Sequences = []*models.Sequence{{Name: "Existing", StartValue: 1, IncrementBy: 1}}
+
+	source := models.InitSchema("public")
+	source.Sequences = []*models.Sequence{
+		{Name: "existing", StartValue: 100, IncrementBy: 10}, // conflicting: must not overwrite
+		{Name: "fresh", StartValue: 5, IncrementBy: 2, MinValue: 1, MaxValue: 99, CacheSize: 3, Cycle: true, OwnedByTable: "t", OwnedByColumn: "id", Comment: "c", Description: "d"},
+	}
+
+	res := &MergeResult{}
+	res.mergeSequences(target, source)
+
+	if res.SequencesAdded != 1 || len(target.Sequences) != 2 {
+		t.Fatalf("added=%d len=%d", res.SequencesAdded, len(target.Sequences))
+	}
+	if target.Sequences[0].StartValue != 1 || target.Sequences[0].IncrementBy != 1 {
+		t.Errorf("existing sequence was modified: %+v", target.Sequences[0])
+	}
+	added := target.Sequences[1]
+	if added.Name != "fresh" || added.StartValue != 5 || added.IncrementBy != 2 || added.MinValue != 1 || added.MaxValue != 99 ||
+		added.CacheSize != 3 || !added.Cycle || added.OwnedByTable != "t" || added.OwnedByColumn != "id" || added.Comment != "c" || added.Description != "d" {
+		t.Errorf("clone lost fields: %+v", added)
+	}
+	if added == source.Sequences[1] {
+		t.Error("sequence must be cloned, not shared")
+	}
+	source.Sequences[1].StartValue = 777
+	if added.StartValue != 5 {
+		t.Error("clone must be independent of source")
+	}
+	if cloneSequence(nil) != nil {
+		t.Error("cloneSequence(nil) must be nil")
+	}
+}
+
+func TestCloneSchemaIsIndependent(t *testing.T) {
+	src := models.InitSchema("public")
+	src.Description, src.Owner, src.Comment, src.Sequence = "d", "o", "c", 4
+	src.Permissions["r"] = "all"
+	src.Metadata["k"] = "v"
+	src.Scripts = []*models.Script{{Name: "s"}}
+
+	tbl := models.InitTable("t", "public")
+	col := models.InitColumn("id", "t", "public")
+	col.Type = "integer"
+	tbl.Columns["id"] = col
+	tbl.Constraints["pk"] = &models.Constraint{Name: "pk", Type: models.PrimaryKeyConstraint, Columns: []string{"id"}}
+	tbl.Indexes["i"] = &models.Index{Name: "i", Columns: []string{"id"}, Include: []string{"x"}}
+	tbl.Metadata["tm"] = 1
+	src.Tables = []*models.Table{tbl}
+
+	v := models.InitView("v", "public")
+	v.Definition = "select 1"
+	v.Columns["c"] = &models.Column{Name: "c"}
+	v.Metadata["vm"] = 1
+	src.Views = []*models.View{v}
+	src.Sequences = []*models.Sequence{{Name: "sq", StartValue: 3}}
+	src.Enums = []*models.Enum{{Name: "e", Values: []string{"a", "b"}}}
+	src.Relations = []*models.Relationship{{Name: "r", FromColumns: []string{"a"}, ToColumns: []string{"b"}, Properties: map[string]string{"p": "q"}}}
+
+	got := cloneSchema(src)
+	if got == src || got.Name != "public" || got.Description != "d" || got.Owner != "o" || got.Comment != "c" || got.Sequence != 4 {
+		t.Fatalf("scalar fields: %+v", got)
+	}
+	if got.Permissions["r"] != "all" || got.Metadata["k"] != "v" || len(got.Scripts) != 1 {
+		t.Errorf("maps/scripts: %+v", got)
+	}
+	if len(got.Tables) != 1 || got.Tables[0] == tbl || got.Tables[0].Columns["id"] == col || got.Tables[0].Columns["id"].Type != "integer" {
+		t.Errorf("tables not deep cloned: %+v", got.Tables)
+	}
+	if len(got.Views) != 1 || got.Views[0] == v || got.Views[0].Definition != "select 1" || got.Views[0].Columns["c"] == v.Columns["c"] || got.Views[0].Metadata["vm"] != 1 {
+		t.Errorf("views not deep cloned: %+v", got.Views)
+	}
+	if len(got.Sequences) != 1 || got.Sequences[0] == src.Sequences[0] || got.Sequences[0].StartValue != 3 {
+		t.Errorf("sequences: %+v", got.Sequences)
+	}
+	if len(got.Enums) != 1 || got.Enums[0] == src.Enums[0] || strings.Join(got.Enums[0].Values, ",") != "a,b" {
+		t.Errorf("enums: %+v", got.Enums)
+	}
+	if len(got.Relations) != 1 || got.Relations[0] == src.Relations[0] || got.Relations[0].Properties["p"] != "q" {
+		t.Errorf("relations: %+v", got.Relations)
+	}
+
+	// Mutating the clone must not touch the source.
+	got.Permissions["r"] = "none"
+	got.Metadata["k"] = "changed"
+	got.Tables[0].Columns["id"].Type = "text"
+	got.Tables[0].Constraints["pk"].Columns[0] = "zzz"
+	got.Tables[0].Indexes["i"].Columns[0] = "zzz"
+	got.Tables[0].Metadata["tm"] = 2
+	got.Enums[0].Values[0] = "zzz"
+	got.Relations[0].FromColumns[0] = "zzz"
+	got.Relations[0].Properties["p"] = "zzz"
+	got.Views[0].Columns["c"].Name = "zzz"
+	if src.Permissions["r"] != "all" || src.Metadata["k"] != "v" || col.Type != "integer" ||
+		tbl.Constraints["pk"].Columns[0] != "id" || tbl.Indexes["i"].Columns[0] != "id" || tbl.Metadata["tm"] != 1 ||
+		src.Enums[0].Values[0] != "a" || src.Relations[0].FromColumns[0] != "a" || src.Relations[0].Properties["p"] != "q" ||
+		v.Columns["c"].Name != "c" {
+		t.Error("clone shares state with the source")
+	}
+
+	if cloneSchema(nil) != nil {
+		t.Error("cloneSchema(nil) must be nil")
+	}
+	bare := cloneSchema(&models.Schema{Name: "bare"})
+	if bare.Permissions != nil || bare.Metadata != nil {
+		t.Errorf("nil maps must stay nil: %+v", bare)
+	}
+}
+
+func TestCloneNilInputs(t *testing.T) {
+	if cloneTable(nil) != nil || cloneColumn(nil) != nil || cloneConstraint(nil) != nil || cloneIndex(nil) != nil ||
+		cloneView(nil) != nil || cloneEnum(nil) != nil || cloneRelation(nil) != nil || cloneDomain(nil) != nil {
+		t.Error("clone of nil must be nil")
+	}
+}
+
+func TestCloneDomainAndRelation(t *testing.T) {
+	d := &models.Domain{Name: "d", Description: "x", Comment: "c", Sequence: 2, Metadata: map[string]any{"k": 1}, Tables: []*models.DomainTable{{TableName: "t", SchemaName: "s"}}}
+	cd := cloneDomain(d)
+	if cd == d || cd.Name != "d" || cd.Description != "x" || cd.Comment != "c" || cd.Sequence != 2 || cd.Metadata["k"] != 1 || len(cd.Tables) != 1 {
+		t.Errorf("domain clone: %+v", cd)
+	}
+	cd.Metadata["k"] = 2
+	if d.Metadata["k"] != 1 {
+		t.Error("domain metadata shared")
+	}
+
+	r := &models.Relationship{Name: "r", Type: "one_to_many", FromTable: "a", FromSchema: "s", ToTable: "b", ToSchema: "s", ForeignKey: "fk", ThroughTable: "l", ThroughSchema: "s", Description: "d", Sequence: 3}
+	cr := cloneRelation(r)
+	if cr == r || cr.Name != "r" || cr.Type != "one_to_many" || cr.FromTable != "a" || cr.ToTable != "b" || cr.ForeignKey != "fk" || cr.ThroughTable != "l" || cr.Description != "d" || cr.Sequence != 3 {
+		t.Errorf("relation clone: %+v", cr)
+	}
+	if cr.Properties != nil {
+		t.Errorf("nil properties must stay nil")
+	}
+}
+
+func TestExtractTypeParts(t *testing.T) {
+	tests := []struct {
+		name                         string
+		col                          models.Column
+		wantType                     string
+		wantLen, wantPrec, wantScale int
+	}{
+		{"plain", models.Column{Type: "TEXT"}, "text", 0, 0, 0},
+		{"trim and lower", models.Column{Type: "  Integer "}, "integer", 0, 0, 0},
+		{"embedded length", models.Column{Type: "varchar(50)"}, "varchar", 50, 0, 0},
+		{"embedded precision and scale", models.Column{Type: "numeric(10,2)"}, "numeric", 0, 10, 2},
+		{"embedded with spaces", models.Column{Type: "numeric( 10 , 2 )"}, "numeric", 0, 10, 2},
+		{"fields win over embedded precision", models.Column{Type: "numeric(10,2)", Precision: 12, Scale: 4}, "numeric", 0, 12, 4},
+		{"fields win over embedded length", models.Column{Type: "varchar(50)", Length: 80}, "varchar", 80, 0, 0},
+		{"precision field blocks embedded length", models.Column{Type: "varchar(50)", Precision: 5}, "varchar", 0, 5, 0},
+		{"non-numeric modifier", models.Column{Type: "varchar(max)"}, "varchar", 0, 0, 0},
+		{"zero modifier", models.Column{Type: "char(0)"}, "char", 0, 0, 0},
+		{"serial sugar", models.Column{Type: "bigserial"}, "bigint", 0, 0, 0},
+		{"smallserial sugar", models.Column{Type: "smallserial"}, "smallint", 0, 0, 0},
+		{"three modifiers ignored", models.Column{Type: "x(1,2,3)"}, "x", 0, 0, 0},
+		{"empty", models.Column{}, "", 0, 0, 0},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			col := tt.col
+			gt, gl, gp, gs := extractTypeParts(&col)
+			if gt != tt.wantType || gl != tt.wantLen || gp != tt.wantPrec || gs != tt.wantScale {
+				t.Errorf("got (%q,%d,%d,%d), want (%q,%d,%d,%d)", gt, gl, gp, gs, tt.wantType, tt.wantLen, tt.wantPrec, tt.wantScale)
+			}
+		})
+	}
+}
+
+func TestColumnTypeConflict(t *testing.T) {
+	c := func(typ string, l, p, s int) *models.Column {
+		return &models.Column{Type: typ, Length: l, Precision: p, Scale: s}
+	}
+	tests := []struct {
+		name string
+		a, b *models.Column
+		want bool
+	}{
+		{"nil target", nil, c("text", 0, 0, 0), false},
+		{"nil source", c("text", 0, 0, 0), nil, false},
+		{"same", c("text", 0, 0, 0), c("TEXT", 0, 0, 0), false},
+		{"different base", c("text", 0, 0, 0), c("integer", 0, 0, 0), true},
+		{"embedded equals field", c("varchar(50)", 0, 0, 0), c("varchar", 50, 0, 0), false},
+		{"different length", c("varchar", 50, 0, 0), c("varchar", 80, 0, 0), true},
+		{"different scale", c("numeric", 0, 10, 2), c("numeric", 0, 10, 3), true},
+		{"serial vs int", c("bigserial", 0, 0, 0), c("bigint", 0, 0, 0), false},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := columnTypeConflict(tt.a, tt.b); got != tt.want {
+				t.Errorf("got %v, want %v", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestDescribeColumnType(t *testing.T) {
+	tests := []struct {
+		col  *models.Column
+		want string
+	}{
+		{nil, ""},
+		{&models.Column{}, ""},
+		{&models.Column{Type: "  "}, ""},
+		{&models.Column{Type: "text"}, "text"},
+		{&models.Column{Type: " numeric ", Precision: 10, Scale: 2}, "numeric(10,2)"},
+		{&models.Column{Type: "numeric", Precision: 10}, "numeric(10)"},
+		{&models.Column{Type: "varchar", Length: 50}, "varchar(50)"},
+		{&models.Column{Type: "varchar", Length: 50, Precision: 7}, "varchar(7)"},
+	}
+	for _, tt := range tests {
+		if got := describeColumnType(tt.col); got != tt.want {
+			t.Errorf("describeColumnType(%+v) = %q, want %q", tt.col, got, tt.want)
+		}
+	}
+}
+
+func TestFirstNonEmpty(t *testing.T) {
+	if got := firstNonEmpty("", "  ", "x", "y"); got != "x" {
+		t.Errorf("got %q", got)
+	}
+	if got := firstNonEmpty(); got != "" {
+		t.Errorf("none: %q", got)
+	}
+	if got := firstNonEmpty("", " "); got != "" {
+		t.Errorf("all blank: %q", got)
+	}
+}
+
+func TestGetColumnTypeConflictSummary(t *testing.T) {
+	conflicts := []ColumnTypeConflict{
+		{Schema: "s", Table: "t", Column: "a", TargetType: "text", SourceType: "integer"},
+		{Schema: "s", Table: "t", Column: "b", TargetType: "int", SourceType: "text"},
+		{Schema: "s", Table: "u", Column: "c", TargetType: "x", SourceType: "y"},
+	}
+	res := &MergeResult{TypeConflicts: conflicts}
+
+	if GetColumnTypeConflictSummary(nil, 5) != "" || GetColumnTypeConflictSummary(&MergeResult{}, 5) != "" {
+		t.Error("no conflicts must yield empty summary")
+	}
+
+	all := GetColumnTypeConflictSummary(res, 0)
+	if !strings.Contains(all, "column type conflicts detected:") || !strings.Contains(all, "s.t.a: target=text source=integer") ||
+		!strings.Contains(all, "s.u.c: target=x source=y") || strings.Contains(all, "more") {
+		t.Errorf("unlimited summary:\n%s", all)
+	}
+	if neg := GetColumnTypeConflictSummary(res, -1); neg != all {
+		t.Error("negative limit must behave as unlimited")
+	}
+
+	limited := GetColumnTypeConflictSummary(res, 2)
+	if !strings.Contains(limited, "s.t.b") || strings.Contains(limited, "s.u.c") || !strings.HasSuffix(limited, "... and 1 more") {
+		t.Errorf("limited summary:\n%s", limited)
+	}
+	exact := GetColumnTypeConflictSummary(res, 3)
+	if strings.Contains(exact, "more") {
+		t.Errorf("limit == len must not truncate:\n%s", exact)
+	}
+}
+
+func TestMinHelper(t *testing.T) {
+	if min(1, 2) != 1 || min(2, 1) != 1 || min(3, 3) != 3 {
+		t.Error("min")
+	}
+}

pkg/merge/merge_regression_test.go

@@ -0,0 +1,58 @@
+package merge
+
+import (
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+)
+
+func sourceWithRelationship() *models.Database {
+	db := models.InitDatabase("src")
+	s := models.InitSchema("sales")
+	orders := models.InitTable("orders", "sales")
+	orders.Tablespace = "fast"
+	orders.GUID = "guid-1"
+	orders.Relationships["fk_cust"] = &models.Relationship{
+		Name: "fk_cust", FromTable: "orders", ToTable: "customers",
+		FromColumns: []string{"cust_id"}, ToColumns: []string{"id"},
+	}
+	s.Tables = append(s.Tables, orders, models.InitTable("Audit", "sales"))
+	db.Schemas = append(db.Schemas, s)
+	return db
+}
+
+func TestCloneTable_CopiesRelationshipsTablespaceGUID(t *testing.T) {
+	src := sourceWithRelationship()
+	target := models.InitDatabase("tgt")
+	MergeDatabases(target, src, nil)
+
+	got := target.Schemas[0].Tables[0]
+	if got.Tablespace != "fast" || got.GUID != "guid-1" {
+		t.Errorf("tablespace/guid lost: %+v", got)
+	}
+	rel := got.Relationships["fk_cust"]
+	if rel == nil || rel.ToTable != "customers" {
+		t.Fatalf("relationship lost: %+v", got.Relationships)
+	}
+	if rel == src.Schemas[0].Tables[0].Relationships["fk_cust"] {
+		t.Error("relationship must be deep-copied")
+	}
+	rel.FromColumns[0] = "changed"
+	if src.Schemas[0].Tables[0].Relationships["fk_cust"].FromColumns[0] != "cust_id" {
+		t.Error("relationship columns shared with source")
+	}
+}
+
+func TestMerge_SkipTablesAppliesToNewSchemas(t *testing.T) {
+	src := sourceWithRelationship()
+	target := models.InitDatabase("tgt")
+	MergeDatabases(target, src, &MergeOptions{SkipTableNames: map[string]bool{"audit": true}})
+
+	tables := target.Schemas[0].Tables
+	if len(tables) != 1 || tables[0].Name != "orders" {
+		t.Errorf("skipped table copied into new schema: %+v", tables)
+	}
+	if len(src.Schemas[0].Tables) != 2 {
+		t.Error("source must not be modified")
+	}
+}

pkg/merge/merge_test.go

@@ -721,3 +721,32 @@ func TestComplexMerge(t *testing.T) {
 		t.Error("Expected ukey_users_guid constraint to exist")
 	}
 }
+
+func TestMergeDatabases_Metadata(t *testing.T) {
+	target := &models.Database{Metadata: map[string]any{
+		"refs": []string{"a", "b"},
+		"name": "target",
+	}}
+	source := &models.Database{Metadata: map[string]any{
+		"refs":  []string{"b", "c"},
+		"name":  "source",
+		"extra": []string{"x"},
+	}}
+
+	MergeDatabases(target, source, nil)
+
+	if got := target.Metadata["refs"].([]string); strings.Join(got, ",") != "a,b,c" {
+		t.Errorf("refs = %v, want [a b c]", got)
+	}
+	if got := target.Metadata["name"]; got != "target" {
+		t.Errorf("name = %v, want target (existing scalar keys are kept)", got)
+	}
+	extra := target.Metadata["extra"].([]string)
+	if strings.Join(extra, ",") != "x" {
+		t.Errorf("extra = %v, want [x]", extra)
+	}
+	source.Metadata["extra"].([]string)[0] = "changed"
+	if extra[0] != "x" {
+		t.Error("copied slice must not alias the source")
+	}
+}

pkg/models/models.go

@@ -20,6 +20,7 @@ const (
 	PostgresqlDatabaseType DatabaseType = "pgsql"  // PostgreSQL database
 	MSSQLDatabaseType      DatabaseType = "mssql"  // Microsoft SQL Server database
 	SqlLiteDatabaseType    DatabaseType = "sqlite" // SQLite database
+	MySQLDatabaseType      DatabaseType = "mysql"  // MySQL/MariaDB database
 )
 
 // Database represents the complete database schema
@@ -227,23 +228,27 @@ func (d *Sequence) SQLName() string {
 
 // Column represents a table column
 type Column struct {
-	Name          string         `json:"name" yaml:"name" xml:"name"`
-	Description   string         `json:"description,omitempty" yaml:"description,omitempty" xml:"description,omitempty"`
-	Table         string         `json:"table" yaml:"table" xml:"table"`
-	Schema        string         `json:"schema" yaml:"schema" xml:"schema"`
-	Type          string         `json:"type" yaml:"type" xml:"type"`
-	Length        int            `json:"length,omitempty" yaml:"length,omitempty" xml:"length,omitempty"`
-	Precision     int            `json:"precision,omitempty" yaml:"precision,omitempty" xml:"precision,omitempty"`
-	Scale         int            `json:"scale,omitempty" yaml:"scale,omitempty" xml:"scale,omitempty"`
-	NotNull       bool           `json:"not_null" yaml:"not_null" xml:"not_null"`
-	Default       any            `json:"default,omitempty" yaml:"default,omitempty" xml:"default,omitempty"`
-	AutoIncrement bool           `json:"auto_increment" yaml:"auto_increment" xml:"auto_increment"`
-	IsPrimaryKey  bool           `json:"is_primary_key" yaml:"is_primary_key" xml:"is_primary_key"`
-	Comment       string         `json:"comment,omitempty" yaml:"comment,omitempty" xml:"comment,omitempty"`
-	Collation     string         `json:"collation,omitempty" yaml:"collation,omitempty" xml:"collation,omitempty"`
-	Metadata      map[string]any `json:"metadata,omitempty" yaml:"metadata,omitempty" xml:"-"`
-	Sequence      uint           `json:"sequence,omitempty" yaml:"sequence,omitempty" xml:"sequence,omitempty"`
-	GUID          string         `json:"guid" yaml:"guid" xml:"guid"`
+	Name                 string         `json:"name" yaml:"name" xml:"name"`
+	Description          string         `json:"description,omitempty" yaml:"description,omitempty" xml:"description,omitempty"`
+	Table                string         `json:"table" yaml:"table" xml:"table"`
+	Schema               string         `json:"schema" yaml:"schema" xml:"schema"`
+	Type                 string         `json:"type" yaml:"type" xml:"type"`
+	Length               int            `json:"length,omitempty" yaml:"length,omitempty" xml:"length,omitempty"`
+	Precision            int            `json:"precision,omitempty" yaml:"precision,omitempty" xml:"precision,omitempty"`
+	Scale                int            `json:"scale,omitempty" yaml:"scale,omitempty" xml:"scale,omitempty"`
+	NotNull              bool           `json:"not_null" yaml:"not_null" xml:"not_null"`
+	Default              any            `json:"default,omitempty" yaml:"default,omitempty" xml:"default,omitempty"`
+	AutoIncrement        bool           `json:"auto_increment" yaml:"auto_increment" xml:"auto_increment"`
+	IsPrimaryKey         bool           `json:"is_primary_key" yaml:"is_primary_key" xml:"is_primary_key"`
+	Comment              string         `json:"comment,omitempty" yaml:"comment,omitempty" xml:"comment,omitempty"`
+	Collation            string         `json:"collation,omitempty" yaml:"collation,omitempty" xml:"collation,omitempty"`
+	Metadata             map[string]any `json:"metadata,omitempty" yaml:"metadata,omitempty" xml:"-"`
+	Sequence             uint           `json:"sequence,omitempty" yaml:"sequence,omitempty" xml:"sequence,omitempty"`
+	GUID                 string         `json:"guid" yaml:"guid" xml:"guid"`
+	Generated            bool           `json:"generated,omitempty" yaml:"generated,omitempty" xml:"generated,omitempty"`
+	GenerationExpression string         `json:"generation_expression,omitempty" yaml:"generation_expression,omitempty" xml:"generation_expression,omitempty"`
+	Identity             bool           `json:"identity,omitempty" yaml:"identity,omitempty" xml:"identity,omitempty"`
+	IdentityGeneration   string         `json:"identity_generation,omitempty" yaml:"identity_generation,omitempty" xml:"identity_generation,omitempty"` // "ALWAYS" or "BY DEFAULT"
 }
 
 // SQLName returns the column name in lowercase for SQL compatibility.

pkg/models/models_test.go

@@ -0,0 +1,232 @@
+package models
+
+import (
+	"testing"
+	"time"
+)
+
+func TestSQLNameLowercases(t *testing.T) {
+	tests := []struct {
+		name string
+		got  string
+	}{
+		{"database", (&Database{Name: "MyDB"}).SQLName()},
+		{"domain", (&Domain{Name: "MyDomain"}).SQLName()},
+		{"schema", (&Schema{Name: "MySchema"}).SQLName()},
+		{"table", (&Table{Name: "MyTable"}).SQLName()},
+		{"view", (&View{Name: "MyView"}).SQLName()},
+		{"sequence", (&Sequence{Name: "MySeq"}).SQLName()},
+		{"column", (&Column{Name: "MyCol"}).SQLName()},
+		{"index", (&Index{Name: "MyIdx"}).SQLName()},
+		{"relationship", (&Relationship{Name: "MyRel"}).SQLName()},
+		{"constraint", (&Constraint{Name: "MyCon"}).SQLName()},
+		{"enum", (&Enum{Name: "MyEnum"}).SQLName()},
+		{"script", (&Script{Name: "MyScript"}).SQLName()},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if tt.got == "" || tt.got != lower(tt.got) {
+				t.Errorf("SQLName not lowercase: %q", tt.got)
+			}
+		})
+	}
+	if got := (&Table{}).SQLName(); got != "" {
+		t.Errorf("empty name: %q", got)
+	}
+	if got := (&Table{Name: "MyTable"}).SQLName(); got != "mytable" {
+		t.Errorf("got %q", got)
+	}
+}
+
+func lower(s string) string {
+	b := []byte(s)
+	for i, c := range b {
+		if c >= 'A' && c <= 'Z' {
+			b[i] = c + 32
+		}
+	}
+	return string(b)
+}
+
+func TestUpdateDatePropagates(t *testing.T) {
+	db := InitDatabase("d")
+	schema := InitSchema("s")
+	schema.RefDatabase = db
+	table := InitTable("t", "s")
+	table.RefSchema = schema
+
+	table.UpdateDate()
+	for name, v := range map[string]string{"table": table.UpdatedAt, "schema": schema.UpdatedAt, "database": db.UpdatedAt} {
+		ts, err := time.Parse(time.RFC3339, v)
+		if err != nil {
+			t.Fatalf("%s UpdatedAt %q: %v", name, v, err)
+		}
+		if time.Since(ts) > time.Minute {
+			t.Errorf("%s UpdatedAt too old: %v", name, ts)
+		}
+	}
+
+	// Without references only the receiver is updated.
+	lone := InitTable("lone", "s")
+	lone.UpdateDate()
+	if lone.UpdatedAt == "" {
+		t.Error("lone table not updated")
+	}
+	loneSchema := InitSchema("x")
+	loneSchema.UpdateDate()
+	if loneSchema.UpdatedAt == "" {
+		t.Error("lone schema not updated")
+	}
+}
+
+func TestGetPrimaryKey(t *testing.T) {
+	tests := []struct {
+		name string
+		cols []*Column
+		want string
+	}{
+		{"none", []*Column{{Name: "a"}}, ""},
+		{"single", []*Column{{Name: "a"}, {Name: "id", IsPrimaryKey: true}}, "id"},
+		{"composite ordered by sequence", []*Column{
+			{Name: "a", IsPrimaryKey: true, Sequence: 2},
+			{Name: "b", IsPrimaryKey: true, Sequence: 1},
+		}, "b"},
+		{"composite without sequence falls back to name", []*Column{
+			{Name: "z", IsPrimaryKey: true},
+			{Name: "m", IsPrimaryKey: true},
+		}, "m"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			tbl := InitTable("t", "s")
+			for _, c := range tt.cols {
+				tbl.Columns[c.Name] = c
+			}
+			got := tbl.GetPrimaryKey()
+			if tt.want == "" {
+				if got != nil {
+					t.Errorf("expected nil, got %s", got.Name)
+				}
+				return
+			}
+			if got == nil || got.Name != tt.want {
+				t.Errorf("got %v, want %s", got, tt.want)
+			}
+		})
+	}
+	if InitTable("empty", "s").GetPrimaryKey() != nil {
+		t.Error("empty table must have no PK")
+	}
+}
+
+func TestColumnLess(t *testing.T) {
+	tests := []struct {
+		a, b *Column
+		want bool
+	}{
+		{&Column{Name: "a", Sequence: 1}, &Column{Name: "b", Sequence: 2}, true},
+		{&Column{Name: "a", Sequence: 2}, &Column{Name: "b", Sequence: 1}, false},
+		{&Column{Name: "a"}, &Column{Name: "b"}, true},
+		{&Column{Name: "b"}, &Column{Name: "a"}, false},
+		{&Column{Name: "b", Sequence: 1}, &Column{Name: "a"}, false}, // one side unsequenced: by name
+		{&Column{Name: "a", Sequence: 1}, &Column{Name: "b"}, true},
+	}
+	for i, tt := range tests {
+		if got := columnLess(tt.a, tt.b); got != tt.want {
+			t.Errorf("case %d: got %v, want %v", i, got, tt.want)
+		}
+	}
+}
+
+func TestGetForeignKeys(t *testing.T) {
+	tbl := InitTable("t", "s")
+	add := func(name string, typ ConstraintType, seq uint) {
+		c := InitConstraint(name, typ)
+		c.Sequence = seq
+		tbl.Constraints[name] = c
+	}
+	add("pk", PrimaryKeyConstraint, 0)
+	add("fk_b", ForeignKeyConstraint, 0)
+	add("fk_a", ForeignKeyConstraint, 0)
+	add("uq", UniqueConstraint, 0)
+
+	got := tbl.GetForeignKeys()
+	if len(got) != 2 || got[0].Name != "fk_a" || got[1].Name != "fk_b" {
+		t.Errorf("by name: %v", got)
+	}
+
+	tbl.Constraints["fk_a"].Sequence = 5
+	tbl.Constraints["fk_b"].Sequence = 2
+	got = tbl.GetForeignKeys()
+	if got[0].Name != "fk_b" || got[1].Name != "fk_a" {
+		t.Errorf("by sequence: %v", got)
+	}
+
+	if got := InitTable("e", "s").GetForeignKeys(); got == nil || len(got) != 0 {
+		t.Errorf("empty table must give non-nil empty slice, got %v", got)
+	}
+}
+
+func TestInitConstructors(t *testing.T) {
+	db := InitDatabase("db")
+	if db.Name != "db" || db.Schemas == nil || db.Domains == nil || db.Metadata == nil || db.GUID == "" {
+		t.Errorf("InitDatabase: %+v", db)
+	}
+	s := InitSchema("s")
+	if s.Name != "s" || s.Tables == nil || s.Views == nil || s.Sequences == nil || s.Permissions == nil || s.Metadata == nil || s.Scripts == nil || s.GUID == "" {
+		t.Errorf("InitSchema: %+v", s)
+	}
+	tb := InitTable("t", "s")
+	if tb.Name != "t" || tb.Schema != "s" || tb.Columns == nil || tb.Constraints == nil || tb.Indexes == nil || tb.Relationships == nil || tb.Metadata == nil || tb.GUID == "" {
+		t.Errorf("InitTable: %+v", tb)
+	}
+	c := InitColumn("c", "t", "s")
+	if c.Name != "c" || c.Table != "t" || c.Schema != "s" || c.Metadata == nil || c.GUID == "" {
+		t.Errorf("InitColumn: %+v", c)
+	}
+	ix := InitIndex("i", "t", "s")
+	if ix.Name != "i" || ix.Table != "t" || ix.Schema != "s" || ix.Columns == nil || ix.Include == nil || ix.Metadata == nil || ix.GUID == "" {
+		t.Errorf("InitIndex: %+v", ix)
+	}
+	r := InitRelation("r", "s")
+	if r.Name != "r" || r.FromSchema != "s" || r.ToSchema != "s" || r.Properties == nil || r.FromColumns == nil || r.ToColumns == nil || r.GUID == "" {
+		t.Errorf("InitRelation: %+v", r)
+	}
+	rel := InitRelationship("rel", RelationType("one_to_many"))
+	if rel.Name != "rel" || rel.Type != "one_to_many" || rel.Properties == nil || rel.GUID == "" {
+		t.Errorf("InitRelationship: %+v", rel)
+	}
+	con := InitConstraint("k", UniqueConstraint)
+	if con.Name != "k" || con.Type != UniqueConstraint || con.Columns == nil || con.ReferencedColumns == nil || con.GUID == "" {
+		t.Errorf("InitConstraint: %+v", con)
+	}
+	sc := InitScript("sc")
+	if sc.Name != "sc" || sc.RunAfter == nil || sc.Metadata == nil || sc.GUID == "" {
+		t.Errorf("InitScript: %+v", sc)
+	}
+	v := InitView("v", "s")
+	if v.Name != "v" || v.Schema != "s" || v.Columns == nil || v.Metadata == nil || v.GUID == "" {
+		t.Errorf("InitView: %+v", v)
+	}
+	sq := InitSequence("sq", "s")
+	if sq.Name != "sq" || sq.Schema != "s" || sq.IncrementBy != 1 || sq.StartValue != 1 || sq.GUID == "" {
+		t.Errorf("InitSequence: %+v", sq)
+	}
+	d := InitDomain("d")
+	if d.Name != "d" || d.Tables == nil || d.Metadata == nil || d.GUID == "" {
+		t.Errorf("InitDomain: %+v", d)
+	}
+	dt := InitDomainTable("t", "s")
+	if dt.TableName != "t" || dt.SchemaName != "s" || dt.GUID == "" {
+		t.Errorf("InitDomainTable: %+v", dt)
+	}
+	e := InitEnum("e", "s")
+	if e.Name != "e" || e.Schema != "s" || e.Values == nil || e.GUID == "" {
+		t.Errorf("InitEnum: %+v", e)
+	}
+
+	// GUIDs are unique per call.
+	if InitTable("t", "s").GUID == InitTable("t", "s").GUID {
+		t.Error("GUIDs must be unique")
+	}
+}

pkg/models/sorting_test.go

@@ -0,0 +1,171 @@
+package models
+
+import (
+	"reflect"
+	"testing"
+)
+
+type sortCase struct {
+	name string
+	seq  uint
+}
+
+var sortFixture = []sortCase{{"Banana", 3}, {"apple", 1}, {"Cherry", 2}}
+
+var (
+	wantNameAsc  = []string{"apple", "Banana", "Cherry"}
+	wantNameDesc = []string{"Cherry", "Banana", "apple"}
+	wantSeqAsc   = []string{"apple", "Cherry", "Banana"}
+	wantSeqDesc  = []string{"Banana", "Cherry", "apple"}
+)
+
+func checkNames(t *testing.T, label string, got, want []string) {
+	t.Helper()
+	if !reflect.DeepEqual(got, want) {
+		t.Errorf("%s: got %v, want %v", label, got, want)
+	}
+}
+
+// runSortSuite exercises a by-name and by-sequence sorter pair over the shared fixture.
+func runSortSuite[T any](t *testing.T, build func(sortCase) T, name func(T) string,
+	byName, bySeq func([]T, bool) error,
+) {
+	t.Helper()
+	mk := func() []T {
+		out := make([]T, 0, len(sortFixture))
+		for _, c := range sortFixture {
+			out = append(out, build(c))
+		}
+		return out
+	}
+	names := func(items []T) []string {
+		out := make([]string, 0, len(items))
+		for _, it := range items {
+			out = append(out, name(it))
+		}
+		return out
+	}
+	if byName != nil {
+		items := mk()
+		_ = byName(items, false)
+		checkNames(t, "name asc", names(items), wantNameAsc)
+		_ = byName(items, true)
+		checkNames(t, "name desc", names(items), wantNameDesc)
+		_ = byName(nil, false)
+		_ = byName([]T{}, true)
+	}
+	if bySeq != nil {
+		items := mk()
+		_ = bySeq(items, false)
+		checkNames(t, "seq asc", names(items), wantSeqAsc)
+		_ = bySeq(items, true)
+		checkNames(t, "seq desc", names(items), wantSeqDesc)
+		_ = bySeq(nil, false)
+	}
+}
+
+func TestSortSchemas(t *testing.T) {
+	runSortSuite(t, func(c sortCase) *Schema { return &Schema{Name: c.name, Sequence: c.seq} },
+		func(s *Schema) string { return s.Name }, SortSchemasByName, SortSchemasBySequence)
+}
+
+func TestSortTables(t *testing.T) {
+	runSortSuite(t, func(c sortCase) *Table { return &Table{Name: c.name, Sequence: c.seq} },
+		func(s *Table) string { return s.Name }, SortTablesByName, SortTablesBySequence)
+}
+
+func TestSortColumns(t *testing.T) {
+	runSortSuite(t, func(c sortCase) *Column { return &Column{Name: c.name, Sequence: c.seq} },
+		func(s *Column) string { return s.Name }, SortColumnsByName, SortColumnsBySequence)
+}
+
+func TestSortViews(t *testing.T) {
+	runSortSuite(t, func(c sortCase) *View { return &View{Name: c.name, Sequence: c.seq} },
+		func(s *View) string { return s.Name }, SortViewsByName, SortViewsBySequence)
+}
+
+func TestSortSequences(t *testing.T) {
+	runSortSuite(t, func(c sortCase) *Sequence { return &Sequence{Name: c.name, Sequence: c.seq} },
+		func(s *Sequence) string { return s.Name }, SortSequencesByName, SortSequencesBySequence)
+}
+
+func TestSortIndexes(t *testing.T) {
+	runSortSuite(t, func(c sortCase) *Index { return &Index{Name: c.name, Sequence: c.seq} },
+		func(s *Index) string { return s.Name }, SortIndexesByName, SortIndexesBySequence)
+}
+
+func TestSortNameOnly(t *testing.T) {
+	runSortSuite(t, func(c sortCase) *Constraint { return &Constraint{Name: c.name} },
+		func(s *Constraint) string { return s.Name }, SortConstraintsByName, nil)
+	runSortSuite(t, func(c sortCase) *Relationship { return &Relationship{Name: c.name} },
+		func(s *Relationship) string { return s.Name }, SortRelationshipsByName, nil)
+	runSortSuite(t, func(c sortCase) *Script { return &Script{Name: c.name} },
+		func(s *Script) string { return s.Name }, SortScriptsByName, nil)
+	runSortSuite(t, func(c sortCase) *Enum { return &Enum{Name: c.name} },
+		func(s *Enum) string { return s.Name }, SortEnumsByName, nil)
+}
+
+func TestSortStableForTies(t *testing.T) {
+	cols := []*Column{{Name: "x", Description: "first"}, {Name: "X", Description: "second"}, {Name: "x", Description: "third"}}
+	_ = SortColumnsByName(cols, false)
+	if cols[0].Description != "first" || cols[1].Description != "second" || cols[2].Description != "third" {
+		t.Errorf("ties must keep input order: %v %v %v", cols[0].Description, cols[1].Description, cols[2].Description)
+	}
+	_ = SortColumnsBySequence(cols, true)
+	if cols[0].Description != "first" || cols[2].Description != "third" {
+		t.Errorf("sequence ties must keep input order")
+	}
+}
+
+func TestSortMapVariants(t *testing.T) {
+	cols := map[string]*Column{}
+	idx := map[string]*Index{}
+	cons := map[string]*Constraint{}
+	rels := map[string]*Relationship{}
+	for _, c := range sortFixture {
+		cols[c.name] = &Column{Name: c.name, Sequence: c.seq}
+		idx[c.name] = &Index{Name: c.name, Sequence: c.seq}
+		cons[c.name] = &Constraint{Name: c.name}
+		rels[c.name] = &Relationship{Name: c.name}
+	}
+	colNames := func(l []*Column) (o []string) {
+		for _, x := range l {
+			o = append(o, x.Name)
+		}
+		return
+	}
+	idxNames := func(l []*Index) (o []string) {
+		for _, x := range l {
+			o = append(o, x.Name)
+		}
+		return
+	}
+	conNames := func(l []*Constraint) (o []string) {
+		for _, x := range l {
+			o = append(o, x.Name)
+		}
+		return
+	}
+	relNames := func(l []*Relationship) (o []string) {
+		for _, x := range l {
+			o = append(o, x.Name)
+		}
+		return
+	}
+
+	checkNames(t, "cols name", colNames(SortColumnsMapByName(cols, false)), wantNameAsc)
+	checkNames(t, "cols name desc", colNames(SortColumnsMapByName(cols, true)), wantNameDesc)
+	checkNames(t, "cols seq", colNames(SortColumnsMapBySequence(cols, false)), wantSeqAsc)
+	checkNames(t, "cols seq desc", colNames(SortColumnsMapBySequence(cols, true)), wantSeqDesc)
+	checkNames(t, "idx name", idxNames(SortIndexesMapByName(idx, false)), wantNameAsc)
+	checkNames(t, "idx seq", idxNames(SortIndexesMapBySequence(idx, true)), wantSeqDesc)
+	checkNames(t, "con name", conNames(SortConstraintsMapByName(cons, false)), wantNameAsc)
+	checkNames(t, "rel name", relNames(SortRelationshipsMapByName(rels, true)), wantNameDesc)
+
+	if got := SortColumnsMapByName(nil, false); got == nil || len(got) != 0 {
+		t.Errorf("nil map must give non-nil empty slice")
+	}
+	if len(cols) != 3 {
+		t.Error("input map must not be modified")
+	}
+}

pkg/models/views_test.go

@@ -0,0 +1,249 @@
+package models
+
+import (
+	"reflect"
+	"testing"
+)
+
+// viewFixture builds a two-schema database whose map contents would randomise output order.
+func viewFixture() *Database {
+	db := InitDatabase("shop")
+	db.Description = "desc"
+	db.DatabaseType = PostgresqlDatabaseType
+	db.DatabaseVersion = "16"
+
+	for _, sn := range []string{"sales", "public"} {
+		s := InitSchema(sn)
+		s.Owner = "owner_" + sn
+		s.Scripts = append(s.Scripts, InitScript("seed"))
+
+		users := InitTable("users", sn)
+		for _, cn := range []string{"id", "email", "name"} {
+			c := InitColumn(cn, "users", sn)
+			c.Type = "text"
+			users.Columns[cn] = c
+		}
+		users.Columns["id"].IsPrimaryKey = true
+		users.Columns["id"].NotNull = true
+
+		pk := InitConstraint("users_pkey", PrimaryKeyConstraint)
+		pk.Columns = []string{"id"}
+		users.Constraints["users_pkey"] = pk
+		ck := InitConstraint("users_ck", CheckConstraint)
+		ck.Expression = "id > 0"
+		users.Constraints["users_ck"] = ck
+		users.Indexes["users_idx"] = InitIndex("users_idx", "users", sn)
+
+		orders := InitTable("orders", sn)
+		oid := InitColumn("id", "orders", sn)
+		orders.Columns["id"] = oid
+		uid := InitColumn("user_id", "orders", sn)
+		orders.Columns["user_id"] = uid
+		fk := InitConstraint("orders_user_fk", ForeignKeyConstraint)
+		fk.Columns = []string{"user_id"}
+		fk.ReferencedSchema = sn
+		fk.ReferencedTable = "users"
+		fk.ReferencedColumns = []string{"id"}
+		fk.OnDelete = "CASCADE"
+		orders.Constraints["orders_user_fk"] = fk
+		rel := InitRelationship("orders_users", RelationType("one_to_many"))
+		rel.FromTable, rel.FromSchema = "orders", sn
+		rel.ToTable, rel.ToSchema = "users", sn
+		rel.ForeignKey = "orders_user_fk"
+		rel.ThroughTable, rel.ThroughSchema = "link", sn
+		orders.Relationships["orders_users"] = rel
+		plain := InitRelationship("plain", RelationType("one_to_one"))
+		plain.FromTable, plain.FromSchema = "orders", sn
+		plain.ToTable, plain.ToSchema = "users", sn
+		orders.Relationships["plain"] = plain
+
+		s.Tables = append(s.Tables, users, orders)
+		db.Schemas = append(db.Schemas, s)
+	}
+	return db
+}
+
+func TestToFlatColumns(t *testing.T) {
+	db := viewFixture()
+	first := db.ToFlatColumns()
+	if len(first) != 2*(3+2) {
+		t.Fatalf("got %d columns", len(first))
+	}
+	for i := 1; i < len(first); i++ {
+		if first[i-1].FullyQualifiedName >= first[i].FullyQualifiedName {
+			t.Fatalf("not sorted at %d: %s >= %s", i, first[i-1].FullyQualifiedName, first[i].FullyQualifiedName)
+		}
+	}
+	if first[0].FullyQualifiedName != "shop.public.orders.id" {
+		t.Errorf("first: %s", first[0].FullyQualifiedName)
+	}
+	var id *FlatColumn
+	for _, c := range first {
+		if c.FullyQualifiedName == "shop.sales.users.id" {
+			id = c
+		}
+	}
+	if id == nil || !id.IsPrimaryKey || !id.NotNull || id.Type != "text" || id.DatabaseName != "shop" || id.SchemaName != "sales" || id.TableName != "users" || id.ColumnName != "id" {
+		t.Errorf("flat id column: %+v", id)
+	}
+	for i := 0; i < 20; i++ {
+		if !reflect.DeepEqual(first, db.ToFlatColumns()) {
+			t.Fatal("ToFlatColumns not deterministic")
+		}
+	}
+	if got := InitDatabase("e").ToFlatColumns(); got == nil || len(got) != 0 {
+		t.Errorf("empty db: %v", got)
+	}
+}
+
+func TestToFlatTables(t *testing.T) {
+	got := viewFixture().ToFlatTables()
+	if len(got) != 4 {
+		t.Fatalf("got %d tables", len(got))
+	}
+	// schema order follows the database slice: sales first
+	if got[0].FullyQualifiedName != "shop.sales.users" || got[0].ColumnCount != 3 || got[0].ConstraintCount != 2 || got[0].IndexCount != 1 {
+		t.Errorf("first: %+v", got[0])
+	}
+	if got[1].FullyQualifiedName != "shop.sales.orders" || got[1].ColumnCount != 2 || got[1].ConstraintCount != 1 {
+		t.Errorf("second: %+v", got[1])
+	}
+	if got := InitDatabase("e").ToFlatTables(); got == nil || len(got) != 0 {
+		t.Errorf("empty db: %v", got)
+	}
+}
+
+func TestToFlatConstraints(t *testing.T) {
+	db := viewFixture()
+	got := db.ToFlatConstraints()
+	if len(got) != 6 {
+		t.Fatalf("got %d constraints", len(got))
+	}
+	for i := 1; i < len(got); i++ {
+		if got[i-1].FullyQualifiedName >= got[i].FullyQualifiedName {
+			t.Fatalf("not sorted: %s >= %s", got[i-1].FullyQualifiedName, got[i].FullyQualifiedName)
+		}
+	}
+	var fk, ck *FlatConstraint
+	for _, c := range got {
+		switch c.FullyQualifiedName {
+		case "shop.sales.orders.orders_user_fk":
+			fk = c
+		case "shop.sales.users.users_ck":
+			ck = c
+		}
+	}
+	if fk == nil || fk.ReferencedFQN != "shop.sales.users" || fk.OnDelete != "CASCADE" || fk.Type != ForeignKeyConstraint {
+		t.Errorf("fk: %+v", fk)
+	}
+	if ck == nil || ck.ReferencedFQN != "" || ck.Expression != "id > 0" {
+		t.Errorf("check: %+v", ck)
+	}
+
+	// FK without a referenced table gets no FQN.
+	db2 := InitDatabase("d")
+	s := InitSchema("s")
+	tb := InitTable("t", "s")
+	tb.Constraints["fk"] = InitConstraint("fk", ForeignKeyConstraint)
+	s.Tables = append(s.Tables, tb)
+	db2.Schemas = append(db2.Schemas, s)
+	if out := db2.ToFlatConstraints(); len(out) != 1 || out[0].ReferencedFQN != "" {
+		t.Errorf("unreferenced fk: %+v", out)
+	}
+	if got := InitDatabase("e").ToFlatConstraints(); got == nil || len(got) != 0 {
+		t.Errorf("empty db: %v", got)
+	}
+}
+
+func TestToFlatRelationships(t *testing.T) {
+	db := viewFixture()
+	got := db.ToFlatRelationships()
+	if len(got) != 4 {
+		t.Fatalf("got %d relationships", len(got))
+	}
+	for i := 1; i < len(got); i++ {
+		a, b := got[i-1], got[i]
+		if a.FromFQN > b.FromFQN || (a.FromFQN == b.FromFQN && a.RelationshipName > b.RelationshipName) {
+			t.Fatalf("not sorted at %d", i)
+		}
+	}
+	var through, plain *FlatRelationship
+	for _, r := range got {
+		if r.FromSchema == "sales" && r.RelationshipName == "orders_users" {
+			through = r
+		}
+		if r.FromSchema == "sales" && r.RelationshipName == "plain" {
+			plain = r
+		}
+	}
+	if through == nil || through.ThroughTableFQN != "shop.sales.link" || through.FromFQN != "shop.sales.orders" || through.ToFQN != "shop.sales.users" || through.ForeignKey != "orders_user_fk" {
+		t.Errorf("through: %+v", through)
+	}
+	if plain == nil || plain.ThroughTableFQN != "" {
+		t.Errorf("plain: %+v", plain)
+	}
+	for i := 0; i < 20; i++ {
+		if !reflect.DeepEqual(got, db.ToFlatRelationships()) {
+			t.Fatal("ToFlatRelationships not deterministic")
+		}
+	}
+	if got := InitDatabase("e").ToFlatRelationships(); got == nil || len(got) != 0 {
+		t.Errorf("empty db: %v", got)
+	}
+}
+
+func TestSummaries(t *testing.T) {
+	db := viewFixture()
+	ds := db.ToSummary()
+	if ds.Name != "shop" || ds.Description != "desc" || ds.DatabaseType != PostgresqlDatabaseType || ds.DatabaseVersion != "16" ||
+		ds.SchemaCount != 2 || ds.TotalTables != 4 || ds.TotalColumns != 10 {
+		t.Errorf("database summary: %+v", ds)
+	}
+	if es := InitDatabase("e").ToSummary(); es.SchemaCount != 0 || es.TotalTables != 0 || es.TotalColumns != 0 {
+		t.Errorf("empty summary: %+v", es)
+	}
+
+	ss := db.Schemas[0].ToSummary()
+	if ss.Name != "sales" || ss.Owner != "owner_sales" || ss.TableCount != 2 || ss.ScriptCount != 1 || ss.TotalColumns != 5 || ss.TotalConstraints != 3 {
+		t.Errorf("schema summary: %+v", ss)
+	}
+
+	users := db.Schemas[0].Tables[0].ToSummary()
+	if users.Name != "users" || users.Schema != "sales" || users.ColumnCount != 3 || users.ConstraintCount != 2 || users.IndexCount != 1 ||
+		users.RelationshipCount != 0 || !users.HasPrimaryKey || users.ForeignKeyCount != 0 {
+		t.Errorf("users summary: %+v", users)
+	}
+	orders := db.Schemas[0].Tables[1].ToSummary()
+	if orders.HasPrimaryKey || orders.ForeignKeyCount != 1 || orders.RelationshipCount != 2 {
+		t.Errorf("orders summary: %+v", orders)
+	}
+}
+
+func TestDirectiveFromAny(t *testing.T) {
+	want := Directive{Namespace: "postgres", Key: "partition", Args: "partition by RANGE (x)", Line: 7}
+	tests := []struct {
+		name string
+		in   any
+		want Directive
+		ok   bool
+	}{
+		{"directive", want, want, true},
+		{"string map int line", map[string]any{"namespace": "postgres", "key": "partition", "args": "partition by RANGE (x)", "line": 7}, want, true},
+		{"string map int64 line", map[string]any{"namespace": "postgres", "key": "partition", "args": "partition by RANGE (x)", "line": int64(7)}, want, true},
+		{"string map float64 line", map[string]any{"namespace": "postgres", "key": "partition", "args": "partition by RANGE (x)", "line": float64(7)}, want, true},
+		{"any map ignores non-string keys", map[any]any{"namespace": "postgres", "key": "partition", "args": "partition by RANGE (x)", "line": 7, 5: "x"}, want, true},
+		{"key derived from args", map[string]any{"namespace": "sqlite", "args": "WITHOUT ROWID extra"}, Directive{Namespace: "sqlite", Key: "without", Args: "WITHOUT ROWID extra"}, true},
+		{"wrong field types ignored", map[string]any{"namespace": 1, "key": 2, "args": 3, "line": "x"}, Directive{}, true},
+		{"unsupported type", "nope", Directive{}, false},
+		{"nil", nil, Directive{}, false},
+		{"int", 5, Directive{}, false},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got, ok := directiveFromAny(tt.in)
+			if ok != tt.ok || got != tt.want {
+				t.Errorf("got (%+v,%v), want (%+v,%v)", got, ok, tt.want, tt.ok)
+			}
+		})
+	}
+}

pkg/readers/bun/README.md

@@ -64,6 +64,8 @@ The reader recognizes the following Bun struct tags:
 - `autoincrement` - Auto-increment column
 - `default` - Default value
 - `unique` - Unique constraint
+- `generated` - Generated column marker (`Generated`; expression is not in the tag)
+- `identity` - Identity column marker (`Identity`, `ALWAYS`)
 - `rel` - Relationship definition
 
 ## Example Bun Model

pkg/readers/bun/generated_column_test.go

@@ -0,0 +1,48 @@
+package bun
+
+import (
+	"os"
+	"path/filepath"
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/readers"
+)
+
+func TestReader_GeneratedAndIdentityMarkers(t *testing.T) {
+	src := `package models
+
+import "github.com/uptrace/bun"
+
+type Person struct {
+	bun.BaseModel ` + "`bun:\"table:people,alias:p\"`" + `
+
+	ID       int64  ` + "`bun:\"id,pk,type:bigint,autoincrement,identity\"`" + `
+	Seq      int64  ` + "`bun:\"seq,type:bigint,scanonly,identity,notnull\"`" + `
+	FullName string ` + "`bun:\"full_name,type:text,scanonly,generated,nullzero\"`" + `
+	First    string ` + "`bun:\"first,type:text,nullzero\"`" + `
+}
+`
+	path := filepath.Join(t.TempDir(), "person.go")
+	if err := os.WriteFile(path, []byte(src), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	db, err := NewReader(&readers.ReaderOptions{FilePath: path}).ReadDatabase()
+	if err != nil {
+		t.Fatalf("ReadDatabase() error = %v", err)
+	}
+	cols := db.Schemas[0].Tables[0].Columns
+
+	if !cols["id"].Identity || cols["id"].IdentityGeneration != "ALWAYS" {
+		t.Errorf("id: identity = %v/%q, want true/ALWAYS", cols["id"].Identity, cols["id"].IdentityGeneration)
+	}
+	if !cols["seq"].Identity || cols["seq"].IdentityGeneration != "ALWAYS" {
+		t.Errorf("seq: identity = %v/%q, want true/ALWAYS", cols["seq"].Identity, cols["seq"].IdentityGeneration)
+	}
+	if !cols["full_name"].Generated {
+		t.Error("full_name should be generated")
+	}
+	if cols["first"].Generated || cols["first"].Identity {
+		t.Error("first must be an ordinary column")
+	}
+}

pkg/readers/bun/helpers_test.go

@@ -0,0 +1,117 @@
+package bun
+
+import (
+	"go/ast"
+	"go/parser"
+	"go/token"
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/readers"
+)
+
+func newTestReader() *Reader { return NewReader(&readers.ReaderOptions{}) }
+
+func mustExpr(t *testing.T, src string) ast.Expr {
+	t.Helper()
+	e, err := parser.ParseExpr(src)
+	if err != nil {
+		t.Fatal(err)
+	}
+	return e
+}
+
+func TestGoTypeToSQL(t *testing.T) {
+	r := newTestReader()
+	tests := []struct{ src, want string }{
+		{"int", "integer"},
+		{"int32", "integer"},
+		{"int64", "bigint"},
+		{"string", "text"},
+		{"bool", "boolean"},
+		{"float32", "real"},
+		{"float64", "double precision"},
+		{"uint8", "text"},
+		{"time.Time", "timestamp"},
+		{"time.Duration", "text"},
+		{"sql_types.SqlString", "text"},
+		{"sql_types.SqlInt", "integer"},
+		{"sql_types.SqlInt64", "bigint"},
+		{"sql_types.SqlFloat", "double precision"},
+		{"sql_types.SqlBool", "boolean"},
+		{"sql_types.SqlTime", "timestamp"},
+		{"sql_types.Other", "text"},
+		{"other.Thing", "text"},
+		{"*int64", "bigint"},
+		{"*time.Time", "timestamp"},
+		{"[]byte", "text"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.src, func(t *testing.T) {
+			if got := r.goTypeToSQL(mustExpr(t, tt.src)); got != tt.want {
+				t.Errorf("got %q want %q", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestDeriveTableName(t *testing.T) {
+	r := newTestReader()
+	for in, want := range map[string]string{
+		"ModelUser":     "user",
+		"ModelUserRole": "user_role",
+		"Account":       "account",
+		"OrderItem":     "order_item",
+	} {
+		if got := r.deriveTableName(in); got != want {
+			t.Errorf("%q: got %q want %q", in, got, want)
+		}
+	}
+}
+
+func TestGetReceiverType(t *testing.T) {
+	r := newTestReader()
+	for src, want := range map[string]string{"User": "User", "*User": "User", "*pkg.User": "", "[]User": ""} {
+		if got := r.getReceiverType(mustExpr(t, src)); got != want {
+			t.Errorf("%s: got %q want %q", src, got, want)
+		}
+	}
+}
+
+func TestGetRelationType(t *testing.T) {
+	r := newTestReader()
+	for tag, want := range map[string]string{
+		`bun:"rel:has-many,join:id=user_id"`:   "has-many",
+		`bun:"rel:belongs-to,join:user_id=id"`: "belongs-to",
+		`bun:"rel:has-one,join:id=user_id"`:    "has-one",
+		`bun:"rel:many-to-many,join_table:x"`:  "many-to-many",
+		`bun:"rel:unknown"`:                    "",
+		`bun:"id,pk"`:                          "",
+	} {
+		if got := r.getRelationType(tag); got != want {
+			t.Errorf("%s: got %q want %q", tag, got, want)
+		}
+	}
+}
+
+func TestParseTableNameMethod(t *testing.T) {
+	r := newTestReader()
+	parse := func(src string) *ast.FuncDecl {
+		f, err := parser.ParseFile(token.NewFileSet(), "x.go", "package p\n"+src, 0)
+		if err != nil {
+			t.Fatal(err)
+		}
+		return f.Decls[0].(*ast.FuncDecl)
+	}
+	if tbl, sch := r.parseTableNameMethod(parse(`func (User) TableName() string { return "public.users" }`)); tbl != "users" || sch != "public" {
+		t.Errorf("qualified: %q %q", tbl, sch)
+	}
+	if tbl, sch := r.parseTableNameMethod(parse(`func (User) TableName() string { return "users" }`)); tbl != "users" || sch != "public" {
+		t.Errorf("plain: %q %q", tbl, sch)
+	}
+	if tbl, _ := r.parseTableNameMethod(parse(`func (User) TableName() string`)); tbl != "" {
+		t.Errorf("no body: %q", tbl)
+	}
+	if tbl, _ := r.parseTableNameMethod(parse(`func (User) TableName() string { x := 1; _ = x; return foo() }`)); tbl != "" {
+		t.Errorf("non-literal: %q", tbl)
+	}
+}

pkg/readers/bun/reader.go

@@ -659,6 +659,13 @@ func (r *Reader) parseColumn(fieldName string, fieldType ast.Expr, tag string, s
 			hasExplicitNullableMarker = true
 		case "autoincrement":
 			column.AutoIncrement = true
+		case "generated":
+			// GENERATED ... STORED marker written by the Bun writer; the expression is not in the tag
+			column.Generated = true
+		case "identity":
+			// GENERATED ALWAYS AS IDENTITY marker written by the Bun writer
+			column.Identity = true
+			column.IdentityGeneration = "ALWAYS"
 		case "default":
 			// Default value from Bun tag (e.g., default:gen_random_uuid())
 			column.Default = value

pkg/readers/dbml/README.md

@@ -90,11 +90,28 @@ Ref: posts.user_id > users.id [delete: cascade]
 - Default values (`default`)
 - Inline references (`ref`)
 - Standalone `Ref` blocks
+- Commented cross-file refs (`// Ref:` — see below)
 - Indexes and composite indexes
 - Table notes and column notes
 - Enums
 - Dialect directives (`@postgres:` / `@sqlite:` — see below)
 
+## Commented cross-file refs
+
+`// Ref:` / `// ref:` lines (ignored by dbdiagram) become FKs + relationships once both ends are loaded.
+
+| Rule | Behaviour |
+|---|---|
+| When | Single file / directory: end of read. `--from-list`, `merge`, jobs: after all inputs are combined |
+| Match | `schema.table.column` on both sides, case-insensitive |
+| Operators | `>`, `<`, `-` (parsed like `Ref:`) |
+| Duplicate of an FK on the same columns | Skipped silently |
+| Target not loaded | Kept pending; skipped with a warning on the final pass |
+| Column type mismatch | Warning, FK still created (`serial`≈`integer`, `bigserial`≈`bigint`) |
+| Pending state | `Database.Metadata["dbml.commented_refs"]` (`[]string`) |
+
+API: `dbml.ResolveCommentedRefs(db, final bool) []string` returns warnings.
+
 ## Dialect directives
 
 Lines of the form `@<namespace>[(<column>)]: <args>` embed database-specific
@@ -140,6 +157,7 @@ grammar and the supported-directive matrix.
 
 ## Notes
 
+- Column notes `GENERATED ALWAYS AS (expr) STORED` and `GENERATED ALWAYS|BY DEFAULT AS IDENTITY` set `Generated`/`GenerationExpression` and `Identity`/`IdentityGeneration`; they are not kept as comments
 - DBML is designed for database documentation and diagramming
 - Schema name defaults to `public`
 - Relationship cardinality is preserved

pkg/readers/dbml/commented_refs.go

@@ -0,0 +1,218 @@
+package dbml
+
+import (
+	"fmt"
+	"regexp"
+	"strings"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+)
+
+// CommentedRefsMetadataKey is the Database.Metadata key holding commented
+// `// Ref:` lines not yet resolved against the model ([]string).
+const CommentedRefsMetadataKey = "dbml.commented_refs"
+
+// commentedRefRegex matches `// Ref: ...` and `// ref: ...`.
+var commentedRefRegex = regexp.MustCompile(`^//\s*[Rr]ef\s*:\s*(.+)$`)
+
+// commentedRef returns the ref body of a trimmed `// Ref:` comment line.
+func commentedRef(line string) (string, bool) {
+	m := commentedRefRegex.FindStringSubmatch(line)
+	if m == nil {
+		return "", false
+	}
+	ref := strings.TrimSpace(m[1])
+	return ref, ref != ""
+}
+
+// PendingCommentedRefs returns the commented refs not yet resolved.
+func PendingCommentedRefs(db *models.Database) []string {
+	if db == nil || db.Metadata == nil {
+		return nil
+	}
+	refs, _ := db.Metadata[CommentedRefsMetadataKey].([]string)
+	return refs
+}
+
+func setPendingCommentedRefs(db *models.Database, refs []string) {
+	if len(refs) == 0 {
+		if db.Metadata != nil {
+			delete(db.Metadata, CommentedRefsMetadataKey)
+		}
+		return
+	}
+	if db.Metadata == nil {
+		db.Metadata = make(map[string]any)
+	}
+	db.Metadata[CommentedRefsMetadataKey] = refs
+}
+
+// addPendingCommentedRef queues a ref, skipping exact repeats.
+func addPendingCommentedRef(db *models.Database, ref string) {
+	refs := PendingCommentedRefs(db)
+	for _, existing := range refs {
+		if existing == ref {
+			return
+		}
+	}
+	setPendingCommentedRefs(db, append(refs, ref))
+}
+
+// ResolveCommentedRefs turns pending commented refs into foreign keys and
+// relationships when both ends (schema.table.column) exist in db. A ref that
+// matches an existing FK on the same columns is dropped as a duplicate.
+// Unmatched refs stay pending unless final is set, in which case they are
+// dropped with a warning. Returns human-readable warnings.
+func ResolveCommentedRefs(db *models.Database, final bool) []string {
+	refs := PendingCommentedRefs(db)
+	if len(refs) == 0 {
+		return nil
+	}
+
+	var warnings []string
+	var pending []string
+	parser := &Reader{}
+
+	for _, ref := range refs {
+		fk := parser.parseRef(ref)
+		if fk == nil || len(fk.Columns) == 0 || len(fk.Columns) != len(fk.ReferencedColumns) {
+			warnings = append(warnings, fmt.Sprintf("skipping commented ref %q: cannot parse", ref))
+			continue
+		}
+
+		srcTable, srcCols, srcMissing := lookupColumns(db, fk.Schema, fk.Table, fk.Columns)
+		dstTable, dstCols, dstMissing := lookupColumns(db, fk.ReferencedSchema, fk.ReferencedTable, fk.ReferencedColumns)
+		if srcMissing != "" || dstMissing != "" {
+			if final {
+				missing := srcMissing
+				if missing == "" {
+					missing = dstMissing
+				}
+				warnings = append(warnings, fmt.Sprintf("skipping commented ref %q: %s not found", ref, missing))
+			} else {
+				pending = append(pending, ref)
+			}
+			continue
+		}
+
+		fk.Schema, fk.Table, fk.Columns = srcTable.Schema, srcTable.Name, columnNames(srcCols)
+		fk.ReferencedSchema, fk.ReferencedTable, fk.ReferencedColumns = dstTable.Schema, dstTable.Name, columnNames(dstCols)
+
+		if hasFKOnColumns(srcTable, fk.Columns) {
+			continue // already declared by an uncommented Ref or inline ref
+		}
+		if _, taken := srcTable.Constraints[fk.Name]; taken {
+			warnings = append(warnings, fmt.Sprintf("skipping commented ref %q: constraint %s already exists", ref, fk.Name))
+			continue
+		}
+
+		for i := range srcCols {
+			if !compatibleFKTypes(srcCols[i].Type, dstCols[i].Type) {
+				warnings = append(warnings, fmt.Sprintf("commented ref %q: type mismatch %s.%s.%s (%s) -> %s.%s.%s (%s)",
+					ref, srcTable.Schema, srcTable.Name, srcCols[i].Name, srcCols[i].Type,
+					dstTable.Schema, dstTable.Name, dstCols[i].Name, dstCols[i].Type))
+			}
+		}
+
+		if srcTable.Constraints == nil {
+			srcTable.Constraints = make(map[string]*models.Constraint)
+		}
+		srcTable.Constraints[fk.Name] = fk
+		addFKRelationship(srcTable, fk)
+	}
+
+	setPendingCommentedRefs(db, pending)
+	return warnings
+}
+
+// lookupColumns finds a table and its columns, case-insensitively. missing
+// names the first object not found, or is empty.
+func lookupColumns(db *models.Database, schemaName, tableName string, cols []string) (*models.Table, []*models.Column, string) {
+	qualified := schemaName + "." + tableName
+	var table *models.Table
+	for _, schema := range db.Schemas {
+		if !strings.EqualFold(schema.Name, schemaName) {
+			continue
+		}
+		for _, t := range schema.Tables {
+			if strings.EqualFold(t.Name, tableName) {
+				table = t
+				break
+			}
+		}
+	}
+	if table == nil {
+		return nil, nil, "table " + qualified
+	}
+
+	found := make([]*models.Column, 0, len(cols))
+	for _, name := range cols {
+		col := table.Columns[name]
+		if col == nil {
+			for _, c := range table.Columns {
+				if strings.EqualFold(c.Name, name) {
+					col = c
+					break
+				}
+			}
+		}
+		if col == nil {
+			return nil, nil, "column " + qualified + "." + name
+		}
+		found = append(found, col)
+	}
+	return table, found, ""
+}
+
+func columnNames(cols []*models.Column) []string {
+	names := make([]string, len(cols))
+	for i, c := range cols {
+		names[i] = c.Name
+	}
+	return names
+}
+
+// hasFKOnColumns reports whether table already has a foreign key over cols.
+func hasFKOnColumns(table *models.Table, cols []string) bool {
+	for _, c := range table.Constraints {
+		if c.Type != models.ForeignKeyConstraint || len(c.Columns) != len(cols) {
+			continue
+		}
+		same := true
+		for i := range cols {
+			if !strings.EqualFold(c.Columns[i], cols[i]) {
+				same = false
+				break
+			}
+		}
+		if same {
+			return true
+		}
+	}
+	return false
+}
+
+// fkTypeAliases maps serial and alias spellings to their storage type.
+var fkTypeAliases = map[string]string{
+	"smallserial": "smallint", "serial2": "smallint", "int2": "smallint",
+	"serial": "integer", "serial4": "integer", "int": "integer", "int4": "integer",
+	"bigserial": "bigint", "serial8": "bigint", "int8": "bigint",
+}
+
+// compatibleFKTypes compares column types ignoring case, length and serial
+// vs. integer spelling. Unknown (empty) types are treated as compatible.
+func compatibleFKTypes(a, b string) bool {
+	na, nb := normalizeFKType(a), normalizeFKType(b)
+	return na == "" || nb == "" || na == nb
+}
+
+func normalizeFKType(t string) string {
+	t = strings.ToLower(strings.TrimSpace(t))
+	if i := strings.Index(t, "("); i >= 0 {
+		t = strings.TrimSpace(t[:i])
+	}
+	if alias, ok := fkTypeAliases[t]; ok {
+		return alias
+	}
+	return t
+}

pkg/readers/dbml/commented_refs_test.go

@@ -0,0 +1,260 @@
+package dbml
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+	"git.warky.dev/wdevs/relspecgo/pkg/readers"
+)
+
+func readDBMLString(t *testing.T, content string) *models.Database {
+	t.Helper()
+	path := filepath.Join(t.TempDir(), "in.dbml")
+	if err := os.WriteFile(path, []byte(content), 0o644); err != nil {
+		t.Fatalf("write fixture: %v", err)
+	}
+	db, err := NewReader(&readers.ReaderOptions{FilePath: path}).ReadDatabase()
+	if err != nil {
+		t.Fatalf("ReadDatabase() error = %v", err)
+	}
+	return db
+}
+
+func findTable(db *models.Database, schema, table string) *models.Table {
+	for _, s := range db.Schemas {
+		if s.Name != schema {
+			continue
+		}
+		for _, t := range s.Tables {
+			if t.Name == table {
+				return t
+			}
+		}
+	}
+	return nil
+}
+
+func fkOn(table *models.Table, col string) *models.Constraint {
+	for _, c := range table.Constraints {
+		if c.Type == models.ForeignKeyConstraint && len(c.Columns) == 1 && c.Columns[0] == col {
+			return c
+		}
+	}
+	return nil
+}
+
+func relFor(table *models.Table, fkName string) *models.Relationship {
+	for _, r := range table.Relationships {
+		if r.ForeignKey == fkName {
+			return r
+		}
+	}
+	return nil
+}
+
+func TestCommentedRef(t *testing.T) {
+	tests := []struct {
+		line string
+		want string
+		ok   bool
+	}{
+		{`// Ref: a.b.c > d.e.f`, `a.b.c > d.e.f`, true},
+		{`// ref: a.b.c - d.e.f`, `a.b.c - d.e.f`, true},
+		{`//Ref:a.b.c > d.e.f`, `a.b.c > d.e.f`, true},
+		{`// Reference notes`, "", false},
+		{`// see Ref: a.b.c > d.e.f`, "", false},
+		{`// Ref:`, "", false},
+	}
+	for _, tt := range tests {
+		got, ok := commentedRef(tt.line)
+		if ok != tt.ok || got != tt.want {
+			t.Errorf("commentedRef(%q) = (%q, %v), want (%q, %v)", tt.line, got, ok, tt.want, tt.ok)
+		}
+	}
+}
+
+// A commented ref whose tables are in the same file resolves on read.
+func TestReader_CommentedRefSameFile(t *testing.T) {
+	db := readDBMLString(t, `Table "org"."department" {
+  "id_department" bigserial [pk]
+}
+Table "entity"."employee" {
+  "id_employee" bigserial [pk]
+  "rid_department" bigint
+}
+// Ref: "entity"."employee"."rid_department" > "org"."department"."id_department" [delete: restrict, update: restrict]
+`)
+	emp := findTable(db, "entity", "employee")
+	fk := fkOn(emp, "rid_department")
+	if fk == nil {
+		t.Fatal("expected FK on rid_department")
+	}
+	if fk.ReferencedSchema != "org" || fk.ReferencedTable != "department" || fk.ReferencedColumns[0] != "id_department" {
+		t.Errorf("FK target = %s.%s.%v", fk.ReferencedSchema, fk.ReferencedTable, fk.ReferencedColumns)
+	}
+	if fk.OnDelete != "restrict" || fk.OnUpdate != "restrict" {
+		t.Errorf("FK actions = %q/%q, want restrict/restrict", fk.OnDelete, fk.OnUpdate)
+	}
+	if relFor(emp, fk.Name) == nil {
+		t.Error("expected relationship for FK")
+	}
+	if refs := PendingCommentedRefs(db); len(refs) != 0 {
+		t.Errorf("pending = %v, want none", refs)
+	}
+}
+
+// A cross-file commented ref stays pending after a single-file read.
+func TestReader_CommentedRefCrossFilePending(t *testing.T) {
+	db := readDBMLString(t, `Table "entity"."employee" {
+  "id_employee" bigserial [pk]
+  "rid_department" bigint
+}
+// Ref: "entity"."employee"."rid_department" > "org"."department"."id_department"
+`)
+	if fk := fkOn(findTable(db, "entity", "employee"), "rid_department"); fk != nil {
+		t.Fatal("FK must not resolve without the target table")
+	}
+	if refs := PendingCommentedRefs(db); len(refs) != 1 {
+		t.Fatalf("pending = %v, want 1", refs)
+	}
+}
+
+// Directory reads resolve commented refs after all files are merged.
+func TestReader_CommentedRefDirectory(t *testing.T) {
+	db, err := NewReader(&readers.ReaderOptions{
+		FilePath: filepath.Join("..", "..", "..", "tests", "assets", "dbml", "multifile"),
+	}).ReadDatabase()
+	if err != nil {
+		t.Fatalf("ReadDatabase() error = %v", err)
+	}
+	fk := fkOn(findTable(db, "public", "posts"), "user_id")
+	if fk == nil {
+		t.Fatal("expected FK posts.user_id from 9_refs.dbml commented ref")
+	}
+	if fk.ReferencedTable != "users" || fk.OnDelete != "CASCADE" {
+		t.Errorf("FK = %s ondelete %s, want users ondelete CASCADE", fk.ReferencedTable, fk.OnDelete)
+	}
+}
+
+func crossFileDB(t *testing.T, refs ...string) *models.Database {
+	t.Helper()
+	db := readDBMLString(t, `Table "org"."department" {
+  "id_department" bigserial [pk]
+}
+Table "entity"."employee" {
+  "id_employee" bigserial [pk]
+  "rid_department" bigint
+  "rid_manager" integer
+  "rid_team" bigint
+}
+`)
+	setPendingCommentedRefs(db, refs)
+	return db
+}
+
+func TestResolveCommentedRefs(t *testing.T) {
+	t.Run("lowercase ref and one-to-one", func(t *testing.T) {
+		db := crossFileDB(t,
+			`"entity"."employee"."rid_department" > "org"."department"."id_department"`,
+			`entity.employee.rid_team - org.department.id_department`,
+		)
+		if w := ResolveCommentedRefs(db, true); len(w) != 0 {
+			t.Errorf("warnings = %v, want none", w)
+		}
+		emp := findTable(db, "entity", "employee")
+		for _, col := range []string{"rid_department", "rid_team"} {
+			fk := fkOn(emp, col)
+			if fk == nil {
+				t.Fatalf("expected FK on %s", col)
+			}
+			if relFor(emp, fk.Name) == nil {
+				t.Errorf("expected relationship for %s", fk.Name)
+			}
+		}
+		if len(emp.Relationships) != 2 {
+			t.Errorf("relationships = %d, want 2 (same target must not overwrite)", len(emp.Relationships))
+		}
+	})
+
+	t.Run("type mismatch warns but resolves", func(t *testing.T) {
+		db := crossFileDB(t, `entity.employee.rid_manager > org.department.id_department`)
+		w := ResolveCommentedRefs(db, true)
+		if len(w) != 1 || !strings.Contains(w[0], "type mismatch") {
+			t.Errorf("warnings = %v, want one type mismatch", w)
+		}
+		if fkOn(findTable(db, "entity", "employee"), "rid_manager") == nil {
+			t.Error("expected FK on rid_manager")
+		}
+	})
+
+	t.Run("missing target stays pending until final", func(t *testing.T) {
+		db := crossFileDB(t, `entity.employee.rid_team > hr.team.id_team`)
+		if w := ResolveCommentedRefs(db, false); len(w) != 0 {
+			t.Errorf("non-final warnings = %v, want none", w)
+		}
+		if len(PendingCommentedRefs(db)) != 1 {
+			t.Fatal("ref should stay pending")
+		}
+		w := ResolveCommentedRefs(db, true)
+		if len(w) != 1 || !strings.Contains(w[0], "table hr.team not found") {
+			t.Errorf("final warnings = %v, want missing table", w)
+		}
+		if len(PendingCommentedRefs(db)) != 0 {
+			t.Error("final pass must clear pending refs")
+		}
+	})
+
+	t.Run("missing column", func(t *testing.T) {
+		db := crossFileDB(t, `entity.employee.rid_nope > org.department.id_department`)
+		w := ResolveCommentedRefs(db, true)
+		if len(w) != 1 || !strings.Contains(w[0], "column entity.employee.rid_nope not found") {
+			t.Errorf("warnings = %v, want missing column", w)
+		}
+	})
+
+	t.Run("deduplicates against uncommented ref", func(t *testing.T) {
+		db := readDBMLString(t, `Table "org"."department" {
+  "id_department" bigserial [pk]
+}
+Table "entity"."employee" {
+  "id_employee" bigserial [pk]
+  "rid_department" bigint
+}
+Ref: "entity"."employee"."rid_department" > "org"."department"."id_department"
+// Ref: "entity"."employee"."rid_department" > "org"."department"."id_department"
+`)
+		emp := findTable(db, "entity", "employee")
+		count := 0
+		for _, c := range emp.Constraints {
+			if c.Type == models.ForeignKeyConstraint {
+				count++
+			}
+		}
+		if count != 1 || len(emp.Relationships) != 1 {
+			t.Errorf("FKs = %d, relationships = %d, want 1 and 1", count, len(emp.Relationships))
+		}
+	})
+}
+
+func TestCompatibleFKTypes(t *testing.T) {
+	tests := []struct {
+		a, b string
+		want bool
+	}{
+		{"bigint", "bigserial", true},
+		{"integer", "serial", true},
+		{"INT4", "integer", true},
+		{"varchar(10)", "varchar(20)", true},
+		{"bigint", "serial", false},
+		{"uuid", "bigint", false},
+		{"", "bigint", true},
+	}
+	for _, tt := range tests {
+		if got := compatibleFKTypes(tt.a, tt.b); got != tt.want {
+			t.Errorf("compatibleFKTypes(%q, %q) = %v, want %v", tt.a, tt.b, got, tt.want)
+		}
+	}
+}

pkg/readers/dbml/generated_columns.go

@@ -0,0 +1,38 @@
+package dbml
+
+import (
+	"regexp"
+	"strings"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+)
+
+var (
+	generatedNoteRegex = regexp.MustCompile(`(?is)^GENERATED\s+ALWAYS\s+AS\s*\((.*)\)\s*STORED$`)
+	identityNoteRegex  = regexp.MustCompile(`(?i)^GENERATED\s+(ALWAYS|BY\s+DEFAULT)\s+AS\s+IDENTITY$`)
+)
+
+// applyColumnMarkerNote interprets a column note written by the DBML writer for
+// GENERATED ... STORED and GENERATED ... AS IDENTITY columns, setting the matching
+// model fields. It returns true when the note was a marker, in which case it is not
+// a user comment.
+func applyColumnMarkerNote(column *models.Column, rawNote string) bool {
+	note := strings.TrimSpace(rawNote)
+	if len(note) >= 2 && (note[0] == '\'' || note[0] == '"') && note[len(note)-1] == note[0] {
+		note = note[1 : len(note)-1]
+	}
+	note = strings.ReplaceAll(strings.ReplaceAll(note, `\'`, `'`), `\\`, `\`)
+	note = strings.TrimSpace(note)
+
+	if m := generatedNoteRegex.FindStringSubmatch(note); m != nil {
+		column.Generated = true
+		column.GenerationExpression = strings.TrimSpace(m[1])
+		return true
+	}
+	if m := identityNoteRegex.FindStringSubmatch(note); m != nil {
+		column.Identity = true
+		column.IdentityGeneration = strings.Join(strings.Fields(strings.ToUpper(m[1])), " ")
+		return true
+	}
+	return false
+}

pkg/readers/dbml/reader.go

@@ -48,7 +48,22 @@ func (r *Reader) ReadDatabase() (*models.Database, error) {
 		return nil, fmt.Errorf("failed to read file: %w", err)
 	}
 
-	return r.parseDBML(string(content))
+	db, err := r.parseDBML(string(content))
+	if err != nil {
+		return nil, err
+	}
+	r.resolveCommentedRefs(db)
+	return db, nil
+}
+
+// resolveCommentedRefs resolves the commented refs whose tables are loaded.
+// Unmatched refs stay pending for a later pass over a combined model.
+func (r *Reader) resolveCommentedRefs(db *models.Database) {
+	for _, w := range ResolveCommentedRefs(db, false) {
+		if r.options.Progress != nil {
+			r.options.Progress("warning: " + w)
+		}
+	}
 }
 
 // ReadSchema reads and parses DBML input, returning a Schema model
@@ -125,6 +140,7 @@ func (r *Reader) readDirectoryDBML(dirPath string) (*models.Database, error) {
 		}
 	}
 
+	r.resolveCommentedRefs(db)
 	return db, nil
 }
 
@@ -305,9 +321,39 @@ func sortDBMLFiles(files []string) []string {
 // Merges: Columns (map), Constraints (map), Indexes (map), Relationships (map)
 // Uses first non-empty Description
 func mergeTable(baseTable, fileTable *models.Table) {
-	// Merge columns (map naturally merges - later keys overwrite)
-	for key, col := range fileTable.Columns {
-		baseTable.Columns[key] = col
+	// Merge columns. Each file numbers its own columns from 1, so a table split
+	// across files would otherwise end up with colliding Column.Sequence values
+	// and writers would fall back to alphabetical order. Re-base the incoming
+	// file's new columns after the highest sequence already present, preserving
+	// their in-file order. Columns that overwrite an existing key keep the
+	// original position.
+	var maxSeq uint
+	for _, col := range baseTable.Columns {
+		if col.Sequence > maxSeq {
+			maxSeq = col.Sequence
+		}
+	}
+
+	incoming := make([]*models.Column, 0, len(fileTable.Columns))
+	for _, col := range fileTable.Columns {
+		incoming = append(incoming, col)
+	}
+	sort.Slice(incoming, func(i, j int) bool {
+		if incoming[i].Sequence != incoming[j].Sequence {
+			return incoming[i].Sequence < incoming[j].Sequence
+		}
+		return incoming[i].Name < incoming[j].Name
+	})
+
+	var added uint
+	for _, col := range incoming {
+		if existing, ok := baseTable.Columns[col.Name]; ok {
+			col.Sequence = existing.Sequence
+		} else {
+			added++
+			col.Sequence = maxSeq + added
+		}
+		baseTable.Columns[col.Name] = col
 	}
 
 	// Merge constraints
@@ -410,6 +456,10 @@ func mergeDatabase(baseDB, fileDB *models.Database) {
 	// Merge domains
 	baseDB.Domains = append(baseDB.Domains, fileDB.Domains...)
 
+	for _, ref := range PendingCommentedRefs(fileDB) {
+		addPendingCommentedRef(baseDB, ref)
+	}
+
 	// Use first non-empty description
 	if baseDB.Description == "" && fileDB.Description != "" {
 		baseDB.Description = fileDB.Description
@@ -463,8 +513,12 @@ func (r *Reader) parseDBML(content string) (*models.Database, error) {
 			continue
 		}
 
-		// Skip empty lines and comments
+		// Skip empty lines and comments. A commented `// Ref:` is kept as a
+		// pending cross-file ref, resolved once every file is loaded.
 		if line == "" || strings.HasPrefix(line, "//") {
+			if ref, ok := commentedRef(line); ok {
+				addPendingCommentedRef(db, ref)
+			}
 			continue
 		}
 
@@ -551,7 +605,13 @@ func (r *Reader) parseDBML(content string) (*models.Database, error) {
 
 			index := r.parseIndex(line, currentTable.Name, currentSchema)
 			if index != nil {
-				currentTable.Indexes[index.Name] = index
+				// Keep a reused name under a unique map key so the duplicate is
+				// not silently dropped; the inspector reports it.
+				key := index.Name
+				for n := 2; currentTable.Indexes[key] != nil; n++ {
+					key = fmt.Sprintf("%s#%d", index.Name, n)
+				}
+				currentTable.Indexes[key] = index
 				lastIndex = index
 			}
 			continue
@@ -629,20 +689,12 @@ func (r *Reader) parseDBML(content string) (*models.Database, error) {
 	// for DBML refs so diffing equivalent schemas compares the same model.
 	for _, schema := range schemaMap {
 		for _, table := range schema.Tables {
-			for _, constraint := range table.Constraints {
+			for _, name := range sortedConstraintNames(table.Constraints) {
+				constraint := table.Constraints[name]
 				if constraint.Type != models.ForeignKeyConstraint {
 					continue
 				}
-				name := fmt.Sprintf("%s_to_%s", table.Name, constraint.ReferencedTable)
-				relationship := models.InitRelationship(name, models.OneToMany)
-				relationship.FromTable = table.Name
-				relationship.FromSchema = table.Schema
-				relationship.FromColumns = append([]string(nil), constraint.Columns...)
-				relationship.ToTable = constraint.ReferencedTable
-				relationship.ToSchema = constraint.ReferencedSchema
-				relationship.ToColumns = append([]string(nil), constraint.ReferencedColumns...)
-				relationship.ForeignKey = constraint.Name
-				table.Relationships[name] = relationship
+				addFKRelationship(table, constraint)
 			}
 		}
 	}
@@ -655,6 +707,39 @@ func (r *Reader) parseDBML(content string) (*models.Database, error) {
 	return db, nil
 }
 
+// sortedConstraintNames returns constraint keys in sorted order so derived
+// relationship names do not depend on map iteration order.
+func sortedConstraintNames(constraints map[string]*models.Constraint) []string {
+	names := make([]string, 0, len(constraints))
+	for name := range constraints {
+		names = append(names, name)
+	}
+	sort.Strings(names)
+	return names
+}
+
+// addFKRelationship derives the relationship for a foreign key, matching how
+// the PostgreSQL reader models FKs.
+func addFKRelationship(table *models.Table, constraint *models.Constraint) {
+	name := fmt.Sprintf("%s_to_%s", table.Name, constraint.ReferencedTable)
+	if existing, taken := table.Relationships[name]; taken && existing.ForeignKey != constraint.Name {
+		// A second FK to the same table must not overwrite the first.
+		name = fmt.Sprintf("%s_%s", name, strings.Join(constraint.Columns, "_"))
+	}
+	relationship := models.InitRelationship(name, models.OneToMany)
+	relationship.FromTable = table.Name
+	relationship.FromSchema = table.Schema
+	relationship.FromColumns = append([]string(nil), constraint.Columns...)
+	relationship.ToTable = constraint.ReferencedTable
+	relationship.ToSchema = constraint.ReferencedSchema
+	relationship.ToColumns = append([]string(nil), constraint.ReferencedColumns...)
+	relationship.ForeignKey = constraint.Name
+	if table.Relationships == nil {
+		table.Relationships = make(map[string]*models.Relationship)
+	}
+	table.Relationships[name] = relationship
+}
+
 // setTableNote preserves multiple table notes. The first maps to Description
 // and the second to Comment, matching the model fields used by code writers.
 func setTableNote(table *models.Table, note string) {
@@ -719,7 +804,9 @@ func (r *Reader) parseColumn(line, tableName, schemaName string) (*models.Column
 			} else if strings.HasPrefix(attr, "note:") {
 				// Parse column note/comment
 				note := strings.TrimSpace(strings.TrimPrefix(attr, "note:"))
-				column.Comment = strings.Trim(note, "'\"")
+				if !applyColumnMarkerNote(column, note) {
+					column.Comment = strings.Trim(note, "'\"")
+				}
 			} else if strings.HasPrefix(attr, "ref:") {
 				// Parse inline reference
 				// DBML semantics depend on context:

pkg/readers/dbml/reader_test.go

@@ -652,6 +652,22 @@ func TestReadDirectory_TableMerging(t *testing.T) {
 	if emailCol.Type != "varchar(255)" {
 		t.Errorf("Expected email type 'varchar(255)', got '%s'", emailCol.Type)
 	}
+
+	// Merged columns must keep declaration order (file 1: id, email; file 3:
+	// name, created_at) via strictly increasing, non-colliding Sequence values
+	// so downstream writers do not fall back to alphabetical order.
+	order := []string{"id", "email", "name", "created_at"}
+	var prev uint
+	for i, name := range order {
+		col := usersTable.Columns[name]
+		if col.Sequence == 0 {
+			t.Fatalf("column %q has zero Sequence after merge", name)
+		}
+		if i > 0 && col.Sequence <= prev {
+			t.Errorf("column %q Sequence %d not greater than previous %d (order not preserved across files)", name, col.Sequence, prev)
+		}
+		prev = col.Sequence
+	}
 }
 
 func TestReadDirectory_CommentedRefsLast(t *testing.T) {
@@ -1069,3 +1085,39 @@ func TestReader_MultilineTableNote(t *testing.T) {
 		t.Errorf("column note = %q, want %q", got, want)
 	}
 }
+
+// A name reused inside one Indexes block must not silently drop the earlier
+// index; both are kept so the inspector can report the duplicate.
+func TestReader_DuplicateIndexNameInTableKept(t *testing.T) {
+	dbmlContent := `Table individual_actor {
+  id bigint [pk]
+  rid_actor bigint
+  kind text
+
+  Indexes {
+    (rid_actor) [name: 'idx_actor', unique]
+    (rid_actor, kind) [name: 'idx_actor']
+  }
+}
+`
+	dir := t.TempDir()
+	path := filepath.Join(dir, "dup_index.dbml")
+	if err := os.WriteFile(path, []byte(dbmlContent), 0o644); err != nil {
+		t.Fatalf("failed to write fixture: %v", err)
+	}
+
+	db, err := NewReader(&readers.ReaderOptions{FilePath: path}).ReadDatabase()
+	if err != nil {
+		t.Fatalf("ReadDatabase() error = %v", err)
+	}
+
+	table := db.Schemas[0].Tables[0]
+	if len(table.Indexes) != 2 {
+		t.Fatalf("expected 2 indexes, got %d: %v", len(table.Indexes), table.Indexes)
+	}
+	for key, idx := range table.Indexes {
+		if idx.Name != "idx_actor" {
+			t.Errorf("index %q: Name = %q, want idx_actor", key, idx.Name)
+		}
+	}
+}

pkg/readers/drizzle/README.md

@@ -85,6 +85,7 @@ export const postsRelations = relations(posts, ({ one }) => ({
 
 ## Notes
 
+- `.generatedAlwaysAs(sql`expr`)` sets `Generated` + `GenerationExpression`; `.generatedAlwaysAsIdentity()` / `.generatedByDefaultAsIdentity()` set `Identity` + `IdentityGeneration`
 - Supports both PostgreSQL and MySQL Drizzle schemas
 - Extracts relationship information from `relations` definitions
 - Schema defaults to `public` for PostgreSQL

pkg/readers/drizzle/reader.go

@@ -15,6 +15,9 @@ import (
 // Reader implements the readers.Reader interface for Drizzle schema format
 type Reader struct {
 	options *readers.ReaderOptions
+	// enumVars maps the constant a pgEnum() is assigned to (e.g. "role") to the
+	// enum's SQL name (e.g. "Role"), so columns declared as role('col') resolve.
+	enumVars map[string]string
 }
 
 // NewReader creates a new Drizzle reader with the given options
@@ -29,6 +32,7 @@ func (r *Reader) ReadDatabase() (*models.Database, error) {
 	if r.options.FilePath == "" {
 		return nil, fmt.Errorf("file path is required for Drizzle reader")
 	}
+	r.enumVars = make(map[string]string)
 
 	// Check if it's a file or directory
 	info, err := os.Stat(r.options.FilePath)
@@ -100,6 +104,13 @@ func (r *Reader) readDirectory(dirPath string) (*models.Database, error) {
 		return nil, fmt.Errorf("failed to glob directory: %w", err)
 	}
 
+	// Enums may be declared in a different file than the tables using them
+	for _, file := range files {
+		if content, err := os.ReadFile(file); err == nil {
+			r.collectEnumVars(string(content))
+		}
+	}
+
 	// Parse each file
 	for _, file := range files {
 		content, err := os.ReadFile(file)
@@ -125,9 +136,22 @@ func (r *Reader) readDirectory(dirPath string) (*models.Database, error) {
 	return db, nil
 }
 
+var enumVarRegex = regexp.MustCompile(`export\s+const\s+(\w+)\s*=\s*pgEnum\s*\(\s*['"](\w+)['"]`)
+
+// collectEnumVars records every pgEnum() constant declared in content.
+func (r *Reader) collectEnumVars(content string) {
+	if r.enumVars == nil {
+		r.enumVars = make(map[string]string)
+	}
+	for _, m := range enumVarRegex.FindAllStringSubmatch(content, -1) {
+		r.enumVars[m[1]] = m[2]
+	}
+}
+
 // parseDrizzle parses Drizzle schema content and returns a Database model
 func (r *Reader) parseDrizzle(content string) (*models.Database, error) {
 	db := models.InitDatabase("database")
+	r.collectEnumVars(content)
 
 	if r.options.Metadata != nil {
 		if name, ok := r.options.Metadata["name"].(string); ok {
@@ -375,6 +399,9 @@ func (r *Reader) parseColumnDefinition(line, fieldName, drizzleType string, tabl
 
 	// Map Drizzle type to SQL type
 	column.Type = r.drizzleTypeToSQL(drizzleType)
+	if enumName, ok := r.enumVars[drizzleType]; ok {
+		column.Type = enumName
+	}
 
 	// Default: columns are nullable unless specified
 	column.NotNull = false
@@ -495,9 +522,24 @@ func (r *Reader) parseColumnModifiers(line string, column *models.Column, table
 		}
 	}
 
-	// Check for .generatedAlwaysAsIdentity()
+	// Check for .generatedAlwaysAsIdentity() / .generatedByDefaultAsIdentity()
 	if strings.Contains(line, ".generatedAlwaysAsIdentity()") {
 		column.AutoIncrement = true
+		column.Identity = true
+		column.IdentityGeneration = "ALWAYS"
+	}
+	if strings.Contains(line, ".generatedByDefaultAsIdentity()") {
+		column.AutoIncrement = true
+		column.Identity = true
+		column.IdentityGeneration = "BY DEFAULT"
+	}
+
+	// Check for .generatedAlwaysAs(sql`expr`) (generated column)
+	if idx := strings.Index(line, ".generatedAlwaysAs("); idx != -1 {
+		if expr, ok := parseGeneratedAlwaysAs(line[idx+len(".generatedAlwaysAs("):]); ok {
+			column.Generated = true
+			column.GenerationExpression = expr
+		}
 	}
 
 	// Check for .references(() => otherTable.column)
@@ -615,3 +657,34 @@ func (r *Reader) varNameToTableName(varName string) string {
 	// For now, assume variable name matches table name
 	return varName
 }
+
+// parseGeneratedAlwaysAs extracts the SQL expression from the argument list of
+// generatedAlwaysAs(...), which follows the opening parenthesis in rest. The expression is
+// the contents of a sql`...` template (backticks inside it are escaped with a backslash)
+// or, failing that, a quoted string.
+func parseGeneratedAlwaysAs(rest string) (string, bool) {
+	rest = strings.TrimSpace(rest)
+	rest = strings.TrimPrefix(rest, "sql")
+	if rest == "" {
+		return "", false
+	}
+	quote := rest[0]
+	if quote != '`' && quote != '\'' && quote != '"' {
+		return "", false
+	}
+
+	var sb strings.Builder
+	for i := 1; i < len(rest); i++ {
+		ch := rest[i]
+		if ch == '\\' && i+1 < len(rest) {
+			i++
+			sb.WriteByte(rest[i])
+			continue
+		}
+		if ch == quote {
+			return strings.TrimSpace(sb.String()), true
+		}
+		sb.WriteByte(ch)
+	}
+	return "", false
+}

pkg/readers/drizzle/reader_test.go

@@ -0,0 +1,114 @@
+package drizzle
+
+import (
+	"os"
+	"path/filepath"
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+	"git.warky.dev/wdevs/relspecgo/pkg/readers"
+)
+
+const fixture = "../../../tests/assets/drizzle/schema.ts"
+
+func readFile(t *testing.T, path string) *models.Database {
+	t.Helper()
+	db, err := NewReader(&readers.ReaderOptions{FilePath: path}).ReadDatabase()
+	if err != nil {
+		t.Fatal(err)
+	}
+	return db
+}
+
+func findTable(db *models.Database, name string) *models.Table {
+	for _, s := range db.Schemas {
+		for _, tb := range s.Tables {
+			if tb.Name == name {
+				return tb
+			}
+		}
+	}
+	return nil
+}
+
+func TestReadFixture(t *testing.T) {
+	db := readFile(t, fixture)
+	if len(db.Schemas) == 0 || len(db.Schemas[0].Tables) == 0 {
+		t.Fatal("expected tables")
+	}
+	if len(db.Schemas[0].Enums) != 1 || db.Schemas[0].Enums[0].Name != "Role" {
+		t.Fatalf("enums = %+v", db.Schemas[0].Enums)
+	}
+	var found bool
+	for _, tb := range db.Schemas[0].Tables {
+		if c, ok := tb.Columns["role"]; ok {
+			found = true
+			if c.Type != "Role" {
+				t.Errorf("role type = %q", c.Type)
+			}
+		}
+		for n := range tb.Columns {
+			if n == "profile" {
+				t.Errorf("relation field leaked as column in %s", tb.Name)
+			}
+		}
+	}
+	if !found {
+		t.Error("no role column")
+	}
+}
+
+func TestEnumColumnSyntax(t *testing.T) {
+	tests := []struct {
+		name string
+		src  string
+	}{
+		{"enum constant", "export const role = pgEnum('Role', ['A','B']);\nexport const users = pgTable('users', {\n  role: role('role').notNull(),\n});\n"},
+		{"legacy", "export const role = pgEnum('Role', ['A','B']);\nexport const users = pgTable('users', {\n  role: pgEnum('Role')('role').notNull(),\n});\n"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			p := filepath.Join(t.TempDir(), "s.ts")
+			if err := os.WriteFile(p, []byte(tt.src), 0o644); err != nil {
+				t.Fatal(err)
+			}
+			tb := findTable(readFile(t, p), "users")
+			if tb == nil {
+				t.Fatal("users missing")
+			}
+			c := tb.Columns["role"]
+			if c == nil || c.Type != "Role" || !c.NotNull {
+				t.Errorf("column = %+v", c)
+			}
+		})
+	}
+}
+
+func TestReadDirectorySeparateEnums(t *testing.T) {
+	dir := t.TempDir()
+	files := map[string]string{
+		"enums.ts":  "export const status = pgEnum('Status', ['on','off']);\n",
+		"tables.ts": "export const items = pgTable('items', {\n  status: status('status'),\n});\n",
+	}
+	for n, c := range files {
+		if err := os.WriteFile(filepath.Join(dir, n), []byte(c), 0o644); err != nil {
+			t.Fatal(err)
+		}
+	}
+	tb := findTable(readFile(t, dir), "items")
+	if tb == nil {
+		t.Fatal("items missing")
+	}
+	if c := tb.Columns["status"]; c == nil || c.Type != "Status" {
+		t.Errorf("column = %+v", c)
+	}
+}
+
+func TestReaderErrors(t *testing.T) {
+	if _, err := NewReader(&readers.ReaderOptions{}).ReadDatabase(); err == nil {
+		t.Error("expected error for empty path")
+	}
+	if _, err := NewReader(&readers.ReaderOptions{FilePath: "/nonexistent.ts"}).ReadDatabase(); err == nil {
+		t.Error("expected error for missing file")
+	}
+}

pkg/readers/gorm/README.md

@@ -78,6 +78,8 @@ The reader recognizes the following GORM struct tags:
 - `not null` - NOT NULL constraint
 - `autoIncrement` - Auto-increment column
 - `default` - Default value
+- `generated` - Generated column marker (`Generated`; expression is not in the tag)
+- `identity` - Identity column marker (`Identity`, `ALWAYS`)
 - `size` - Column size/length
 - `index` - Create index
 - `uniqueIndex` - Create unique index

pkg/readers/gorm/generated_column_test.go

@@ -0,0 +1,48 @@
+package gorm
+
+import (
+	"os"
+	"path/filepath"
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/readers"
+)
+
+func TestReader_GeneratedAndIdentityMarkers(t *testing.T) {
+	src := `package models
+
+type Person struct {
+	ID       int64  ` + "`gorm:\"column:id;primaryKey;autoIncrement;type:bigint;identity\"`" + `
+	Seq      int64  ` + "`gorm:\"column:seq;type:bigint;<-:false;identity;not null\"`" + `
+	FullName string ` + "`gorm:\"column:full_name;type:text;<-:false;generated\"`" + `
+	First    string ` + "`gorm:\"column:first;type:text\"`" + `
+}
+
+func (Person) TableName() string {
+	return "people"
+}
+`
+	path := filepath.Join(t.TempDir(), "person.go")
+	if err := os.WriteFile(path, []byte(src), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	db, err := NewReader(&readers.ReaderOptions{FilePath: path}).ReadDatabase()
+	if err != nil {
+		t.Fatalf("ReadDatabase() error = %v", err)
+	}
+	cols := db.Schemas[0].Tables[0].Columns
+
+	if !cols["id"].Identity || cols["id"].IdentityGeneration != "ALWAYS" {
+		t.Errorf("id: identity = %v/%q, want true/ALWAYS", cols["id"].Identity, cols["id"].IdentityGeneration)
+	}
+	if !cols["seq"].Identity {
+		t.Error("seq should be identity")
+	}
+	if !cols["full_name"].Generated {
+		t.Error("full_name should be generated")
+	}
+	if cols["first"].Generated || cols["first"].Identity {
+		t.Error("first must be an ordinary column")
+	}
+}

pkg/readers/gorm/helpers_test.go

@@ -0,0 +1,164 @@
+package gorm
+
+import (
+	"go/ast"
+	"go/parser"
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+	"git.warky.dev/wdevs/relspecgo/pkg/readers"
+)
+
+func newTestReader() *Reader { return NewReader(&readers.ReaderOptions{}) }
+
+func mustExpr(t *testing.T, src string) ast.Expr {
+	t.Helper()
+	e, err := parser.ParseExpr(src)
+	if err != nil {
+		t.Fatal(err)
+	}
+	return e
+}
+
+func TestGoTypeToSQL(t *testing.T) {
+	r := newTestReader()
+	tests := []struct{ src, want string }{
+		{"int", "integer"},
+		{"int32", "integer"},
+		{"int64", "bigint"},
+		{"string", "text"},
+		{"bool", "boolean"},
+		{"float32", "real"},
+		{"float64", "double precision"},
+		{"uint8", "text"},
+		{"time.Time", "timestamp"},
+		{"time.Duration", "text"},
+		{"sql_types.SqlString", "text"},
+		{"sql_types.SqlInt", "integer"},
+		{"sql_types.SqlInt64", "bigint"},
+		{"sql_types.SqlFloat", "double precision"},
+		{"sql_types.SqlBool", "boolean"},
+		{"sql_types.SqlTime", "timestamp"},
+		{"sql_types.Other", "text"},
+		{"other.Thing", "text"},
+		{"*int64", "bigint"},
+		{"*time.Time", "timestamp"},
+		{"[]byte", "text"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.src, func(t *testing.T) {
+			if got := r.goTypeToSQL(mustExpr(t, tt.src)); got != tt.want {
+				t.Errorf("got %q want %q", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestFieldNameToColumnName(t *testing.T) {
+	r := newTestReader()
+	for in, want := range map[string]string{"ID": "i_d", "UserName": "user_name", "name": "name", "": ""} {
+		if got := r.fieldNameToColumnName(in); got != want {
+			t.Errorf("%q: got %q want %q", in, got, want)
+		}
+	}
+}
+
+func TestGetReceiverType(t *testing.T) {
+	r := newTestReader()
+	tests := []struct{ src, want string }{
+		{"User", "User"}, {"*User", "User"}, {"*pkg.User", ""}, {"[]User", ""},
+	}
+	for _, tt := range tests {
+		if got := r.getReceiverType(mustExpr(t, tt.src)); got != tt.want {
+			t.Errorf("%s: got %q want %q", tt.src, got, tt.want)
+		}
+	}
+}
+
+func TestIsGORMModel(t *testing.T) {
+	r := newTestReader()
+	tests := []struct {
+		name  string
+		field *ast.Field
+		want  bool
+	}{
+		{"embedded gorm.Model", &ast.Field{Type: mustExpr(t, "gorm.Model")}, true},
+		{"named field", &ast.Field{Names: []*ast.Ident{ast.NewIdent("M")}, Type: mustExpr(t, "gorm.Model")}, false},
+		{"plain ident", &ast.Field{Type: mustExpr(t, "Model")}, false},
+		{"other package", &ast.Field{Type: mustExpr(t, "other.Model")}, false},
+		{"gorm other", &ast.Field{Type: mustExpr(t, "gorm.DB")}, false},
+		{"non-ident selector base", &ast.Field{Type: mustExpr(t, "a.b.Model")}, false},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := r.isGORMModel(tt.field); got != tt.want {
+				t.Errorf("got %v want %v", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestParseTypeWithReferences(t *testing.T) {
+	r := newTestReader()
+	tests := []struct {
+		in      string
+		base    string
+		length  int
+		refInfo string
+	}{
+		{"bigint", "bigint", 0, ""},
+		{"varchar(50)", "varchar", 50, ""},
+		{"bigint references mainaccount(id) ON DELETE CASCADE", "bigint", 0, "mainaccount(id) ON DELETE CASCADE"},
+		{"varchar(20) REFERENCES t(c)", "varchar", 20, "t(c)"},
+	}
+	for _, tt := range tests {
+		base, length, ref := r.parseTypeWithReferences(tt.in)
+		if base != tt.base || length != tt.length || ref != tt.refInfo {
+			t.Errorf("%q: got (%q,%d,%q)", tt.in, base, length, ref)
+		}
+	}
+}
+
+func TestCreateInlineReferenceConstraint(t *testing.T) {
+	tests := []struct {
+		name     string
+		ref      string
+		wantNone bool
+		schema   string
+		table    string
+		col      string
+		onDelete string
+		onUpdate string
+	}{
+		{"simple", "accounts(id)", false, "public", "accounts", "id", "NO ACTION", "NO ACTION"},
+		{"schema qualified", "billing.accounts(id)", false, "billing", "accounts", "id", "NO ACTION", "NO ACTION"},
+		{"cascade restrict", "accounts(id) ON DELETE CASCADE ON UPDATE RESTRICT", false, "public", "accounts", "id", "CASCADE", "RESTRICT"},
+		{"set null no action", "accounts(id) on delete set null on update no action", false, "public", "accounts", "id", "SET NULL", "NO ACTION"},
+		{"restrict delete cascade update", "accounts(id) ON DELETE RESTRICT ON UPDATE CASCADE", false, "public", "accounts", "id", "RESTRICT", "CASCADE"},
+		{"update set null", "accounts(id) ON DELETE NO ACTION ON UPDATE SET NULL", false, "public", "accounts", "id", "NO ACTION", "SET NULL"},
+		{"no parens", "accounts", true, "", "", "", "", ""},
+		{"reversed parens", "accounts)id(", true, "", "", "", "", ""},
+	}
+	r := newTestReader()
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			table := models.InitTable("orders", "public")
+			col := models.InitColumn("account_id", "orders", "public")
+			r.createInlineReferenceConstraint(table, col, tt.ref)
+			if tt.wantNone {
+				if len(table.Constraints) != 0 {
+					t.Fatalf("unexpected constraints: %v", table.Constraints)
+				}
+				return
+			}
+			c := table.Constraints["fk_orders_account_id"]
+			if c == nil {
+				t.Fatal("constraint missing")
+			}
+			if c.ReferencedSchema != tt.schema || c.ReferencedTable != tt.table ||
+				c.ReferencedColumns[0] != tt.col || c.OnDelete != tt.onDelete || c.OnUpdate != tt.onUpdate {
+				t.Errorf("constraint = %+v", c)
+			}
+		})
+	}
+}

pkg/readers/gorm/reader.go

@@ -726,6 +726,15 @@ func (r *Reader) parseColumn(fieldName string, fieldType ast.Expr, tag string, s
 	if _, ok := parts["autoincrement"]; ok {
 		column.AutoIncrement = true
 	}
+	if _, ok := parts["generated"]; ok {
+		// GENERATED ... STORED marker written by the GORM writer; the expression is not in the tag
+		column.Generated = true
+	}
+	if _, ok := parts["identity"]; ok {
+		// GENERATED ALWAYS AS IDENTITY marker written by the GORM writer
+		column.Identity = true
+		column.IdentityGeneration = "ALWAYS"
+	}
 	if def, ok := parts["default"]; ok {
 		// Default value from GORM tag (e.g., default:gen_random_uuid())
 		column.Default = def

pkg/readers/mssql/README.md

@@ -25,6 +25,11 @@ sqlserver://user:pass@192.168.1.100:1433/production
 sqlserver://localhost/testdb?encrypt=disable
 ```
 
+## Computed and Identity Columns
+
+- `sys.computed_columns` -> `Generated` + `GenerationExpression` (outer parentheses removed; persisted or not)
+- Identity columns -> `Identity` + `IdentityGeneration = ALWAYS`
+
 ## Supported Constraints
 
 - Primary Keys

pkg/readers/mssql/queries.go

@@ -104,8 +104,12 @@ func (r *Reader) queryColumns(schemaName string) (map[string]map[string]*models.
 			c.numeric_precision,
 			c.numeric_scale,
 			ISNULL(ep.value, '') as description,
-			COLUMNPROPERTY(OBJECT_ID(QUOTENAME(c.table_schema) + '.' + QUOTENAME(c.table_name)), c.column_name, 'IsIdentity') as is_identity
+			COLUMNPROPERTY(OBJECT_ID(QUOTENAME(c.table_schema) + '.' + QUOTENAME(c.table_name)), c.column_name, 'IsIdentity') as is_identity,
+			cc.definition as computed_definition
 		FROM information_schema.columns c
+		LEFT JOIN sys.computed_columns cc
+			ON cc.object_id = OBJECT_ID(QUOTENAME(c.table_schema) + '.' + QUOTENAME(c.table_name))
+			AND cc.name = c.column_name
 		LEFT JOIN sys.extended_properties ep
 			ON ep.major_id = OBJECT_ID(QUOTENAME(c.table_schema) + '.' + QUOTENAME(c.table_name))
 			AND ep.minor_id = COLUMNPROPERTY(OBJECT_ID(QUOTENAME(c.table_schema) + '.' + QUOTENAME(c.table_name)), c.column_name, 'ColumnId')
@@ -127,8 +131,9 @@ func (r *Reader) queryColumns(schemaName string) (map[string]map[string]*models.
 		var schema, tableName, columnName, isNullable, dataType, description string
 		var ordinalPosition int
 		var columnDefault, charMaxLength, numPrecision, numScale, isIdentity *int
+		var computedDefinition *string
 
-		if err := rows.Scan(&schema, &tableName, &columnName, &ordinalPosition, &columnDefault, &isNullable, &dataType, &charMaxLength, &numPrecision, &numScale, &description, &isIdentity); err != nil {
+		if err := rows.Scan(&schema, &tableName, &columnName, &ordinalPosition, &columnDefault, &isNullable, &dataType, &charMaxLength, &numPrecision, &numScale, &description, &isIdentity, &computedDefinition); err != nil {
 			return nil, err
 		}
 
@@ -144,6 +149,14 @@ func (r *Reader) queryColumns(schemaName string) (map[string]map[string]*models.
 		// Check if this is an identity column (auto-increment)
 		if isIdentity != nil && *isIdentity == 1 {
 			column.AutoIncrement = true
+			column.Identity = true
+			column.IdentityGeneration = "ALWAYS"
+		}
+
+		// Computed columns report their expression wrapped in an extra pair of parentheses
+		if computedDefinition != nil && strings.TrimSpace(*computedDefinition) != "" {
+			column.Generated = true
+			column.GenerationExpression = trimOuterParens(*computedDefinition)
 		}
 
 		if charMaxLength != nil && *charMaxLength > 0 {
@@ -414,3 +427,25 @@ func (r *Reader) queryIndexes(schemaName string) (map[string][]*models.Index, er
 
 	return indexes, rows.Err()
 }
+
+// trimOuterParens removes one pair of parentheses wrapping the whole expression, as
+// SQL Server stores computed column definitions.
+func trimOuterParens(expr string) string {
+	expr = strings.TrimSpace(expr)
+	if len(expr) < 2 || expr[0] != '(' || expr[len(expr)-1] != ')' {
+		return expr
+	}
+	depth := 0
+	for i := 0; i < len(expr); i++ {
+		switch expr[i] {
+		case '(':
+			depth++
+		case ')':
+			depth--
+			if depth == 0 && i != len(expr)-1 {
+				return expr
+			}
+		}
+	}
+	return strings.TrimSpace(expr[1 : len(expr)-1])
+}

pkg/readers/mssql/reader_test.go

@@ -85,3 +85,18 @@ func TestConvertMSSQLToCanonical(t *testing.T) {
 		})
 	}
 }
+
+func TestTrimOuterParens(t *testing.T) {
+	tests := map[string]string{
+		"(([a])+([b]))":   "([a])+([b])",
+		"([a]+[b])":       "[a]+[b]",
+		"([a])+([b])":     "([a])+([b])",
+		"[a]+[b]":         "[a]+[b]",
+		"  (upper([a])) ": "upper([a])",
+	}
+	for in, want := range tests {
+		if got := trimOuterParens(in); got != want {
+			t.Errorf("trimOuterParens(%q) = %q, want %q", in, got, want)
+		}
+	}
+}

pkg/readers/mysql/reader.go

@@ -0,0 +1,254 @@
+package mysql
+
+import (
+	"context"
+	"database/sql"
+	"fmt"
+	"strings"
+
+	_ "github.com/go-sql-driver/mysql"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/mariadb"
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+	"git.warky.dev/wdevs/relspecgo/pkg/readers"
+)
+
+type Reader struct {
+	options *readers.ReaderOptions
+	db      *sql.DB
+	ctx     context.Context
+}
+
+func NewReader(options *readers.ReaderOptions) *Reader {
+	return &Reader{options: options, ctx: context.Background()}
+}
+
+func (r *Reader) ReadDatabase() (*models.Database, error) {
+	if r.options == nil || r.options.ConnectionString == "" {
+		return nil, fmt.Errorf("connection string is required")
+	}
+	if err := r.connect(); err != nil {
+		return nil, fmt.Errorf("failed to connect: %w", err)
+	}
+	defer r.close()
+	var name, version string
+	if err := r.db.QueryRowContext(r.ctx, "SELECT DATABASE()").Scan(&name); err != nil {
+		return nil, fmt.Errorf("failed to get database name: %w", err)
+	}
+	_ = r.db.QueryRowContext(r.ctx, "SELECT VERSION()").Scan(&version)
+	db := models.InitDatabase(name)
+	db.DatabaseType, db.SourceFormat, db.DatabaseVersion = models.MySQLDatabaseType, "mysql", version
+	schemas, err := r.querySchemas(name)
+	if err != nil {
+		return nil, fmt.Errorf("failed to query schemas: %w", err)
+	}
+	for _, schema := range schemas {
+		tables, err := r.queryTables(schema.Name)
+		if err != nil {
+			return nil, err
+		}
+		schema.Tables = tables
+		for _, table := range tables {
+			table.Columns, err = r.queryColumns(schema.Name, table.Name)
+			if err != nil {
+				return nil, err
+			}
+			table.Constraints, err = r.queryConstraints(schema.Name, table.Name)
+			if err != nil {
+				return nil, err
+			}
+			table.Indexes, err = r.queryIndexes(schema.Name, table.Name)
+			if err != nil {
+				return nil, err
+			}
+			table.RefSchema = schema
+			for _, c := range table.Constraints {
+				if c.Type == models.ForeignKeyConstraint {
+					r.deriveRelationship(table, c)
+				}
+			}
+		}
+		schema.RefDatabase = db
+		db.Schemas = append(db.Schemas, schema)
+	}
+	return db, nil
+}
+
+func (r *Reader) ReadSchema() (*models.Schema, error) {
+	db, err := r.ReadDatabase()
+	if err != nil {
+		return nil, err
+	}
+	if len(db.Schemas) == 0 {
+		return nil, fmt.Errorf("no schemas found in database")
+	}
+	return db.Schemas[0], nil
+}
+
+func (r *Reader) ReadTable() (*models.Table, error) {
+	s, err := r.ReadSchema()
+	if err != nil {
+		return nil, err
+	}
+	if len(s.Tables) == 0 {
+		return nil, fmt.Errorf("no tables found in schema")
+	}
+	return s.Tables[0], nil
+}
+
+func (r *Reader) connect() error {
+	db, err := sql.Open("mysql", r.options.ConnectionString)
+	if err != nil {
+		return err
+	}
+	if err = db.PingContext(r.ctx); err != nil {
+		db.Close()
+		return err
+	}
+	r.db = db
+	return nil
+}
+
+func (r *Reader) close() {
+	if r.db != nil {
+		_ = r.db.Close()
+	}
+}
+func (r *Reader) mapDataType(t string) string { return mariadb.ConvertMariaDBToCanonical(t) }
+
+func (r *Reader) querySchemas(current string) ([]*models.Schema, error) {
+	rows, err := r.db.QueryContext(r.ctx, "SELECT SCHEMA_NAME FROM information_schema.SCHEMATA WHERE SCHEMA_NAME = ?", current)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	var out []*models.Schema
+	for rows.Next() {
+		var n string
+		if err := rows.Scan(&n); err != nil {
+			return nil, err
+		}
+		out = append(out, models.InitSchema(n))
+	}
+	return out, rows.Err()
+}
+
+func (r *Reader) queryTables(schema string) ([]*models.Table, error) {
+	rows, err := r.db.QueryContext(r.ctx, "SELECT TABLE_NAME FROM information_schema.TABLES WHERE TABLE_SCHEMA = ? AND TABLE_TYPE = 'BASE TABLE' ORDER BY TABLE_NAME", schema)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	var out []*models.Table
+	for rows.Next() {
+		var n string
+		if err := rows.Scan(&n); err != nil {
+			return nil, err
+		}
+		out = append(out, models.InitTable(n, schema))
+	}
+	return out, rows.Err()
+}
+
+func (r *Reader) queryColumns(schema, table string) (map[string]*models.Column, error) {
+	rows, err := r.db.QueryContext(r.ctx, `SELECT COLUMN_NAME, COLUMN_TYPE, IS_NULLABLE, COLUMN_DEFAULT, ORDINAL_POSITION, EXTRA, COLUMN_COMMENT FROM information_schema.COLUMNS WHERE TABLE_SCHEMA = ? AND TABLE_NAME = ? ORDER BY ORDINAL_POSITION`, schema, table)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	out := map[string]*models.Column{}
+	for rows.Next() {
+		var name, typ, nullable, extra, comment string
+		var def sql.NullString
+		var pos int
+		if err := rows.Scan(&name, &typ, &nullable, &def, &pos, &extra, &comment); err != nil {
+			return nil, err
+		}
+		c := models.InitColumn(name, table, schema)
+		c.Type = r.mapDataType(typ)
+		c.NotNull = strings.EqualFold(nullable, "NO")
+		c.Sequence = uint(pos)
+		c.Comment = comment
+		if def.Valid {
+			c.Default = def.String
+		}
+		c.AutoIncrement = strings.Contains(strings.ToLower(extra), "auto_increment")
+		out[name] = c
+	}
+	return out, rows.Err()
+}
+
+func (r *Reader) queryConstraints(schema, table string) (map[string]*models.Constraint, error) {
+	rows, err := r.db.QueryContext(r.ctx, `SELECT CONSTRAINT_NAME, CONSTRAINT_TYPE, COLUMN_NAME, REFERENCED_TABLE_SCHEMA, REFERENCED_TABLE_NAME, REFERENCED_COLUMN_NAME, ORDINAL_POSITION FROM information_schema.KEY_COLUMN_USAGE k JOIN information_schema.TABLE_CONSTRAINTS t USING (CONSTRAINT_SCHEMA, TABLE_NAME, CONSTRAINT_NAME) WHERE k.TABLE_SCHEMA=? AND k.TABLE_NAME=? ORDER BY CONSTRAINT_NAME, ORDINAL_POSITION`, schema, table)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	out := map[string]*models.Constraint{}
+	for rows.Next() {
+		var name, typ, col, rs, rt, rc string
+		var pos int
+		if err := rows.Scan(&name, &typ, &col, &rs, &rt, &rc, &pos); err != nil {
+			return nil, err
+		}
+		c := out[name]
+		if c == nil {
+			ct := models.UniqueConstraint
+			if typ == "PRIMARY KEY" {
+				ct = models.PrimaryKeyConstraint
+			}
+			if typ == "FOREIGN KEY" {
+				ct = models.ForeignKeyConstraint
+			}
+			c = models.InitConstraint(name, ct)
+			c.Schema = schema
+			c.Table = table
+			c.ReferencedSchema = rs
+			c.ReferencedTable = rt
+			out[name] = c
+		}
+		c.Columns = append(c.Columns, col)
+		if rc != "" {
+			c.ReferencedColumns = append(c.ReferencedColumns, rc)
+		}
+	}
+	return out, rows.Err()
+}
+
+func (r *Reader) queryIndexes(schema, table string) (map[string]*models.Index, error) {
+	rows, err := r.db.QueryContext(r.ctx, `SELECT INDEX_NAME, NON_UNIQUE, COLUMN_NAME, SEQ_IN_INDEX, INDEX_TYPE FROM information_schema.STATISTICS WHERE TABLE_SCHEMA=? AND TABLE_NAME=? AND INDEX_NAME <> 'PRIMARY' ORDER BY INDEX_NAME, SEQ_IN_INDEX`, schema, table)
+	if err != nil {
+		return nil, err
+	}
+	defer rows.Close()
+	out := map[string]*models.Index{}
+	for rows.Next() {
+		var name, col, typ string
+		var non, seq int
+		if err := rows.Scan(&name, &non, &col, &seq, &typ); err != nil {
+			return nil, err
+		}
+		i := out[name]
+		if i == nil {
+			i = models.InitIndex(name, table, schema)
+			i.Unique = non == 0
+			i.Type = strings.ToLower(typ)
+			out[name] = i
+		}
+		i.Columns = append(i.Columns, col)
+	}
+	return out, rows.Err()
+}
+
+func (r *Reader) deriveRelationship(t *models.Table, c *models.Constraint) {
+	n := fmt.Sprintf("%s_to_%s", t.Name, c.ReferencedTable)
+	rel := models.InitRelationship(n, models.OneToMany)
+	rel.FromTable = t.Name
+	rel.FromSchema = t.Schema
+	rel.FromColumns = append([]string(nil), c.Columns...)
+	rel.ToTable = c.ReferencedTable
+	rel.ToSchema = c.ReferencedSchema
+	rel.ToColumns = append([]string(nil), c.ReferencedColumns...)
+	rel.ForeignKey = c.Name
+	t.Relationships[n] = rel
+}

pkg/readers/mysql/reader_test.go

@@ -0,0 +1,22 @@
+package mysql
+
+import (
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/readers"
+)
+
+func TestReaderMapDataType(t *testing.T) {
+	r := NewReader(&readers.ReaderOptions{})
+	for _, tc := range []struct{ input, want string }{{"varchar(64)", "string"}, {"bigint unsigned", "int64"}, {"datetime", "timestamp"}, {"json", "json"}} {
+		if got := r.mapDataType(tc.input); got != tc.want {
+			t.Errorf("mapDataType(%q) = %q, want %q", tc.input, got, tc.want)
+		}
+	}
+}
+
+func TestReaderRequiresConnectionString(t *testing.T) {
+	if _, err := NewReader(&readers.ReaderOptions{}).ReadDatabase(); err == nil {
+		t.Fatal("expected missing connection string error")
+	}
+}

pkg/readers/pgsql/queries.go

@@ -243,7 +243,11 @@ func (r *Reader) queryColumns(schemaName string) (map[string]map[string]*models.
 			c.numeric_scale,
 			c.udt_name,
 			pg_catalog.format_type(a.atttypid, a.atttypmod) as formatted_data_type,
-			col_description((c.table_schema||'.'||c.table_name)::regclass, c.ordinal_position) as description
+			col_description((c.table_schema||'.'||c.table_name)::regclass, c.ordinal_position) as description,
+			c.is_generated,
+			c.generation_expression,
+			c.is_identity,
+			c.identity_generation
 		FROM information_schema.columns c
 		JOIN pg_catalog.pg_namespace n
 			ON n.nspname = c.table_schema
@@ -268,17 +272,36 @@ func (r *Reader) queryColumns(schemaName string) (map[string]map[string]*models.
 	columnsMap := make(map[string]map[string]*models.Column)
 
 	for rows.Next() {
-		var schema, tableName, columnName, isNullable, dataType, udtName, formattedDataType string
+		var schema, tableName, columnName, isNullable, dataType, udtName, formattedDataType, isGenerated, isIdentity string
 		var ordinalPosition int
-		var columnDefault, description *string
+		var columnDefault, description, generationExpression, identityGeneration *string
 		var charMaxLength, numPrecision, numScale *int
 
-		if err := rows.Scan(&schema, &tableName, &columnName, &ordinalPosition, &columnDefault, &isNullable, &dataType, &charMaxLength, &numPrecision, &numScale, &udtName, &formattedDataType, &description); err != nil {
+		if err := rows.Scan(&schema, &tableName, &columnName, &ordinalPosition, &columnDefault, &isNullable, &dataType, &charMaxLength, &numPrecision, &numScale, &udtName, &formattedDataType, &description, &isGenerated, &generationExpression, &isIdentity, &identityGeneration); err != nil {
 			return nil, err
 		}
 
 		column := models.InitColumn(columnName, tableName, schema)
 
+		// GENERATED ALWAYS ... STORED columns are computed by Postgres and cannot
+		// have their default altered/dropped like a regular column.
+		column.Generated = isGenerated == "ALWAYS"
+		if generationExpression != nil {
+			column.GenerationExpression = *generationExpression
+		}
+
+		// GENERATED { ALWAYS | BY DEFAULT } AS IDENTITY columns are driven by an
+		// internal sequence rather than a literal default (unlike serial columns,
+		// they carry no pg_attrdef row at all), so they need the same DB-side
+		// handling as generated columns even though the underlying mechanism differs.
+		column.Identity = isIdentity == "YES"
+		if identityGeneration != nil {
+			column.IdentityGeneration = strings.ToUpper(strings.TrimSpace(*identityGeneration))
+		}
+		if column.Identity {
+			column.AutoIncrement = true
+		}
+
 		// Check if this is a serial type (has nextval default)
 		hasNextval := false
 		if columnDefault != nil {
@@ -301,6 +324,24 @@ func (r *Reader) queryColumns(schemaName string) (map[string]map[string]*models.
 			column.Description = *description
 		}
 
+		if column.Generated {
+			note := "GENERATED ALWAYS AS (" + column.GenerationExpression + ") STORED"
+			if column.Description != "" {
+				column.Description = column.Description + " " + note
+			} else {
+				column.Description = note
+			}
+		}
+
+		if column.Identity {
+			note := "GENERATED " + column.IdentityGeneration + " AS IDENTITY"
+			if column.Description != "" {
+				column.Description = column.Description + " " + note
+			} else {
+				column.Description = note
+			}
+		}
+
 		if charMaxLength != nil {
 			column.Length = *charMaxLength
 		}
@@ -381,31 +422,35 @@ func (r *Reader) queryPrimaryKeys(schemaName string) (map[string]*models.Constra
 // queryForeignKeys retrieves all foreign key constraints for a schema
 // Returns map[schema.table][]*Constraint
 func (r *Reader) queryForeignKeys(schemaName string) (map[string][]*models.Constraint, error) {
+	// Columns are paired by position from pg_constraint so composite keys are not
+	// cross-joined (information_schema.constraint_column_usage has no ordering).
+	actionCase := func(col string) string {
+		return `CASE ` + col + ` WHEN 'a' THEN 'NO ACTION' WHEN 'r' THEN 'RESTRICT' WHEN 'c' THEN 'CASCADE' WHEN 'n' THEN 'SET NULL' WHEN 'd' THEN 'SET DEFAULT' END`
+	}
 	query := `
 		SELECT
-			tc.table_schema,
-			tc.table_name,
-			tc.constraint_name,
-			kcu.table_schema as foreign_table_schema,
-			kcu.table_name as foreign_table_name,
-			kcu.column_name as foreign_column,
-			ccu.table_schema as referenced_table_schema,
-			ccu.table_name as referenced_table_name,
-			ccu.column_name as referenced_column,
-			rc.update_rule,
-			rc.delete_rule
-		FROM information_schema.table_constraints tc
-		JOIN information_schema.key_column_usage kcu
-			ON tc.constraint_name = kcu.constraint_name
-			AND tc.table_schema = kcu.table_schema
-		JOIN information_schema.constraint_column_usage ccu
-			ON ccu.constraint_name = tc.constraint_name
-		JOIN information_schema.referential_constraints rc
-			ON rc.constraint_name = tc.constraint_name
-			AND rc.constraint_schema = tc.table_schema
-		WHERE tc.constraint_type = 'FOREIGN KEY'
-		  AND tc.table_schema = $1
-		ORDER BY tc.table_schema, tc.table_name, tc.constraint_name, kcu.ordinal_position
+			ns.nspname AS table_schema,
+			cl.relname AS table_name,
+			con.conname AS constraint_name,
+			ns.nspname AS foreign_table_schema,
+			cl.relname AS foreign_table_name,
+			att.attname AS foreign_column,
+			fns.nspname AS referenced_table_schema,
+			fcl.relname AS referenced_table_name,
+			fatt.attname AS referenced_column,
+			` + actionCase("con.confupdtype") + ` AS update_rule,
+			` + actionCase("con.confdeltype") + ` AS delete_rule
+		FROM pg_catalog.pg_constraint con
+		JOIN pg_catalog.pg_class cl ON cl.oid = con.conrelid
+		JOIN pg_catalog.pg_namespace ns ON ns.oid = cl.relnamespace
+		JOIN pg_catalog.pg_class fcl ON fcl.oid = con.confrelid
+		JOIN pg_catalog.pg_namespace fns ON fns.oid = fcl.relnamespace
+		CROSS JOIN LATERAL unnest(con.conkey, con.confkey) WITH ORDINALITY AS k(attnum, fattnum, ord)
+		JOIN pg_catalog.pg_attribute att ON att.attrelid = con.conrelid AND att.attnum = k.attnum
+		JOIN pg_catalog.pg_attribute fatt ON fatt.attrelid = con.confrelid AND fatt.attnum = k.fattnum
+		WHERE con.contype = 'f'
+		  AND ns.nspname = $1
+		ORDER BY ns.nspname, cl.relname, con.conname, k.ord
 	`
 
 	rows, err := r.conn.Query(r.ctx, query, schemaName)

pkg/readers/pgsql/queries_pure_test.go

@@ -0,0 +1,113 @@
+package pgsql
+
+import (
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+)
+
+func TestNormalizePostgresDefault(t *testing.T) {
+	tests := []struct {
+		name string
+		in   string
+		want string
+	}{
+		{"empty", "", ""},
+		{"function", "now()", "now()"},
+		{"nextval passthrough", "nextval('seq'::regclass)", "nextval('seq'::regclass)"},
+		{"number", "42", "42"},
+		{"null cast", "NULL::text", "NULL::text"},
+		{"quoted literal", "'abc'", "abc"},
+		{"quoted with cast", "'abc'::character varying", "abc"},
+		{"escaped quote", "'it''s'::text", "it's"},
+		{"empty literal", "''::text", ""},
+		{"only escaped quotes", "''''", "'"},
+		{"unterminated", "'abc", "abc"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := normalizePostgresDefault(tt.in); got != tt.want {
+				t.Errorf("normalizePostgresDefault(%q) = %q, want %q", tt.in, got, tt.want)
+			}
+		})
+	}
+}
+
+func TestCountHelpers(t *testing.T) {
+	cols := map[string]map[string]*models.Column{
+		"a": {"x": {}, "y": {}},
+		"b": {"z": {}},
+		"c": {},
+	}
+	if got := countColumns(cols); got != 3 {
+		t.Errorf("countColumns = %d, want 3", got)
+	}
+	if got := countColumns(nil); got != 0 {
+		t.Errorf("countColumns(nil) = %d, want 0", got)
+	}
+
+	cons := map[string][]*models.Constraint{"a": {{}, {}}, "b": {{}}}
+	if got := countConstraints(cons); got != 3 {
+		t.Errorf("countConstraints = %d, want 3", got)
+	}
+	if got := countConstraints(nil); got != 0 {
+		t.Errorf("countConstraints(nil) = %d, want 0", got)
+	}
+
+	idx := map[string][]*models.Index{"a": {{}}, "b": {{}, {}, {}}}
+	if got := countIndexes(idx); got != 4 {
+		t.Errorf("countIndexes = %d, want 4", got)
+	}
+	if got := countIndexes(nil); got != 0 {
+		t.Errorf("countIndexes(nil) = %d, want 0", got)
+	}
+}
+
+func TestExtractIndexOperatorClass(t *testing.T) {
+	tests := []struct {
+		name string
+		in   []string
+		want string
+	}{
+		{"none", nil, ""},
+		{"sort modifiers only", []string{"DESC", "NULLS", "LAST"}, ""},
+		{"opclass", []string{"", " Vector_Cosine_Ops "}, "vector_cosine_ops"},
+		{"opclass after ordering", []string{"desc", "gin_trgm_ops"}, "gin_trgm_ops"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := extractIndexOperatorClass(tt.in); got != tt.want {
+				t.Errorf("got %q, want %q", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestBuildIndexHint(t *testing.T) {
+	tests := []struct {
+		opClass, params, want string
+	}{
+		{"", "", ""},
+		{"vector_cosine_ops", "", "opclass=vector_cosine_ops"},
+		{"", "m=16", "with (m=16)"},
+		{"vector_cosine_ops", "m=16", "opclass=vector_cosine_ops; with (m=16)"},
+	}
+	for _, tt := range tests {
+		if got := buildIndexHint(tt.opClass, tt.params); got != tt.want {
+			t.Errorf("buildIndexHint(%q,%q) = %q, want %q", tt.opClass, tt.params, got, tt.want)
+		}
+	}
+}
+
+func TestNormalizeIndexStorageParams(t *testing.T) {
+	tests := []struct{ in, want string }{
+		{"", ""},
+		{"m='16', ef_construction='64'", "m=16, ef_construction=64"},
+		{"key_field='id'", "key_field='id'"},
+	}
+	for _, tt := range tests {
+		if got := normalizeIndexStorageParams(tt.in); got != tt.want {
+			t.Errorf("normalizeIndexStorageParams(%q) = %q, want %q", tt.in, got, tt.want)
+		}
+	}
+}

pkg/readers/pgsql/reader_test.go

@@ -1,10 +1,14 @@
 package pgsql
 
 import (
+	"context"
 	"os"
 	"reflect"
+	"strings"
 	"testing"
 
+	"github.com/jackc/pgx/v5"
+
 	"git.warky.dev/wdevs/relspecgo/pkg/models"
 	"git.warky.dev/wdevs/relspecgo/pkg/readers"
 )
@@ -499,3 +503,60 @@ func TestMapDataType_ExtensionTypesPreserveModifiers(t *testing.T) {
 		})
 	}
 }
+
+func TestReader_CompositeForeignKeyColumnsArePairedOnce(t *testing.T) {
+	connStr := getTestConnectionString(t)
+
+	ctx := context.Background()
+	conn, err := pgx.Connect(ctx, connStr)
+	if err != nil {
+		t.Fatalf("connect: %v", err)
+	}
+	defer conn.Close(ctx)
+
+	const schema = "relspec_fk_test"
+	setup := []string{
+		"DROP SCHEMA IF EXISTS " + schema + " CASCADE",
+		"CREATE SCHEMA " + schema,
+		"CREATE TABLE " + schema + ".parent (a int, b int, PRIMARY KEY (a, b))",
+		"CREATE TABLE " + schema + ".child (x int, y int, CONSTRAINT fk_child_parent FOREIGN KEY (x, y) REFERENCES " + schema + ".parent (a, b) ON DELETE CASCADE)",
+	}
+	for _, stmt := range setup {
+		if _, err := conn.Exec(ctx, stmt); err != nil {
+			t.Fatalf("setup %q: %v", stmt, err)
+		}
+	}
+	defer conn.Exec(ctx, "DROP SCHEMA IF EXISTS "+schema+" CASCADE")
+
+	reader := NewReader(&readers.ReaderOptions{ConnectionString: connStr})
+	db, err := reader.ReadDatabase()
+	if err != nil {
+		t.Fatalf("ReadDatabase: %v", err)
+	}
+
+	for _, s := range db.Schemas {
+		if s.Name != schema {
+			continue
+		}
+		for _, tbl := range s.Tables {
+			if tbl.Name != "child" {
+				continue
+			}
+			fk := tbl.Constraints["fk_child_parent"]
+			if fk == nil {
+				t.Fatal("foreign key fk_child_parent not read")
+			}
+			if got := strings.Join(fk.Columns, ","); got != "x,y" {
+				t.Errorf("columns = %q, want x,y", got)
+			}
+			if got := strings.Join(fk.ReferencedColumns, ","); got != "a,b" {
+				t.Errorf("referenced columns = %q, want a,b", got)
+			}
+			if fk.OnDelete != "CASCADE" || fk.OnUpdate != "NO ACTION" {
+				t.Errorf("rules = %s/%s, want CASCADE/NO ACTION", fk.OnDelete, fk.OnUpdate)
+			}
+			return
+		}
+	}
+	t.Fatal("test schema/table not found in read result")
+}

pkg/readers/prisma/reader.go

@@ -13,7 +13,8 @@ import (
 
 // Reader implements the readers.Reader interface for Prisma schema format
 type Reader struct {
-	options *readers.ReaderOptions
+	options   *readers.ReaderOptions
+	enumNames map[string]bool // enum names declared in the schema being parsed
 }
 
 // NewReader creates a new Prisma reader with the given options
@@ -82,6 +83,8 @@ func (r *Reader) parsePrisma(content string) (*models.Database, error) {
 	schema := models.InitSchema("public")
 	schema.Enums = make([]*models.Enum, 0)
 
+	r.enumNames = collectEnumNames(content)
+
 	scanner := bufio.NewScanner(strings.NewReader(content))
 
 	// State tracking
@@ -600,20 +603,21 @@ func (r *Reader) isPrimitiveType(typeName string) bool {
 	return false
 }
 
-// isEnumType checks if a type name might be an enum
-// Note: We can't definitively check against schema.Enums at parse time
-// because enums might be defined after the model, so we just check
-// if it starts with uppercase (Prisma convention for enums)
-func (r *Reader) isEnumType(typeName string, table *models.Table) bool {
-	// Simple heuristic: enum types start with uppercase letter
-	// and are not known model names (though we can't check that yet)
-	if len(typeName) > 0 && typeName[0] >= 'A' && typeName[0] <= 'Z' {
-		// Additional check: primitive types are already handled above
-		// So if it's uppercase and not primitive, it's likely an enum or model
-		// We'll assume it's an enum if it's a single word
-		return !strings.Contains(typeName, "_")
+// isEnumType reports whether typeName is an enum declared in the schema.
+// Enum names are collected up front because enums may be declared after the
+// models that use them.
+func (r *Reader) isEnumType(typeName string, _ *models.Table) bool {
+	return r.enumNames[typeName]
+}
+
+var enumDeclRegex = regexp.MustCompile(`(?m)^\s*enum\s+(\w+)\s*{`)
+
+func collectEnumNames(content string) map[string]bool {
+	names := make(map[string]bool)
+	for _, m := range enumDeclRegex.FindAllStringSubmatch(content, -1) {
+		names[m[1]] = true
 	}
-	return false
+	return names
 }
 
 // createConstraintFromRelation creates a FK constraint from a @relation attribute

pkg/readers/prisma/reader_full_test.go

@@ -0,0 +1,351 @@
+package prisma
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+	"git.warky.dev/wdevs/relspecgo/pkg/readers"
+)
+
+const examplePrisma = "../../../tests/assets/prisma/example.prisma"
+
+func readFixture(t *testing.T) *models.Schema {
+	t.Helper()
+	db, err := NewReader(&readers.ReaderOptions{FilePath: examplePrisma}).ReadDatabase()
+	if err != nil {
+		t.Fatal(err)
+	}
+	return db.Schemas[0]
+}
+
+func readSource(t *testing.T, src string) *models.Database {
+	t.Helper()
+	p := filepath.Join(t.TempDir(), "schema.prisma")
+	if err := os.WriteFile(p, []byte(src), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	db, err := NewReader(&readers.ReaderOptions{FilePath: p}).ReadDatabase()
+	if err != nil {
+		t.Fatal(err)
+	}
+	return db
+}
+
+func table(s *models.Schema, name string) *models.Table {
+	for _, tb := range s.Tables {
+		if tb.Name == name {
+			return tb
+		}
+	}
+	return nil
+}
+
+func TestFixture_NoRelationFieldColumns(t *testing.T) {
+	s := readFixture(t)
+	// Relation fields (user, author, posts, profile, categories) are not columns.
+	for tbl, fields := range map[string][]string{
+		"User": {"posts", "profile"}, "Profile": {"user"}, "Post": {"author", "categories"}, "Category": {"posts"},
+	} {
+		for _, f := range fields {
+			if _, ok := table(s, tbl).Columns[f]; ok {
+				t.Errorf("%s.%s is a relation field and must not be a column", tbl, f)
+			}
+		}
+	}
+	// Enum-typed fields stay columns.
+	if c := table(s, "User").Columns["role"]; c == nil || c.Type != "Role" || c.Default != "USER" {
+		t.Errorf("User.role: %+v", c)
+	}
+}
+
+func TestFixture_Structure(t *testing.T) {
+	s := readFixture(t)
+	if len(s.Enums) != 1 || s.Enums[0].Name != "Role" || len(s.Enums[0].Values) != 2 {
+		t.Errorf("enums: %+v", s.Enums)
+	}
+	for _, n := range []string{"User", "Profile", "Post", "Category", "_CategoryToPost"} {
+		if table(s, n) == nil {
+			t.Errorf("table %s missing", n)
+		}
+	}
+
+	user := table(s, "User")
+	if id := user.Columns["id"]; id == nil || !id.IsPrimaryKey || !id.AutoIncrement || id.Type != "integer" {
+		t.Errorf("User.id: %+v", id)
+	}
+	if c := user.Columns["name"]; c == nil || c.NotNull {
+		t.Errorf("optional name: %+v", c)
+	}
+	if uq := user.Constraints["uq_email"]; uq == nil || uq.Columns[0] != "email" {
+		t.Errorf("unique: %+v", user.Constraints)
+	}
+
+	post := table(s, "Post")
+	if c := post.Columns["createdAt"]; c == nil || c.Type != "timestamp" || c.Default != "now()" {
+		t.Errorf("createdAt: %+v", c)
+	}
+	if c := post.Columns["updatedAt"]; c == nil || !strings.Contains(c.Comment, "@updatedAt") {
+		t.Errorf("updatedAt: %+v", c)
+	}
+	if c := post.Columns["published"]; c == nil || c.Default != false {
+		t.Errorf("published default: %+v", c)
+	}
+}
+
+func TestFixture_Relations(t *testing.T) {
+	s := readFixture(t)
+
+	fk := table(s, "Post").Constraints["fk_Post_authorId"]
+	if fk == nil || fk.Type != models.ForeignKeyConstraint || fk.Columns[0] != "authorId" || fk.ReferencedTable != "User" || fk.ReferencedColumns[0] != "id" {
+		t.Errorf("Post.author fk: %+v", fk)
+	}
+
+	jt := table(s, "_CategoryToPost")
+	if len(jt.Columns) != 2 {
+		t.Fatalf("join columns: %v", jt.Columns)
+	}
+	var pk, fks int
+	for _, c := range jt.Constraints {
+		switch c.Type {
+		case models.PrimaryKeyConstraint:
+			pk++
+		case models.ForeignKeyConstraint:
+			fks++
+			if c.OnDelete != "Cascade" {
+				t.Errorf("join fk on delete: %q", c.OnDelete)
+			}
+		}
+	}
+	if pk != 1 || fks != 2 {
+		t.Errorf("join constraints: pk=%d fks=%d", pk, fks)
+	}
+}
+
+func TestBlockAttributesAndDefaults(t *testing.T) {
+	db := readSource(t, `datasource db {
+  provider = "mysql"
+}
+
+model Membership {
+  userId  Int
+  groupId Int
+  role    String @default("member")
+  alias   String @default('x')
+  score   Float  @default(1.5)
+  tag     String @default(cuid())
+  token   String @default(uuid())
+  user    User   @relation(fields: [userId], references: [id], onDelete: Cascade, onUpdate: Restrict)
+  @@id([userId, groupId])
+  @@unique([userId, role])
+  @@index([groupId])
+  @@map("memberships")
+}
+
+model User {
+  id          Int @id
+  memberships Membership[]
+  slug        String @unique @default(dbgenerated("abc(1)"))
+}
+`)
+	if db.DatabaseType != "mysql" {
+		t.Errorf("db type: %q", db.DatabaseType)
+	}
+	m := table(db.Schemas[0], "Membership")
+
+	pk := m.Constraints["pk_Membership"]
+	if pk == nil || len(pk.Columns) != 2 || !m.Columns["userId"].IsPrimaryKey || !m.Columns["groupId"].NotNull {
+		t.Errorf("composite pk: %+v", pk)
+	}
+	if uq := m.Constraints["uq_Membership_userId_role"]; uq == nil || len(uq.Columns) != 2 {
+		t.Errorf("composite unique: %+v", m.Constraints)
+	}
+	if ix := m.Indexes["idx_Membership_groupId"]; ix == nil || ix.Columns[0] != "groupId" {
+		t.Errorf("index: %+v", m.Indexes)
+	}
+
+	checks := map[string]any{"role": "member", "alias": "x", "score": "1.5"}
+	for col, want := range checks {
+		if got := m.Columns[col].Default; got != want {
+			t.Errorf("%s default = %#v, want %#v", col, got, want)
+		}
+	}
+	if m.Columns["tag"].Comment != "default(cuid())" {
+		t.Errorf("cuid comment: %q", m.Columns["tag"].Comment)
+	}
+	if m.Columns["token"].Default != "gen_random_uuid()" {
+		t.Errorf("uuid default: %v", m.Columns["token"].Default)
+	}
+	if m.Columns["score"].Type != "double precision" {
+		t.Errorf("score type: %s", m.Columns["score"].Type)
+	}
+
+	fk := m.Constraints["fk_Membership_userId"]
+	if fk == nil || fk.OnDelete != "Cascade" || fk.OnUpdate != "Restrict" {
+		t.Errorf("fk actions: %+v", fk)
+	}
+
+	// Default with nested parentheses is extracted whole.
+	if got := table(db.Schemas[0], "User").Columns["slug"].Default; got != `dbgenerated("abc(1)")` {
+		t.Errorf("nested default: %#v", got)
+	}
+}
+
+func TestEnumDeclaredAfterModel(t *testing.T) {
+	db := readSource(t, `model Account {
+  id     Int    @id
+  status Status @default(ACTIVE)
+  owner  Owner?
+}
+
+model Owner {
+  id Int @id
+}
+
+enum Status {
+  ACTIVE
+  CLOSED
+}
+`)
+	a := table(db.Schemas[0], "Account")
+	if c := a.Columns["status"]; c == nil || c.Type != "Status" {
+		t.Errorf("enum column declared before enum: %+v", c)
+	}
+	if _, ok := a.Columns["owner"]; ok {
+		t.Error("model-typed field must not be a column")
+	}
+}
+
+func TestParseDatasourceProviders(t *testing.T) {
+	r := &Reader{}
+	tests := []struct {
+		provider string
+		want     models.DatabaseType
+	}{
+		{`"postgresql"`, models.PostgresqlDatabaseType},
+		{`"postgres"`, models.PostgresqlDatabaseType},
+		{`"mysql"`, "mysql"},
+		{`"sqlite"`, models.SqlLiteDatabaseType},
+		{`"sqlserver"`, models.MSSQLDatabaseType},
+		{`"cockroachdb"`, models.PostgresqlDatabaseType},
+	}
+	for _, tt := range tests {
+		db := models.InitDatabase("d")
+		r.parseDatasource([]string{"  provider = " + tt.provider}, db)
+		if db.DatabaseType != tt.want {
+			t.Errorf("%s -> %q, want %q", tt.provider, db.DatabaseType, tt.want)
+		}
+	}
+}
+
+func TestParseGenerator(t *testing.T) {
+	tests := []struct {
+		name  string
+		lines []string
+		opts  *readers.ReaderOptions
+		want  string
+	}{
+		{"js client", []string{`provider = "prisma-client-js"`}, &readers.ReaderOptions{}, "prisma"},
+		{"new client", []string{`provider = "prisma-client"`}, &readers.ReaderOptions{}, "prisma7"},
+		{"no provider, flag", []string{`output = "x"`}, &readers.ReaderOptions{Prisma7: true}, "prisma7"},
+		{"no provider, nil options", nil, nil, ""},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			db := models.InitDatabase("d")
+			db.SourceFormat = ""
+			(&Reader{options: tt.opts}).parseGenerator(tt.lines, db)
+			if db.SourceFormat != tt.want {
+				t.Errorf("got %q, want %q", db.SourceFormat, tt.want)
+			}
+		})
+	}
+}
+
+func TestPrisma7FlagWithoutGeneratorBlock(t *testing.T) {
+	p := filepath.Join(t.TempDir(), "s.prisma")
+	if err := os.WriteFile(p, []byte("model A {\n  id Int @id\n}\n"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	db, err := NewReader(&readers.ReaderOptions{FilePath: p, Prisma7: true}).ReadDatabase()
+	if err != nil || db.SourceFormat != "prisma7" {
+		t.Errorf("%v %q", err, db.SourceFormat)
+	}
+}
+
+func TestMetadataNameAndComments(t *testing.T) {
+	p := filepath.Join(t.TempDir(), "s.prisma")
+	src := "// leading comment\nmodel A {\n  // inner comment\n  id Int @id\n}\n"
+	if err := os.WriteFile(p, []byte(src), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	db, err := NewReader(&readers.ReaderOptions{FilePath: p, Metadata: map[string]any{"name": "shop"}}).ReadDatabase()
+	if err != nil || db.Name != "shop" || len(db.Schemas[0].Tables[0].Columns) != 1 {
+		t.Errorf("%v %+v", err, db)
+	}
+}
+
+func TestReadSchemaAndTable(t *testing.T) {
+	r := NewReader(&readers.ReaderOptions{FilePath: examplePrisma})
+	s, err := r.ReadSchema()
+	if err != nil || s.Name != "public" {
+		t.Fatalf("schema: %v", err)
+	}
+	tbl, err := r.ReadTable()
+	if err != nil || tbl.Name != "User" {
+		t.Fatalf("table: %v %+v", err, tbl)
+	}
+}
+
+func TestReader_Errors(t *testing.T) {
+	if _, err := NewReader(&readers.ReaderOptions{}).ReadDatabase(); err == nil || !strings.Contains(err.Error(), "file path is required") {
+		t.Errorf("empty path: %v", err)
+	}
+	if _, err := NewReader(&readers.ReaderOptions{FilePath: filepath.Join(t.TempDir(), "x")}).ReadDatabase(); err == nil || !strings.Contains(err.Error(), "failed to read file") {
+		t.Errorf("missing file: %v", err)
+	}
+	if _, err := NewReader(&readers.ReaderOptions{}).ReadSchema(); err == nil {
+		t.Error("ReadSchema without path")
+	}
+	if _, err := NewReader(&readers.ReaderOptions{}).ReadTable(); err == nil {
+		t.Error("ReadTable without path")
+	}
+	empty := filepath.Join(t.TempDir(), "e.prisma")
+	if err := os.WriteFile(empty, []byte("// nothing\n"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	if _, err := NewReader(&readers.ReaderOptions{FilePath: empty}).ReadTable(); err == nil || !strings.Contains(err.Error(), "no tables found") {
+		t.Errorf("ReadTable on empty: %v", err)
+	}
+}
+
+func TestExtractDefaultValue(t *testing.T) {
+	r := &Reader{}
+	tests := []struct{ in, want string }{
+		{"@id @default(autoincrement())", "autoincrement()"},
+		{`@default("a(b)")`, `"a(b)"`},
+		{"@unique", ""},
+		{"@default(unclosed(", ""},
+	}
+	for _, tt := range tests {
+		if got := r.extractDefaultValue(tt.in); got != tt.want {
+			t.Errorf("%q = %q, want %q", tt.in, got, tt.want)
+		}
+	}
+}
+
+func TestPrismaTypeToSQL(t *testing.T) {
+	r := &Reader{}
+	tests := map[string]string{
+		"String": "text", "Boolean": "boolean", "Int": "integer", "BigInt": "bigint",
+		"Float": "double precision", "Decimal": "decimal", "DateTime": "timestamp",
+		"Json": "jsonb", "Bytes": "bytea", "Custom": "Custom",
+	}
+	for in, want := range tests {
+		if got := r.prismaTypeToSQL(in); got != want {
+			t.Errorf("%s = %s, want %s", in, got, want)
+		}
+	}
+}

pkg/readers/sqlite/README.md

@@ -40,6 +40,7 @@ options := &readers.ReaderOptions{
 - Uses pure Go driver (modernc.org/sqlite) - no CGo required
 - Supports both file path and connection string
 - Auto-increment detection for INTEGER PRIMARY KEY columns
+- Generated columns (virtual and stored) are read via `PRAGMA table_xinfo`; the expression is parsed from the `CREATE TABLE` SQL
 - Foreign keys require `PRAGMA foreign_keys = ON` to be set
 
 ## Example Schema

pkg/readers/sqlite/generated.go

@@ -0,0 +1,116 @@
+package sqlite
+
+import (
+	"regexp"
+	"strings"
+)
+
+var generatedAsRegex = regexp.MustCompile(`(?is)\bAS\s*\(`)
+
+// parseGeneratedExpression extracts the expression of a generated column from a CREATE
+// TABLE statement, e.g. `full TEXT GENERATED ALWAYS AS (a || b) STORED` yields `a || b`.
+// It returns "" when the column or its expression cannot be found.
+func parseGeneratedExpression(createSQL, columnName string) string {
+	open := strings.Index(createSQL, "(")
+	if open < 0 {
+		return ""
+	}
+
+	for _, def := range splitTopLevel(createSQL[open+1:]) {
+		if !strings.EqualFold(firstIdentifier(def), columnName) {
+			continue
+		}
+		loc := generatedAsRegex.FindStringIndex(def)
+		if loc == nil {
+			return ""
+		}
+		return balancedParens(def[loc[1]-1:])
+	}
+	return ""
+}
+
+// splitTopLevel splits a table body on commas that are outside parentheses and quotes,
+// stopping at the parenthesis that closes the body.
+func splitTopLevel(body string) []string {
+	var parts []string
+	depth := 0
+	var quote byte
+	start := 0
+	for i := 0; i < len(body); i++ {
+		ch := body[i]
+		if quote != 0 {
+			if ch == quote {
+				quote = 0
+			}
+			continue
+		}
+		switch ch {
+		case '\'', '"', '`':
+			quote = ch
+		case '[':
+			quote = ']'
+		case '(':
+			depth++
+		case ')':
+			if depth == 0 {
+				return append(parts, body[start:i])
+			}
+			depth--
+		case ',':
+			if depth == 0 {
+				parts = append(parts, body[start:i])
+				start = i + 1
+			}
+		}
+	}
+	return append(parts, body[start:])
+}
+
+// firstIdentifier returns the first (possibly quoted) identifier of a column definition.
+func firstIdentifier(def string) string {
+	def = strings.TrimSpace(def)
+	if def == "" {
+		return ""
+	}
+	switch def[0] {
+	case '"', '\'', '`':
+		if end := strings.IndexByte(def[1:], def[0]); end >= 0 {
+			return def[1 : 1+end]
+		}
+	case '[':
+		if end := strings.IndexByte(def, ']'); end >= 0 {
+			return def[1:end]
+		}
+	}
+	if end := strings.IndexAny(def, " \t\r\n"); end >= 0 {
+		return def[:end]
+	}
+	return def
+}
+
+// balancedParens returns the text inside the parenthesis group that starts at s[0].
+func balancedParens(s string) string {
+	depth := 0
+	var quote byte
+	for i := 0; i < len(s); i++ {
+		ch := s[i]
+		if quote != 0 {
+			if ch == quote {
+				quote = 0
+			}
+			continue
+		}
+		switch ch {
+		case '\'', '"':
+			quote = ch
+		case '(':
+			depth++
+		case ')':
+			depth--
+			if depth == 0 {
+				return strings.TrimSpace(s[1:i])
+			}
+		}
+	}
+	return ""
+}

pkg/readers/sqlite/generated_test.go

@@ -0,0 +1,64 @@
+package sqlite
+
+import (
+	"database/sql"
+	"path/filepath"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/readers"
+)
+
+func TestParseGeneratedExpression(t *testing.T) {
+	createSQL := `CREATE TABLE people (
+		id INTEGER PRIMARY KEY,
+		"first" TEXT,
+		last TEXT,
+		full_name TEXT GENERATED ALWAYS AS (coalesce("first", '') || ' ' || last) STORED,
+		initials TEXT AS (substr("first", 1, 1) || substr(last, 1, 1)),
+		plain TEXT NOT NULL
+	)`
+
+	tests := []struct {
+		column string
+		want   string
+	}{
+		{"full_name", `coalesce("first", '') || ' ' || last`},
+		{"initials", `substr("first", 1, 1) || substr(last, 1, 1)`},
+		{"plain", ""},
+		{"missing", ""},
+	}
+	for _, tt := range tests {
+		t.Run(tt.column, func(t *testing.T) {
+			assert.Equal(t, tt.want, parseGeneratedExpression(createSQL, tt.column))
+		})
+	}
+}
+
+func TestReader_GeneratedColumns(t *testing.T) {
+	dbPath := filepath.Join(t.TempDir(), "gen.db")
+	db, err := sql.Open("sqlite", dbPath)
+	require.NoError(t, err)
+	_, err = db.Exec(`CREATE TABLE people (
+		id INTEGER PRIMARY KEY,
+		first TEXT,
+		last TEXT,
+		full_name TEXT GENERATED ALWAYS AS (first || ' ' || last) STORED,
+		initials TEXT AS (substr(first, 1, 1) || substr(last, 1, 1))
+	)`)
+	require.NoError(t, err)
+	require.NoError(t, db.Close())
+
+	got, err := NewReader(&readers.ReaderOptions{FilePath: dbPath}).ReadDatabase()
+	require.NoError(t, err)
+
+	cols := got.Schemas[0].Tables[0].Columns
+	require.Contains(t, cols, "full_name", "generated columns must be read")
+	assert.True(t, cols["full_name"].Generated)
+	assert.Equal(t, "first || ' ' || last", cols["full_name"].GenerationExpression)
+	assert.True(t, cols["initials"].Generated)
+	assert.Equal(t, "substr(first, 1, 1) || substr(last, 1, 1)", cols["initials"].GenerationExpression)
+	assert.False(t, cols["first"].Generated)
+}

pkg/readers/sqlite/queries.go

@@ -75,7 +75,8 @@ func (r *Reader) queryViews() ([]*models.View, error) {
 
 // queryColumns retrieves all columns for a given table or view
 func (r *Reader) queryColumns(tableName string) (map[string]*models.Column, error) {
-	query := fmt.Sprintf("PRAGMA table_info(%s)", tableName)
+	// table_xinfo, unlike table_info, also lists generated columns (hidden = 2 virtual, 3 stored)
+	query := fmt.Sprintf("PRAGMA table_xinfo(%s)", tableName)
 
 	rows, err := r.db.QueryContext(r.ctx, query)
 	if err != nil {
@@ -84,24 +85,38 @@ func (r *Reader) queryColumns(tableName string) (map[string]*models.Column, erro
 	defer rows.Close()
 
 	columns := make(map[string]*models.Column)
+	var tableSQL string
+	tableSQLLoaded := false
 
 	for rows.Next() {
 		var cid int
 		var name, dataType string
-		var notNull, pk int
+		var notNull, pk, hidden int
 		var defaultValue *string
 
-		if err := rows.Scan(&cid, &name, &dataType, &notNull, &defaultValue, &pk); err != nil {
+		if err := rows.Scan(&cid, &name, &dataType, &notNull, &defaultValue, &pk, &hidden); err != nil {
 			return nil, err
 		}
 
+		// Hidden virtual-table columns (hidden = 1) are not part of the schema
+		if hidden == 1 {
+			continue
+		}
+
 		column := models.InitColumn(name, tableName, "main")
 		column.Type = r.mapDataType(strings.ToUpper(dataType))
 		column.NotNull = (notNull == 1)
 		column.IsPrimaryKey = (pk > 0)
 		column.Sequence = uint(cid + 1)
 
-		if defaultValue != nil {
+		if hidden == 2 || hidden == 3 {
+			column.Generated = true
+			if !tableSQLLoaded {
+				tableSQL = r.tableSQL(tableName)
+				tableSQLLoaded = true
+			}
+			column.GenerationExpression = parseGeneratedExpression(tableSQL, name)
+		} else if defaultValue != nil {
 			column.Default = *defaultValue
 		}
 
@@ -116,6 +131,16 @@ func (r *Reader) queryColumns(tableName string) (map[string]*models.Column, erro
 	return columns, rows.Err()
 }
 
+// tableSQL returns the CREATE TABLE statement of a table, or "" when it cannot be read.
+func (r *Reader) tableSQL(tableName string) string {
+	var sql string
+	err := r.db.QueryRowContext(r.ctx, `SELECT sql FROM sqlite_master WHERE type = 'table' AND name = ?`, tableName).Scan(&sql)
+	if err != nil {
+		return ""
+	}
+	return sql
+}
+
 // isAutoIncrement checks if a column is autoincrement
 func (r *Reader) isAutoIncrement(tableName, columnName string) bool {
 	// Check sqlite_sequence table or parse CREATE TABLE statement

pkg/readers/typeorm/README.md

@@ -114,6 +114,8 @@ export class Post {
 - `@JoinColumn()` - Foreign key column
 - `@Index()` - Index definition
 - `@Unique()` - Unique constraint
+- `asExpression` / `generatedType` - Generated column (`Generated` + `GenerationExpression`)
+- `@PrimaryGeneratedColumn('identity')`, `@Generated('identity')`, `generatedIdentity` - Identity column
 
 ## Notes
 

pkg/readers/typeorm/reader.go

@@ -128,7 +128,6 @@ func (r *Reader) extractEntities(content string) []entityInfo {
 	scanner := bufio.NewScanner(strings.NewReader(content))
 
 	entityRegex := regexp.MustCompile(`^export\s+class\s+(\w+)`)
-	decoratorRegex := regexp.MustCompile(`^\s*@(\w+)(\([^)]*\))?`)
 	fieldRegex := regexp.MustCompile(`^\s*(\w+):\s*([^;]+);`)
 
 	var currentEntity *entityInfo
@@ -145,8 +144,7 @@ func (r *Reader) extractEntities(content string) []entityInfo {
 		}
 
 		// Check for decorator
-		if matches := decoratorRegex.FindStringSubmatch(trimmed); matches != nil {
-			decorator := matches[0]
+		if decorator, ok := matchDecorator(trimmed); ok {
 			pendingDecorators = append(pendingDecorators, decorator)
 			continue
 		}
@@ -488,7 +486,11 @@ func (r *Reader) parseColumnDecorator(decorator string, column *models.Column, t
 		column.IsPrimaryKey = true
 		column.NotNull = true
 
-		if strings.Contains(decorator, "'uuid'") {
+		if strings.Contains(decorator, "'identity'") {
+			column.AutoIncrement = true
+			column.Identity = true
+			column.IdentityGeneration = parseGeneratedIdentity(decorator)
+		} else if strings.Contains(decorator, "'uuid'") {
 			column.Type = "uuid"
 			column.Default = "gen_random_uuid()"
 		} else if strings.Contains(decorator, "'increment'") || strings.Contains(decorator, "()") {
@@ -497,6 +499,17 @@ func (r *Reader) parseColumnDecorator(decorator string, column *models.Column, t
 		return
 	}
 
+	// @Generated('identity') on a non-key column; generatedIdentity is read from @Column
+	if strings.HasPrefix(decorator, "@Generated") {
+		if strings.Contains(decorator, "'identity'") {
+			column.Identity = true
+			if column.IdentityGeneration == "" {
+				column.IdentityGeneration = "BY DEFAULT"
+			}
+		}
+		return
+	}
+
 	// @Column
 	if strings.HasPrefix(decorator, "@Column") {
 		r.parseColumnOptions(decorator, column, table)
@@ -586,6 +599,14 @@ func (r *Reader) parseColumnOptions(decorator string, column *models.Column, tab
 		}
 	}
 
+	if matches := asExpressionRegex.FindStringSubmatch(content); matches != nil {
+		column.Generated = true
+		column.GenerationExpression = unescapeSingleQuoted(matches[1])
+	}
+	if strings.Contains(content, "generatedIdentity") {
+		column.IdentityGeneration = parseGeneratedIdentity(content)
+	}
+
 	if strings.Contains(content, "nullable: true") || strings.Contains(content, "nullable:true") {
 		column.NotNull = false
 	}
@@ -834,3 +855,72 @@ func (r *Reader) getPrimaryKeyColumn(table *models.Table) *models.Column {
 
 	return pk
 }
+
+var (
+	asExpressionRegex       = regexp.MustCompile(`asExpression:\s*'((?:\\.|[^\\'])*)'`)
+	generatedIdentityRegexp = regexp.MustCompile(`generatedIdentity:\s*['"](ALWAYS|BY DEFAULT)['"]`)
+)
+
+// parseGeneratedIdentity returns the identity generation mode named in a decorator
+// ("ALWAYS" or "BY DEFAULT"), defaulting to "BY DEFAULT" as TypeORM does.
+func parseGeneratedIdentity(decorator string) string {
+	if matches := generatedIdentityRegexp.FindStringSubmatch(decorator); matches != nil {
+		return matches[1]
+	}
+	return "BY DEFAULT"
+}
+
+// unescapeSingleQuoted reverses escaping applied inside a quoted TypeScript string.
+func unescapeSingleQuoted(s string) string {
+	var sb strings.Builder
+	for i := 0; i < len(s); i++ {
+		if s[i] == '\\' && i+1 < len(s) {
+			i++
+		}
+		sb.WriteByte(s[i])
+	}
+	return sb.String()
+}
+
+var decoratorNameRegex = regexp.MustCompile(`^@\w+`)
+
+// matchDecorator returns the decorator at the start of line, including its argument list.
+// Parentheses inside quoted strings (e.g. a generated column expression) do not end it.
+func matchDecorator(line string) (string, bool) {
+	name := decoratorNameRegex.FindString(line)
+	if name == "" {
+		return "", false
+	}
+	rest := line[len(name):]
+	if !strings.HasPrefix(rest, "(") {
+		return name, true
+	}
+
+	depth := 0
+	var quote byte
+	for i := 0; i < len(rest); i++ {
+		ch := rest[i]
+		if quote != 0 {
+			switch ch {
+			case '\\':
+				i++
+			case quote:
+				quote = 0
+			}
+			continue
+		}
+		switch ch {
+		case '\'', '"', '`':
+			quote = ch
+		case '(':
+			depth++
+		case ')':
+			depth--
+			if depth == 0 {
+				return name + rest[:i+1], true
+			}
+		}
+	}
+	// Unterminated argument list: keep the whole line, as a best effort
+	return line, true
+}

pkg/readers/typeorm/reader_full_test.go

@@ -0,0 +1,381 @@
+package typeorm
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+	"git.warky.dev/wdevs/relspecgo/pkg/readers"
+)
+
+const exampleTS = "../../../tests/assets/typeorm/example.ts"
+
+func readFixture(t *testing.T) *models.Schema {
+	t.Helper()
+	db, err := NewReader(&readers.ReaderOptions{FilePath: exampleTS}).ReadDatabase()
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(db.Schemas) != 1 {
+		t.Fatalf("schemas: %d", len(db.Schemas))
+	}
+	return db.Schemas[0]
+}
+
+func tableByName(s *models.Schema, name string) *models.Table {
+	for _, t := range s.Tables {
+		if t.Name == name {
+			return t
+		}
+	}
+	return nil
+}
+
+func parseSource(t *testing.T, src string) *models.Schema {
+	t.Helper()
+	db, err := NewReader(&readers.ReaderOptions{}).parseTypeORM(src)
+	if err != nil {
+		t.Fatal(err)
+	}
+	return db.Schemas[0]
+}
+
+func TestReadFixture_Tables(t *testing.T) {
+	s := readFixture(t)
+	for _, name := range []string{"User", "Project", "Task", "Comment", "Tag", "user_project", "tag_task"} {
+		if tableByName(s, name) == nil {
+			t.Errorf("table %q missing", name)
+		}
+	}
+	if len(s.Tables) != 7 {
+		t.Errorf("tables: %d, want 7 (5 entities + 2 join tables)", len(s.Tables))
+	}
+}
+
+func TestReadFixture_ColumnsAndKeys(t *testing.T) {
+	s := readFixture(t)
+
+	user := tableByName(s, "User")
+	id := user.Columns["id"]
+	if id == nil || id.Type != "uuid" || !id.IsPrimaryKey || id.Default != "gen_random_uuid()" {
+		t.Errorf("User.id: %+v", id)
+	}
+	if c := user.Columns["createdAt"]; c == nil || c.Type != "timestamp" || c.Default != "now()" {
+		t.Errorf("User.createdAt: %+v", c)
+	}
+	if c := user.Columns["updatedAt"]; c == nil || c.Type != "timestamp" || !strings.Contains(c.Comment, "auto-update") {
+		t.Errorf("User.updatedAt: %+v", c)
+	}
+	if uq := user.Constraints["uq_email"]; uq == nil || uq.Type != models.UniqueConstraint || uq.Columns[0] != "email" {
+		t.Errorf("unique email: %+v", user.Constraints)
+	}
+	if _, ok := user.Columns["ownedProjects"]; ok {
+		t.Error("relation fields must not become columns")
+	}
+
+	project := tableByName(s, "Project")
+	if c := project.Columns["description"]; c == nil || c.NotNull {
+		t.Errorf("nullable description: %+v", c)
+	}
+	if c := project.Columns["status"]; c == nil || c.Default != "active" {
+		t.Errorf("status default: %+v", c)
+	}
+	if c := tableByName(s, "Task").Columns["description"]; c == nil || c.Type != "text" || c.NotNull {
+		t.Errorf("Task.description: %+v", c)
+	}
+	if c := tableByName(s, "Comment").Columns["content"]; c == nil || c.Type != "text" {
+		t.Errorf("shorthand type: %+v", c)
+	}
+}
+
+func TestReadFixture_Relationships(t *testing.T) {
+	s := readFixture(t)
+
+	fk := tableByName(s, "Project").Constraints["fk_Project_owner"]
+	if fk == nil || fk.Type != models.ForeignKeyConstraint || fk.Columns[0] != "ownerId" || fk.ReferencedTable != "User" {
+		t.Errorf("Project.owner fk: %+v", fk)
+	}
+	if c := tableByName(s, "Project").Columns["ownerId"]; c == nil || c.Type != "uuid" || !c.NotNull {
+		t.Errorf("ownerId column: %+v", c)
+	}
+	// ManyToOne with { nullable: true } produces a nullable FK column.
+	if c := tableByName(s, "Task").Columns["assigneeId"]; c == nil || c.NotNull {
+		t.Errorf("assigneeId must be nullable: %+v", c)
+	}
+
+	for _, jt := range []string{"user_project", "tag_task"} {
+		tbl := tableByName(s, jt)
+		if len(tbl.Columns) != 2 {
+			t.Errorf("%s columns: %d", jt, len(tbl.Columns))
+		}
+		pk := 0
+		fks := 0
+		for _, c := range tbl.Constraints {
+			switch c.Type {
+			case models.PrimaryKeyConstraint:
+				pk++
+				if len(c.Columns) != 2 {
+					t.Errorf("%s composite pk: %v", jt, c.Columns)
+				}
+			case models.ForeignKeyConstraint:
+				fks++
+			}
+		}
+		if pk != 1 || fks != 2 {
+			t.Errorf("%s: pk=%d fks=%d", jt, pk, fks)
+		}
+	}
+}
+
+func TestReadSchemaAndTable(t *testing.T) {
+	r := NewReader(&readers.ReaderOptions{FilePath: exampleTS})
+	s, err := r.ReadSchema()
+	if err != nil || s.Name != "public" {
+		t.Fatalf("schema: %v %+v", err, s)
+	}
+	tbl, err := r.ReadTable()
+	if err != nil || tbl.Name != "User" {
+		t.Fatalf("table: %v %+v", err, tbl)
+	}
+}
+
+func TestReader_Errors(t *testing.T) {
+	if _, err := NewReader(&readers.ReaderOptions{}).ReadDatabase(); err == nil || !strings.Contains(err.Error(), "file path is required") {
+		t.Errorf("empty path: %v", err)
+	}
+	if _, err := NewReader(&readers.ReaderOptions{FilePath: filepath.Join(t.TempDir(), "x.ts")}).ReadDatabase(); err == nil || !strings.Contains(err.Error(), "failed to read file") {
+		t.Errorf("missing file: %v", err)
+	}
+	if _, err := NewReader(&readers.ReaderOptions{}).ReadSchema(); err == nil {
+		t.Error("ReadSchema without path must fail")
+	}
+	if _, err := NewReader(&readers.ReaderOptions{}).ReadTable(); err == nil {
+		t.Error("ReadTable without path must fail")
+	}
+
+	empty := filepath.Join(t.TempDir(), "empty.ts")
+	if err := os.WriteFile(empty, []byte("// nothing here\n"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	r := NewReader(&readers.ReaderOptions{FilePath: empty})
+	if db, err := r.ReadDatabase(); err != nil || len(db.Schemas[0].Tables) != 0 {
+		t.Errorf("empty file: %v %+v", err, db)
+	}
+	if _, err := r.ReadTable(); err == nil || !strings.Contains(err.Error(), "no tables found") {
+		t.Errorf("ReadTable on empty: %v", err)
+	}
+}
+
+func TestEntityOptions(t *testing.T) {
+	s := parseSource(t, `
+@Entity({ name: "app_users", schema: "auth", database: "main", engine: "InnoDB" })
+export class User {
+  @PrimaryGeneratedColumn()
+  id: number;
+
+  @Column({ type: 'varchar', length: 100, nullable: true })
+  login: string;
+
+  @Column({ type: 'numeric', precision: 12, scale: 4 })
+  balance: number;
+
+  @Column({ type: 'boolean' })
+  active: boolean;
+}
+
+@Entity('legacy')
+export class Legacy {
+  @PrimaryGeneratedColumn('increment')
+  id: number;
+
+  @Column('jsonb')
+  payload: any;
+}
+`)
+	user := tableByName(s, "app_users")
+	if user == nil || user.Schema != "auth" {
+		t.Fatalf("tables: %+v", s.Tables)
+	}
+	if c := user.Columns["id"]; c == nil || !c.AutoIncrement || c.Type != "integer" {
+		t.Errorf("id: %+v", c)
+	}
+	if c := user.Columns["login"]; c == nil || c.Type != "varchar(100)" || c.Length != 100 || c.NotNull {
+		t.Errorf("login: %+v", c)
+	}
+	if c := user.Columns["balance"]; c == nil || c.Type != "numeric(12,4)" {
+		t.Errorf("balance: %+v", c)
+	}
+	if c := user.Columns["active"]; c == nil || c.Type != "boolean" {
+		t.Errorf("active: %+v", c)
+	}
+	if c := tableByName(s, "Legacy").Columns["payload"]; c == nil || c.Type != "jsonb" {
+		t.Errorf("payload: %+v", c)
+	}
+}
+
+func TestViewEntity(t *testing.T) {
+	s := parseSource(t, `
+@ViewEntity({
+  name: "active_users",
+  schema: "reporting",
+  expression: `+"`"+`SELECT id, email FROM users WHERE active`+"`"+`
+})
+export class ActiveUsers {
+  id: number;
+  email: string;
+}
+
+@ViewEntity({ expression: "SELECT 1" })
+export class OneView {
+  n: number;
+}
+`)
+	if len(s.Views) != 2 || len(s.Tables) != 0 {
+		t.Fatalf("views=%d tables=%d", len(s.Views), len(s.Tables))
+	}
+	v := s.Views[0]
+	if v.Name != "active_users" || v.Schema != "reporting" || !strings.Contains(v.Definition, "SELECT id, email FROM users") {
+		t.Errorf("view: %+v", v)
+	}
+	if c := v.Columns["email"]; c == nil || c.Type != "text" {
+		t.Errorf("view column: %+v", v.Columns)
+	}
+	if s.Views[1].Name != "OneView" || s.Views[1].Definition != "SELECT 1" {
+		t.Errorf("second view: %+v", s.Views[1])
+	}
+}
+
+func TestParseColumnDecorator_IdentityAndGenerated(t *testing.T) {
+	r := &Reader{}
+	tbl := models.InitTable("t", "public")
+
+	col := models.InitColumn("id", "t", "public")
+	r.parseColumnDecorator(`@PrimaryGeneratedColumn('identity', { generatedIdentity: 'ALWAYS' })`, col, tbl)
+	if !col.IsPrimaryKey || !col.Identity || !col.AutoIncrement {
+		t.Errorf("identity pk: %+v", col)
+	}
+
+	other := models.InitColumn("seq", "t", "public")
+	r.parseColumnDecorator(`@Generated('identity')`, other, tbl)
+	if !other.Identity || other.IdentityGeneration != "BY DEFAULT" {
+		t.Errorf("@Generated: %+v", other)
+	}
+	r.parseColumnDecorator(`@Generated('uuid')`, models.InitColumn("u", "t", "public"), tbl) // no-op, no panic
+
+	gen := models.InitColumn("full", "t", "public")
+	r.parseColumnOptions(`@Column({ type: 'text', generatedType: 'STORED', asExpression: 'a || \'x\'' })`, gen, tbl)
+	if !gen.Generated || !strings.Contains(gen.GenerationExpression, "a ||") {
+		t.Errorf("generated column: %+v", gen)
+	}
+}
+
+func TestParseGeneratedIdentity(t *testing.T) {
+	tests := []struct{ in, want string }{
+		{`{ generatedIdentity: 'ALWAYS' }`, "ALWAYS"},
+		{`{ generatedIdentity: 'BY DEFAULT' }`, "BY DEFAULT"},
+		{`no option`, "BY DEFAULT"},
+	}
+	for _, tt := range tests {
+		if got := parseGeneratedIdentity(tt.in); got != tt.want {
+			t.Errorf("%q = %q, want %q", tt.in, got, tt.want)
+		}
+	}
+}
+
+func TestUnescapeSingleQuoted(t *testing.T) {
+	tests := []struct{ in, want string }{
+		{"plain", "plain"}, {`it\'s`, "it's"}, {`a\\b`, `a\b`}, {`trailing\`, `trailing\`}, {"", ""},
+	}
+	for _, tt := range tests {
+		if got := unescapeSingleQuoted(tt.in); got != tt.want {
+			t.Errorf("unescape(%q) = %q, want %q", tt.in, got, tt.want)
+		}
+	}
+}
+
+func TestMatchDecorator(t *testing.T) {
+	tests := []struct {
+		line   string
+		want   string
+		wantOK bool
+	}{
+		{"@Entity()", "@Entity()", true},
+		{"@Column() name: string;", "@Column()", true},
+		{"@Column({ type: 'text' })", "@Column({ type: 'text' })", true},
+		{`@Column({ asExpression: 'f(a)' }) x: string;`, `@Column({ asExpression: 'f(a)' })`, true},
+		{"@Generated", "@Generated", true},
+		{"@Column({ unterminated", "@Column({ unterminated", true},
+		{"name: string;", "", false},
+		{"", "", false},
+	}
+	for _, tt := range tests {
+		got, ok := matchDecorator(tt.line)
+		if got != tt.want || ok != tt.wantOK {
+			t.Errorf("matchDecorator(%q) = (%q,%v), want (%q,%v)", tt.line, got, ok, tt.want, tt.wantOK)
+		}
+	}
+}
+
+func TestTypeScriptTypeToSQL(t *testing.T) {
+	r := &Reader{}
+	tests := []struct{ in, want string }{
+		{"string", "text"},
+		{"number", "integer"},
+		{"boolean", "boolean"},
+		{"Date", "timestamp"},
+		{"any", "jsonb"},
+		{"string[]", "text"},
+		{"string | null", "text"},
+		{"Unknown", "text"},
+	}
+	for _, tt := range tests {
+		if got := r.typeScriptTypeToSQL(tt.in); got != tt.want {
+			t.Errorf("%q = %q, want %q", tt.in, got, tt.want)
+		}
+	}
+}
+
+func TestIsRelationField(t *testing.T) {
+	r := &Reader{}
+	for _, d := range []string{"@ManyToOne(() => A)", "@OneToMany(() => A, a => a.b)", "@ManyToMany(() => A)", "@OneToOne(() => A)"} {
+		if !r.isRelationField(fieldInfo{decorators: []string{d}}) {
+			t.Errorf("%s should be a relation", d)
+		}
+	}
+	if r.isRelationField(fieldInfo{decorators: []string{"@Column()"}}) || r.isRelationField(fieldInfo{}) {
+		t.Error("non-relation misdetected")
+	}
+}
+
+func TestOneToOne_And_MultiLineDecorators(t *testing.T) {
+	s := parseSource(t, `
+@Entity()
+export class Profile {
+  @PrimaryGeneratedColumn()
+  id: number;
+
+  @Column({
+    type: 'varchar',
+    length: 50,
+    nullable: true,
+  })
+  bio: string;
+
+  @OneToOne(() => Account)
+  @JoinColumn()
+  account: Account;
+}
+
+@Entity()
+export class Account {
+  @PrimaryGeneratedColumn()
+  id: number;
+}
+`)
+	p := tableByName(s, "Profile")
+	if c := p.Columns["bio"]; c == nil || c.Type != "varchar(50)" || c.NotNull {
+		t.Errorf("multi-line @Column not parsed: %+v", c)
+	}
+}

pkg/sqltypes/sql_array_types_roundtrip_test.go

@@ -0,0 +1,207 @@
+package sqltypes
+
+import (
+	"database/sql/driver"
+	"encoding/json"
+	"encoding/xml"
+	"reflect"
+	"strings"
+	"testing"
+
+	"github.com/google/uuid"
+	"gopkg.in/yaml.v3"
+)
+
+// arrayPtr is the pointer-receiver surface shared by every nullable array type.
+type arrayPtr[T any] interface {
+	*T
+	Scan(any) error
+	UnmarshalJSON([]byte) error
+	UnmarshalYAML(*yaml.Node) error
+	UnmarshalXML(*xml.Decoder, xml.StartElement) error
+}
+
+// arrayValue is the value-receiver surface shared by every nullable array type.
+type arrayValue interface {
+	Value() (driver.Value, error)
+	MarshalJSON() ([]byte, error)
+	MarshalYAML() (any, error)
+	MarshalXML(*xml.Encoder, xml.StartElement) error
+}
+
+type wrapped[T any] struct {
+	XMLName xml.Name `yaml:"-" xml:"w"`
+	V       T        `yaml:"v" xml:"v"`
+}
+
+// arrayRoundTrip runs the full Scan/Value/JSON/YAML/XML contract for one array type.
+// badScan is a literal the type's Scan must reject ("" skips the check).
+func arrayRoundTrip[T any, P arrayPtr[T]](t *testing.T, sample, null T, badScan string) {
+	t.Helper()
+	sv, ok := any(sample).(arrayValue)
+	if !ok {
+		t.Fatalf("%T does not implement the array value surface", sample)
+	}
+	nv := any(null).(arrayValue)
+
+	t.Run("scan-value", func(t *testing.T) {
+		val, err := sv.Value()
+		if err != nil || val == nil {
+			t.Fatalf("Value: %v %v", val, err)
+		}
+		for _, in := range []any{val, []byte(val.(string))} {
+			var got T
+			if err := P(&got).Scan(in); err != nil {
+				t.Fatalf("Scan(%T): %v", in, err)
+			}
+			if !reflect.DeepEqual(got, sample) {
+				t.Errorf("Scan(%T) = %+v, want %+v", in, got, sample)
+			}
+		}
+		if v, err := nv.Value(); v != nil || err != nil {
+			t.Errorf("null Value = %v, %v", v, err)
+		}
+		got := sample
+		if err := P(&got).Scan(nil); err != nil || !reflect.DeepEqual(got, null) {
+			t.Errorf("Scan(nil) = %+v, %v", got, err)
+		}
+		if err := P(&got).Scan(12345); err == nil {
+			t.Error("Scan(int) must fail")
+		}
+		if badScan != "" {
+			var bad T
+			if err := P(&bad).Scan(badScan); err == nil {
+				t.Errorf("Scan(%q) must fail", badScan)
+			}
+		}
+	})
+
+	t.Run("json", func(t *testing.T) {
+		b, err := sv.MarshalJSON()
+		if err != nil {
+			t.Fatal(err)
+		}
+		var got T
+		if err := P(&got).UnmarshalJSON(b); err != nil || !reflect.DeepEqual(got, sample) {
+			t.Errorf("round trip = %+v, %v", got, err)
+		}
+		nb, _ := nv.MarshalJSON()
+		if string(nb) != "null" {
+			t.Errorf("null marshals to %s", nb)
+		}
+		got = sample
+		if err := P(&got).UnmarshalJSON([]byte(" null ")); err != nil || !reflect.DeepEqual(got, null) {
+			t.Errorf("null unmarshal = %+v, %v", got, err)
+		}
+		if err := P(&got).UnmarshalJSON([]byte(`{}`)); err == nil {
+			t.Error("object must be rejected")
+		}
+	})
+
+	t.Run("yaml", func(t *testing.T) {
+		b, err := yaml.Marshal(wrapped[T]{V: sample})
+		if err != nil {
+			t.Fatal(err)
+		}
+		var got wrapped[T]
+		if err := yaml.Unmarshal(b, &got); err != nil || !reflect.DeepEqual(got.V, sample) {
+			t.Errorf("round trip = %+v, %v\n%s", got.V, err, b)
+		}
+		nb, err := yaml.Marshal(wrapped[T]{V: null})
+		if err != nil || !strings.Contains(string(nb), "null") {
+			t.Errorf("null marshal = %q, %v", nb, err)
+		}
+		// yaml.v3 skips UnmarshalYAML for null, so decode into a fresh value.
+		got = wrapped[T]{}
+		if err := yaml.Unmarshal(nb, &got); err != nil || !reflect.DeepEqual(got.V, null) {
+			t.Errorf("null unmarshal = %+v, %v", got.V, err)
+		}
+		var bad wrapped[T]
+		if err := yaml.Unmarshal([]byte("v: {a: b}\n"), &bad); err == nil {
+			t.Error("mapping must be rejected")
+		}
+	})
+
+	t.Run("xml", func(t *testing.T) {
+		b, err := xml.Marshal(wrapped[T]{V: sample})
+		if err != nil {
+			t.Fatal(err)
+		}
+		var got wrapped[T]
+		if err := xml.Unmarshal(b, &got); err != nil || !reflect.DeepEqual(got.V, sample) {
+			t.Errorf("round trip = %+v, %v\n%s", got.V, err, b)
+		}
+		if _, err := xml.Marshal(wrapped[T]{V: null}); err != nil {
+			t.Errorf("null marshal: %v", err)
+		}
+		var bad wrapped[T]
+		if err := xml.Unmarshal([]byte("<w><v><item>1</item>"), &bad); err == nil {
+			t.Error("truncated xml must fail")
+		}
+	})
+}
+
+func TestArrayTypes_FullContract(t *testing.T) {
+	u1, u2 := uuid.New(), uuid.New()
+	t.Run("string", func(t *testing.T) {
+		arrayRoundTrip(t, NewSqlStringArray([]string{"a", "b c", `q"uote`, "x,y"}), SqlStringArray{}, "")
+	})
+	t.Run("int16", func(t *testing.T) {
+		arrayRoundTrip(t, NewSqlInt16Array([]int16{1, -2, 300}), SqlInt16Array{}, "{99999}")
+	})
+	t.Run("int32", func(t *testing.T) {
+		arrayRoundTrip(t, NewSqlInt32Array([]int32{1, -2, 300000}), SqlInt32Array{}, "{x}")
+	})
+	t.Run("int64", func(t *testing.T) {
+		arrayRoundTrip(t, NewSqlInt64Array([]int64{1, -2, 1 << 40}), SqlInt64Array{}, "{x}")
+	})
+	t.Run("float32", func(t *testing.T) {
+		arrayRoundTrip(t, NewSqlFloat32Array([]float32{1.5, -2.25}), SqlFloat32Array{}, "{x}")
+	})
+	t.Run("float64", func(t *testing.T) {
+		arrayRoundTrip(t, NewSqlFloat64Array([]float64{1.5, -2.25, 1e10}), SqlFloat64Array{}, "{x}")
+	})
+	t.Run("bool", func(t *testing.T) {
+		arrayRoundTrip(t, NewSqlBoolArray([]bool{true, false, true}), SqlBoolArray{}, "not an array")
+	})
+	t.Run("uuid", func(t *testing.T) {
+		arrayRoundTrip(t, NewSqlUUIDArray([]uuid.UUID{u1, u2}), SqlUUIDArray{}, "{not-a-uuid}")
+	})
+	t.Run("vector", func(t *testing.T) {
+		arrayRoundTrip(t, NewSqlVector([]float32{1, 2.5, -3}), SqlVector{}, "1,2,3")
+	})
+}
+
+func TestArrayTypes_EmptyAndMalformedScan(t *testing.T) {
+	var s SqlStringArray
+	if err := s.Scan("{}"); err != nil || !s.Valid || len(s.Val) != 0 {
+		t.Errorf("empty array: %+v %v", s, err)
+	}
+	var i SqlInt32Array
+	if err := i.Scan("{}"); err != nil || !i.Valid || len(i.Val) != 0 {
+		t.Errorf("empty int array: %+v %v", i, err)
+	}
+	var v SqlVector
+	if err := v.Scan("[]"); err != nil || !v.Valid || len(v.Val) != 0 {
+		t.Errorf("empty vector: %+v %v", v, err)
+	}
+	if err := v.Scan("[1,x]"); err == nil {
+		t.Error("bad vector element must fail")
+	}
+	if err := v.Scan(42); err == nil {
+		t.Error("vector Scan(int) must fail")
+	}
+	for _, bad := range []string{"not an array", "{unterminated"} {
+		var a SqlInt32Array
+		if err := a.Scan(bad); err == nil {
+			t.Errorf("Scan(%q) must fail", bad)
+		}
+	}
+}
+
+func TestArrayJSONIsPlainSlice(t *testing.T) {
+	b, err := json.Marshal(NewSqlInt32Array([]int32{1, 2}))
+	if err != nil || string(b) != "[1,2]" {
+		t.Errorf("got %s, %v", b, err)
+	}
+}

pkg/sqltypes/sql_types_scalar_test.go

@@ -0,0 +1,157 @@
+package sqltypes
+
+import (
+	"database/sql/driver"
+	"encoding/json"
+	"math"
+	"testing"
+	"time"
+)
+
+func TestSqlNull_ValueScalarCases(t *testing.T) {
+	tests := []struct {
+		name  string
+		input SqlNull[any]
+		want  driver.Value
+	}{
+		{name: "invalid", input: SqlNull[any]{}, want: nil},
+		{name: "integer", input: Null[any](int64(42), true), want: int64(42)},
+		{name: "string", input: Null[any]("hello", true), want: "hello"},
+		{name: "boolean", input: Null[any](true, true), want: true},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got, err := tt.input.Value()
+			if err != nil {
+				t.Fatalf("Value returned error: %v", err)
+			}
+			if got != tt.want {
+				t.Errorf("Value() = %v (%T), want %v (%T)", got, got, tt.want, tt.want)
+			}
+		})
+	}
+}
+
+func TestSqlNull_Int64Conversions(t *testing.T) {
+	tests := []struct {
+		name  string
+		input SqlNull[any]
+		want  int64
+	}{
+		{name: "invalid", input: SqlNull[any]{}, want: 0},
+		{name: "signed integer", input: Null[any](int32(-12), true), want: -12},
+		{name: "unsigned integer", input: Null[any](uint16(12), true), want: 12},
+		{name: "float truncates", input: Null[any](float64(12.9), true), want: 12},
+		{name: "numeric string", input: Null[any]("123", true), want: 123},
+		{name: "invalid string", input: Null[any]("not a number", true), want: 0},
+		{name: "true", input: Null[any](true, true), want: 1},
+		{name: "false", input: Null[any](false, true), want: 0},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := tt.input.Int64(); got != tt.want {
+				t.Errorf("Int64() = %d, want %d", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestSqlNull_Float64Conversions(t *testing.T) {
+	tests := []struct {
+		name  string
+		input SqlNull[any]
+		want  float64
+	}{
+		{name: "invalid", input: SqlNull[any]{}, want: 0},
+		{name: "float", input: Null[any](float32(1.25), true), want: 1.25},
+		{name: "signed integer", input: Null[any](int64(-12), true), want: -12},
+		{name: "unsigned integer", input: Null[any](uint16(12), true), want: 12},
+		{name: "numeric string", input: Null[any]("12.5", true), want: 12.5},
+		{name: "invalid string", input: Null[any]("not a number", true), want: 0},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := tt.input.Float64(); got != tt.want {
+				t.Errorf("Float64() = %v, want %v", got, tt.want)
+			}
+		})
+	}
+}
+
+func TestSqlDate_JSONNullAndInvalid(t *testing.T) {
+	tests := []struct {
+		name  string
+		json  string
+		valid bool
+	}{
+		{name: "null", json: "null", valid: false},
+		{name: "invalid date", json: `"not-a-date"`, valid: false},
+		{name: "valid date", json: `"2024-01-15"`, valid: true},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			var got SqlDate
+			if err := json.Unmarshal([]byte(tt.json), &got); err != nil {
+				t.Fatalf("UnmarshalJSON returned error: %v", err)
+			}
+			if got.Valid != tt.valid {
+				t.Errorf("Valid = %v, want %v", got.Valid, tt.valid)
+			}
+		})
+	}
+
+	if data, err := json.Marshal(SqlDate{}); err != nil {
+		t.Fatalf("MarshalJSON returned error: %v", err)
+	} else if string(data) != "null" {
+		t.Errorf("MarshalJSON() = %s, want null", data)
+	}
+}
+
+func TestSqlTypeNowConstructors(t *testing.T) {
+	before := time.Now()
+	timestamp := SqlTimeStampNow()
+	date := SqlDateNow()
+	tm := SqlTimeNow()
+	after := time.Now()
+
+	for name, got := range map[string]time.Time{
+		"timestamp": timestamp.Time(),
+		"date":      date.Time(),
+		"time":      tm.Time(),
+	} {
+		if !got.After(before) && !got.Equal(before) || got.After(after) {
+			t.Errorf("%s constructor returned %v outside [%v, %v]", name, got, before, after)
+		}
+	}
+	if !timestamp.Valid || !date.Valid || !tm.Valid {
+		t.Fatal("Now constructors must return valid values")
+	}
+}
+
+func TestNewSqlAndToJSONDT(t *testing.T) {
+	if got := NewSql[int64]("42"); !got.Valid || got.Val != 42 {
+		t.Errorf("NewSql[int64](\"42\") = %#v, want valid 42", got)
+	}
+	if got := NewSql[int64](nil); got.Valid {
+		t.Errorf("NewSql[int64](nil) = %#v, want invalid", got)
+	}
+	if got := NewSqlFloat32(1.5); !got.Valid || got.Val != 1.5 {
+		t.Errorf("NewSqlFloat32(1.5) = %#v, want valid 1.5", got)
+	}
+
+	when := time.Date(2024, 1, 15, 10, 30, 45, 0, time.UTC)
+	if got := ToJSONDT(when); got != "2024-01-15T10:30:45Z" {
+		t.Errorf("ToJSONDT() = %q, want RFC3339 timestamp", got)
+	}
+}
+
+func TestSqlNull_Float64PreservesInfinity(t *testing.T) {
+	got := Null[float64](math.Inf(1), true).Float64()
+	if !math.IsInf(got, 1) {
+		t.Errorf("Float64() = %v, want +Inf", got)
+	}
+}

pkg/transform/transform_test.go

@@ -0,0 +1,38 @@
+package transform
+
+import (
+	"testing"
+
+	"git.warky.dev/wdevs/relspecgo/pkg/models"
+)
+
+// Validation and normalization are currently pass-through stubs; these tests
+// pin that contract (no error, input returned unchanged).
+func TestTransformerStubs(t *testing.T) {
+	tr := NewTransformer()
+	if tr == nil {
+		t.Fatal("nil transformer")
+	}
+	db := models.InitDatabase("d")
+	schema := models.InitSchema("public")
+	table := models.InitTable("t", "public")
+
+	if err := tr.ValidateDatabase(db); err != nil {
+		t.Error(err)
+	}
+	if err := tr.ValidateSchema(schema); err != nil {
+		t.Error(err)
+	}
+	if err := tr.ValidateTable(table); err != nil {
+		t.Error(err)
+	}
+	if got, err := tr.NormalizeDatabase(db); err != nil || got != db {
+		t.Errorf("NormalizeDatabase = %v, %v", got, err)
+	}
+	if got, err := tr.NormalizeSchema(schema); err != nil || got != schema {
+		t.Errorf("NormalizeSchema = %v, %v", got, err)
+	}
+	if got, err := tr.NormalizeTable(table); err != nil || got != table {
+		t.Errorf("NormalizeTable = %v, %v", got, err)
+	}
+}

pkg/ui/connstring.go

@@ -0,0 +1,186 @@
+package ui
+
+import (
+	"fmt"
+	"net"
+	"net/url"
+	"strings"
+)
+
+// ConnKind identifies the database type a connection string targets.
+type ConnKind string
+
+const (
+	ConnPostgres ConnKind = "postgres"
+	ConnMSSQL    ConnKind = "mssql"
+	ConnSQLite   ConnKind = "sqlite"
+)
+
+// connKinds lists the kinds offered by the builder dialog, in display order.
+var connKinds = []ConnKind{ConnPostgres, ConnMSSQL, ConnSQLite}
+
+// maskedPassword is substituted for the password in previews.
+const maskedPassword = "****"
+
+// ConnFields holds the editable parts of a connection string.
+type ConnFields struct {
+	Kind     ConnKind
+	Host     string
+	Port     string
+	Database string
+	User     string
+	Password string
+	SSLMode  string
+	FilePath string // SQLite only
+
+	// Extra keeps query parameters the builder has no field for, so that
+	// parsing and rebuilding an existing string does not drop them.
+	Extra url.Values
+}
+
+// DefaultConnFields returns sensible defaults for the given kind.
+func DefaultConnFields(kind ConnKind) ConnFields {
+	f := ConnFields{Kind: kind}
+	switch kind {
+	case ConnPostgres:
+		f.Host, f.Port, f.User, f.SSLMode = "localhost", "5432", "postgres", "disable"
+	case ConnMSSQL:
+		f.Host, f.Port, f.User, f.SSLMode = "localhost", "1433", "sa", "disable"
+	}
+	return f
+}
+
+// SSLModes returns the valid SSL/encryption options for a kind.
+func SSLModes(kind ConnKind) []string {
+	switch kind {
+	case ConnPostgres:
+		return []string{"disable", "allow", "prefer", "require", "verify-ca", "verify-full"}
+	case ConnMSSQL:
+		return []string{"disable", "false", "true"}
+	}
+	return nil
+}
+
+func (f ConnFields) sslParam() string {
+	if f.Kind == ConnMSSQL {
+		return "encrypt"
+	}
+	return "sslmode"
+}
+
+// BuildConnString renders the fields as a connection string. With mask set,
+// a non-empty password is replaced by asterisks (for previews).
+func BuildConnString(f ConnFields, mask bool) string {
+	if f.Kind == ConnSQLite {
+		return f.FilePath
+	}
+
+	u := &url.URL{Scheme: "postgres"}
+	if f.Kind == ConnMSSQL {
+		u.Scheme = "sqlserver"
+	}
+
+	if f.Port != "" {
+		u.Host = net.JoinHostPort(f.Host, f.Port)
+	} else {
+		u.Host = f.Host
+	}
+
+	if f.User != "" {
+		if f.Password != "" {
+			pw := f.Password
+			if mask {
+				pw = maskedPassword
+			}
+			u.User = url.UserPassword(f.User, pw)
+		} else {
+			u.User = url.User(f.User)
+		}
+	}
+
+	query := url.Values{}
+	for k, v := range f.Extra {
+		query[k] = v
+	}
+	if f.Kind == ConnMSSQL {
+		if f.Database != "" {
+			query.Set("database", f.Database)
+		}
+	} else if f.Database != "" {
+		u.Path = "/" + f.Database
+	}
+	if f.SSLMode != "" {
+		query.Set(f.sslParam(), f.SSLMode)
+	}
+	u.RawQuery = query.Encode()
+
+	out := u.String()
+	if mask {
+		// url escapes '*' in the userinfo; keep the preview readable.
+		out = strings.Replace(out, url.QueryEscape(maskedPassword), maskedPassword, 1)
+	}
+	return out
+}
+
+// DetectConnKind guesses the kind from a connection string's scheme. Anything
+// that is not a recognised URL is treated as a SQLite file path.
+func DetectConnKind(s string) ConnKind {
+	lower := strings.ToLower(strings.TrimSpace(s))
+	switch {
+	case strings.HasPrefix(lower, "postgres://"), strings.HasPrefix(lower, "postgresql://"):
+		return ConnPostgres
+	case strings.HasPrefix(lower, "sqlserver://"), strings.HasPrefix(lower, "mssql://"):
+		return ConnMSSQL
+	}
+	return ConnSQLite
+}
+
+// ParseConnString splits a connection string into fields. An empty string
+// yields the defaults for hint. Missing ports fall back to the kind default.
+func ParseConnString(s string, hint ConnKind) (ConnFields, error) {
+	s = strings.TrimSpace(s)
+	if s == "" {
+		return DefaultConnFields(hint), nil
+	}
+
+	kind := DetectConnKind(s)
+	if kind == ConnSQLite {
+		path := s
+		for _, prefix := range []string{"sqlite://", "sqlite3://"} {
+			path = strings.TrimPrefix(path, prefix)
+		}
+		return ConnFields{Kind: ConnSQLite, FilePath: path}, nil
+	}
+
+	u, err := url.Parse(s)
+	if err != nil {
+		return DefaultConnFields(kind), fmt.Errorf("invalid connection string: %w", err)
+	}
+
+	f := ConnFields{
+		Kind: kind,
+		Host: u.Hostname(),
+		Port: u.Port(),
+	}
+	if f.Port == "" {
+		f.Port = DefaultConnFields(kind).Port
+	}
+	if u.User != nil {
+		f.User = u.User.Username()
+		f.Password, _ = u.User.Password()
+	}
+
+	query := u.Query()
+	if kind == ConnMSSQL {
+		f.Database = query.Get("database")
+		query.Del("database")
+	} else {
+		f.Database = strings.TrimPrefix(u.Path, "/")
+	}
+	f.SSLMode = query.Get(f.sslParam())
+	query.Del(f.sslParam())
+	if len(query) > 0 {
+		f.Extra = query
+	}
+	return f, nil
+}

pkg/ui/connstring_check.go

@@ -0,0 +1,62 @@
+package ui
+
+import (
+	"context"
+	"database/sql"
+	"fmt"
+	"os"
+	"strings"
+	"time"
+
+	"github.com/jackc/pgx/v5"
+	_ "github.com/microsoft/go-mssqldb"
+	_ "modernc.org/sqlite"
+)
+
+// connTestTimeout bounds how long "Test connection" may block.
+const connTestTimeout = 5 * time.Second
+
+// TestConnection opens and pings the database described by f. Any occurrence
+// of the password in the returned error is masked.
+func TestConnection(f ConnFields) error {
+	ctx, cancel := context.WithTimeout(context.Background(), connTestTimeout)
+	defer cancel()
+
+	err := testConnection(ctx, f)
+	if err != nil && f.Password != "" {
+		err = fmt.Errorf("%s", strings.ReplaceAll(err.Error(), f.Password, maskedPassword))
+	}
+	return err
+}
+
+func testConnection(ctx context.Context, f ConnFields) error {
+	switch f.Kind {
+	case ConnPostgres:
+		conn, err := pgx.Connect(ctx, BuildConnString(f, false))
+		if err != nil {
+			return err
+		}
+		return conn.Close(ctx)
+	case ConnMSSQL:
+		return pingSQL(ctx, "sqlserver", BuildConnString(f, false))
+	case ConnSQLite:
+		if f.FilePath == "" {
+			return fmt.Errorf("file path is required")
+		}
+		// Opening a missing SQLite file would silently create it.
+		if _, err := os.Stat(f.FilePath); err != nil {
+			return err
+		}
+		return pingSQL(ctx, "sqlite", f.FilePath)
+	}
+	return fmt.Errorf("unsupported connection type %q", f.Kind)
+}
+
+func pingSQL(ctx context.Context, driver, dsn string) error {
+	db, err := sql.Open(driver, dsn)
+	if err != nil {
+		return err
+	}
+	defer db.Close()
+	return db.PingContext(ctx)
+}

pkg/ui/connstring_screens.go

@@ -0,0 +1,210 @@
+package ui
+
+import (
+	"fmt"
+	"strings"
+
+	"github.com/gdamore/tcell/v2"
+	"github.com/rivo/tview"
+)
+
+// connBuilderPage is the page name of the connection string builder dialog.
+const connBuilderPage = "conn-builder"
+
+// showConnStringBuilder opens the connection string builder, pre-filled by
+// parsing current. Save calls onDone with the built string; Esc/Back leaves
+// the caller's input untouched.
+func (se *SchemaEditor) showConnStringBuilder(current string, hint ConnKind, returnPage string, onDone func(connString string)) {
+	fields, err := ParseConnString(current, hint)
+	if err != nil {
+		se.showErrorDialog("Error", err.Error()+"\nStarting from defaults.")
+	}
+
+	title := tview.NewTextView().
+		SetText("[::b]Connection String Builder").
+		SetTextAlign(tview.AlignCenter).
+		SetDynamicColors(true)
+
+	preview := tview.NewTextView()
+	preview.SetBorder(true).SetTitle(" Preview (password masked) ").SetTitleAlign(tview.AlignLeft)
+
+	form := tview.NewForm()
+	form.SetBorder(true).SetTitle(" Connection ").SetTitleAlign(tview.AlignLeft)
+
+	updatePreview := func() {
+		preview.SetText(tview.Escape(BuildConnString(fields, true)))
+	}
+
+	closeBuilder := func() {
+		se.pages.RemovePage(connBuilderPage)
+		se.pages.SwitchToPage(returnPage)
+	}
+
+	var render func(focus int)
+	render = func(focus int) {
+		form.Clear(false)
+
+		kindIndex := 0
+		kindLabels := make([]string, len(connKinds))
+		for i, k := range connKinds {
+			kindLabels[i] = string(k)
+			if k == fields.Kind {
+				kindIndex = i
+			}
+		}
+		form.AddDropDown("Type", kindLabels, kindIndex, func(_ string, index int) {
+			if connKinds[index] == fields.Kind {
+				return
+			}
+			fields = DefaultConnFields(connKinds[index])
+			render(0)
+		})
+
+		if fields.Kind == ConnSQLite {
+			form.AddInputField("File Path", fields.FilePath, 50, nil, func(v string) {
+				fields.FilePath = v
+				updatePreview()
+			})
+			if item, ok := form.GetFormItemByLabel("File Path").(*tview.InputField); ok {
+				item.SetInputCapture(func(event *tcell.EventKey) *tcell.EventKey {
+					if event.Key() != tcell.KeyEnter {
+						return event
+					}
+					se.showFileBrowser(FileBrowserConfig{
+						Mode:       FileBrowserLoad,
+						StartPath:  fields.FilePath,
+						Extensions: FormatExtensions("sqlite"),
+						ReturnPage: connBuilderPage,
+						OnSelect:   func(path string) { item.SetText(path) },
+					})
+					return nil
+				})
+			}
+		} else {
+			form.AddInputField("Host", fields.Host, 50, nil, func(v string) { fields.Host = v; updatePreview() })
+			form.AddInputField("Port", fields.Port, 10, tview.InputFieldInteger, func(v string) { fields.Port = v; updatePreview() })
+			form.AddInputField("Database", fields.Database, 50, nil, func(v string) { fields.Database = v; updatePreview() })
+			form.AddInputField("User", fields.User, 50, nil, func(v string) { fields.User = v; updatePreview() })
+			form.AddPasswordField("Password", fields.Password, 50, '*', func(v string) { fields.Password = v; updatePreview() })
+
+			label := "SSL Mode"
+			if fields.Kind == ConnMSSQL {
+				label = "Encrypt"
+			}
+			modes := SSLModes(fields.Kind)
+			modeIndex := -1
+			for i, m := range modes {
+				if m == fields.SSLMode {
+					modeIndex = i
+				}
+			}
+			if modeIndex < 0 {
+				// Keep a value parsed from an existing string even if it is not a listed option.
+				modes = append([]string{fields.SSLMode}, modes...)
+				modeIndex = 0
+			}
+			form.AddDropDown(label, modes, modeIndex, func(option string, _ int) {
+				fields.SSLMode = option
+				updatePreview()
+			})
+		}
+
+		form.AddButton("Save [F2]", connBuilderSave(se, &fields, closeBuilder, onDone))
+		form.AddButton("Test [F3]", func() { se.testConnectionDialog(fields) })
+		form.AddButton("Back [Esc]", closeBuilder)
+
+		updatePreview()
+		form.SetFocus(focus)
+		se.app.SetFocus(form)
+	}
+
+	form.SetInputCapture(func(event *tcell.EventKey) *tcell.EventKey {
+		switch event.Key() {
+		case tcell.KeyEscape:
+			closeBuilder()
+			return nil
+		case tcell.KeyF2:
+			connBuilderSave(se, &fields, closeBuilder, onDone)()
+			return nil
+		case tcell.KeyF3:
+			se.testConnectionDialog(fields)
+			return nil
+		}
+		return event
+	})
+
+	render(0)
+
+	flex := tview.NewFlex().SetDirection(tview.FlexRow).
+		AddItem(title, 1, 0, false).
+		AddItem(form, 0, 1, true).
+		AddItem(preview, 4, 0, false)
+
+	se.pages.AddAndSwitchToPage(connBuilderPage, flex, true)
+	se.app.SetFocus(form)
+}
+
+// connBuilderSave returns the Save action: validate, write back, close.
+func connBuilderSave(se *SchemaEditor, fields *ConnFields, closeBuilder func(), onDone func(string)) func() {
+	return func() {
+		if msg := validateConnFields(*fields); msg != "" {
+			se.showErrorDialog("Error", msg)
+			return
+		}
+		result := BuildConnString(*fields, false)
+		closeBuilder()
+		onDone(result)
+	}
+}
+
+// validateConnFields returns a message describing the first missing required field, or "".
+func validateConnFields(f ConnFields) string {
+	if f.Kind == ConnSQLite {
+		if strings.TrimSpace(f.FilePath) == "" {
+			return "File path is required"
+		}
+		return ""
+	}
+	if strings.TrimSpace(f.Host) == "" {
+		return "Host is required"
+	}
+	return ""
+}
+
+// testConnectionDialog runs TestConnection in the background and reports the result.
+func (se *SchemaEditor) testConnectionDialog(fields ConnFields) {
+	if msg := validateConnFields(fields); msg != "" {
+		se.showErrorDialog("Error", msg)
+		return
+	}
+
+	go func() {
+		err := TestConnection(fields)
+		se.app.QueueUpdateDraw(func() {
+			if err != nil {
+				se.showErrorDialog("Connection Failed", fmt.Sprintf("Connection failed:\n%v", err))
+				return
+			}
+			se.showSuccessDialog("Connection OK", "Connection successful", nil)
+		})
+	}()
+}
+
+// attachConnStringBuilder makes Enter on the named input open the builder.
+func (se *SchemaEditor) attachConnStringBuilder(form *tview.Form, label, returnPage string, format func() string) {
+	item, ok := form.GetFormItemByLabel(label).(*tview.InputField)
+	if !ok {
+		return
+	}
+	item.SetInputCapture(func(event *tcell.EventKey) *tcell.EventKey {
+		if event.Key() != tcell.KeyEnter {
+			return event
+		}
+		hint := ConnPostgres
+		if format != nil && format() == "sqlite" {
+			hint = ConnSQLite
+		}
+		se.showConnStringBuilder(item.GetText(), hint, returnPage, func(s string) { item.SetText(s) })
+		return nil
+	})
+}