Merge pull request 'feat(scripts): support external file embedding' (#12) from issue-6-external-file-embedding into master

Reviewed-on: #12 Reviewed-by: Warky <2+warkanum@noreply@warky.dev>
2026-07-20 11:09:39 +00:00
parent 5c31deb630 1bcdf29206
commit 1c217b546c
9 changed files with 499 additions and 24 deletions
@@ -0,0 +1,165 @@
+package assetloader
+
+import (
+	"encoding/base64"
+	"fmt"
+	"os"
+	"path/filepath"
+	"regexp"
+	"strconv"
+	"strings"
+	"unicode/utf8"
+)
+
+const ScriptSourcePathMetadataKey = "source_path"
+
+var (
+	embedDirectivePattern = regexp.MustCompile(`(?m)^\s*--\s*@embed:\s*(.+?)\s*$`)
+	embedAttrPattern      = regexp.MustCompile(`([a-zA-Z_][a-zA-Z0-9_]*)=("[^"]*"|'[^']*'|\S+)`)
+	embedVarPattern       = regexp.MustCompile(`^:[a-zA-Z_][a-zA-Z0-9_]*$`)
+)
+
+// ProcessEmbedDirectives expands SQL comments in the form:
+//
+//	-- @embed: path=... var=:... mode=text|base64
+//
+// Paths are resolved relative to sqlPath. Text mode embeds a quoted UTF-8 SQL
+// string literal. Base64 mode embeds a quoted base64 literal suitable for
+// decode(:var, 'base64').
+func ProcessEmbedDirectives(sqlPath, sql string) (string, error) {
+	directives := embedDirectivePattern.FindAllStringSubmatch(sql, -1)
+	if len(directives) == 0 {
+		return sql, nil
+	}
+
+	if sqlPath == "" {
+		return "", fmt.Errorf("sql path is required for embed directives")
+	}
+
+	result := embedDirectivePattern.ReplaceAllString(sql, "")
+	for i, directive := range directives {
+		literal, placeholder, err := embedDirectiveLiteral(sqlPath, directive[1], i+1)
+		if err != nil {
+			return "", err
+		}
+		if !embedPlaceholderPattern(placeholder).MatchString(result) {
+			return "", fmt.Errorf("%s embed directive %d: placeholder %s not found", sqlPath, i+1, placeholder)
+		}
+		result = replaceEmbedPlaceholder(result, placeholder, literal)
+	}
+
+	return result, nil
+}
+
+func embedDirectiveLiteral(sqlPath, raw string, directiveNumber int) (string, string, error) {
+	attrs, err := parseEmbedAttrs(raw)
+	if err != nil {
+		return "", "", fmt.Errorf("%s embed directive %d: %w", sqlPath, directiveNumber, err)
+	}
+
+	pathValue := attrs["path"]
+	varValue := attrs["var"]
+	modeValue := attrs["mode"]
+	if pathValue == "" {
+		return "", "", fmt.Errorf("%s embed directive %d: missing path", sqlPath, directiveNumber)
+	}
+	if !embedVarPattern.MatchString(varValue) {
+		return "", "", fmt.Errorf("%s embed directive %d: var must be a named placeholder like :asset", sqlPath, directiveNumber)
+	}
+	if modeValue != "text" && modeValue != "base64" {
+		return "", "", fmt.Errorf("%s embed directive %d: mode must be text or base64", sqlPath, directiveNumber)
+	}
+
+	resolved := filepath.Join(filepath.Dir(sqlPath), filepath.Clean(pathValue))
+	data, err := os.ReadFile(resolved)
+	if err != nil {
+		return "", "", fmt.Errorf("%s embed directive %d: reading %s: %w", sqlPath, directiveNumber, resolved, err)
+	}
+
+	switch modeValue {
+	case "text":
+		if !utf8.Valid(data) {
+			return "", "", fmt.Errorf("%s embed directive %d: %s is not valid UTF-8", sqlPath, directiveNumber, resolved)
+		}
+		literal, err := sqlStringLiteral(string(data))
+		if err != nil {
+			return "", "", fmt.Errorf("%s embed directive %d: %w", sqlPath, directiveNumber, err)
+		}
+		return literal, varValue, nil
+	case "base64":
+		literal, err := sqlStringLiteral(base64.StdEncoding.EncodeToString(data))
+		if err != nil {
+			return "", "", fmt.Errorf("%s embed directive %d: %w", sqlPath, directiveNumber, err)
+		}
+		return literal, varValue, nil
+	default:
+		return "", "", fmt.Errorf("%s embed directive %d: mode must be text or base64", sqlPath, directiveNumber)
+	}
+}
+
+func parseEmbedAttrs(raw string) (map[string]string, error) {
+	attrs := map[string]string{}
+	matches := embedAttrPattern.FindAllStringSubmatchIndex(raw, -1)
+	if len(matches) == 0 {
+		return nil, fmt.Errorf("expected path, var, and mode attributes")
+	}
+
+	lastEnd := 0
+	for _, match := range matches {
+		gap := strings.TrimSpace(raw[lastEnd:match[0]])
+		if gap != "" {
+			return nil, fmt.Errorf("invalid attribute syntax near %q", gap)
+		}
+
+		key := raw[match[2]:match[3]]
+		value := raw[match[4]:match[5]]
+		if _, exists := attrs[key]; exists {
+			return nil, fmt.Errorf("duplicate attribute %q", key)
+		}
+		unquoted, err := unquoteEmbedValue(value)
+		if err != nil {
+			return nil, fmt.Errorf("invalid %s value: %w", key, err)
+		}
+		attrs[key] = unquoted
+		lastEnd = match[1]
+	}
+	if tail := strings.TrimSpace(raw[lastEnd:]); tail != "" {
+		return nil, fmt.Errorf("invalid attribute syntax near %q", tail)
+	}
+
+	for key := range attrs {
+		if key != "path" && key != "var" && key != "mode" {
+			return nil, fmt.Errorf("unknown attribute %q", key)
+		}
+	}
+
+	return attrs, nil
+}
+
+func unquoteEmbedValue(value string) (string, error) {
+	if len(value) < 2 {
+		return value, nil
+	}
+	if value[0] == '"' {
+		return strconv.Unquote(value)
+	}
+	if value[0] == '\'' && value[len(value)-1] == '\'' {
+		return value[1 : len(value)-1], nil
+	}
+	return value, nil
+}
+
+func sqlStringLiteral(value string) (string, error) {
+	if strings.ContainsRune(value, '\x00') {
+		return "", fmt.Errorf("embedded text contains NUL byte")
+	}
+	return "'" + strings.ReplaceAll(value, "'", "''") + "'", nil
+}
+
+func replaceEmbedPlaceholder(sql, placeholder, literal string) string {
+	return embedPlaceholderPattern(placeholder).ReplaceAllString(sql, "${1}"+literal+"${2}")
+}
+
+func embedPlaceholderPattern(placeholder string) *regexp.Regexp {
+	return regexp.MustCompile(`(^|[^a-zA-Z0-9_:])` + regexp.QuoteMeta(placeholder) + `([^a-zA-Z0-9_]|$)`)
+}
@@ -0,0 +1,143 @@
+package assetloader
+
+import (
+	"encoding/base64"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+)
+
+func TestProcessEmbedDirectives_TextLiteralEscapesQuotes(t *testing.T) {
+	dir := t.TempDir()
+	sqlPath := filepath.Join(dir, "1_001_seed.sql")
+	if err := os.WriteFile(filepath.Join(dir, "body.txt"), []byte("Line 1\nIt's fine"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	got, err := ProcessEmbedDirectives(sqlPath, `
+-- @embed: path=body.txt var=:body mode=text
+INSERT INTO notes (body) VALUES (:body);
+`)
+	if err != nil {
+		t.Fatalf("ProcessEmbedDirectives failed: %v", err)
+	}
+	if !strings.Contains(got, "VALUES ('Line 1\nIt''s fine');") {
+		t.Fatalf("embedded SQL did not contain escaped text literal:\n%s", got)
+	}
+	if strings.Contains(got, "VALUES (:body);") {
+		t.Fatalf("placeholder was not replaced:\n%s", got)
+	}
+}
+
+func TestProcessEmbedDirectives_Base64Literal(t *testing.T) {
+	dir := t.TempDir()
+	sqlPath := filepath.Join(dir, "1_001_seed.sql")
+	binary := []byte{0x00, 0xff, 0x10, 0x20}
+	if err := os.WriteFile(filepath.Join(dir, "blob.bin"), binary, 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	got, err := ProcessEmbedDirectives(sqlPath, `
+-- @embed: path=blob.bin var=:payload mode=base64
+INSERT INTO files (payload) VALUES (decode(:payload, 'base64')::bytea);
+`)
+	if err != nil {
+		t.Fatalf("ProcessEmbedDirectives failed: %v", err)
+	}
+	want := "decode('" + base64.StdEncoding.EncodeToString(binary) + "', 'base64')::bytea"
+	if !strings.Contains(got, want) {
+		t.Fatalf("embedded SQL did not contain base64 literal %q:\n%s", want, got)
+	}
+}
+
+func TestProcessEmbedDirectives_RelativeToSQLFile(t *testing.T) {
+	root := t.TempDir()
+	sqlDir := filepath.Join(root, "nested", "seed")
+	if err := os.MkdirAll(filepath.Join(sqlDir, "assets"), 0o755); err != nil {
+		t.Fatal(err)
+	}
+	sqlPath := filepath.Join(sqlDir, "1_001_seed.sql")
+	if err := os.WriteFile(filepath.Join(sqlDir, "assets", "body.txt"), []byte("relative body"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	got, err := ProcessEmbedDirectives(sqlPath, `
+-- @embed: path=assets/body.txt var=:body mode=text
+SELECT :body;
+`)
+	if err != nil {
+		t.Fatalf("ProcessEmbedDirectives failed: %v", err)
+	}
+	if !strings.Contains(got, "SELECT 'relative body';") {
+		t.Fatalf("path was not resolved relative to SQL file:\n%s", got)
+	}
+}
+
+func TestProcessEmbedDirectives_InvalidDirectiveAndFiles(t *testing.T) {
+	dir := t.TempDir()
+	sqlPath := filepath.Join(dir, "1_001_seed.sql")
+	if err := os.WriteFile(filepath.Join(dir, "body.txt"), []byte("ok"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+	if err := os.WriteFile(filepath.Join(dir, "binary.txt"), []byte{0xff, 0xfe}, 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	tests := []struct {
+		name string
+		sql  string
+	}{
+		{
+			name: "missing mode",
+			sql:  "-- @embed: path=body.txt var=:body\nSELECT :body;",
+		},
+		{
+			name: "invalid var",
+			sql:  "-- @embed: path=body.txt var=body mode=text\nSELECT :body;",
+		},
+		{
+			name: "missing file",
+			sql:  "-- @embed: path=missing.txt var=:body mode=text\nSELECT :body;",
+		},
+		{
+			name: "invalid utf8 text",
+			sql:  "-- @embed: path=binary.txt var=:body mode=text\nSELECT :body;",
+		},
+		{
+			name: "placeholder not found",
+			sql:  "-- @embed: path=body.txt var=:body mode=text\nSELECT 1;",
+		},
+		{
+			name: "unknown attribute",
+			sql:  "-- @embed: path=body.txt var=:body mode=text extra=yes\nSELECT :body;",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if _, err := ProcessEmbedDirectives(sqlPath, tt.sql); err == nil {
+				t.Fatal("expected error, got nil")
+			}
+		})
+	}
+}
+
+func TestProcessEmbedDirectives_DoesNotReplacePlaceholderPrefix(t *testing.T) {
+	dir := t.TempDir()
+	sqlPath := filepath.Join(dir, "1_001_seed.sql")
+	if err := os.WriteFile(filepath.Join(dir, "body.txt"), []byte("ok"), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	got, err := ProcessEmbedDirectives(sqlPath, `
+-- @embed: path=body.txt var=:body mode=text
+SELECT :body, :body_extra;
+`)
+	if err != nil {
+		t.Fatalf("ProcessEmbedDirectives failed: %v", err)
+	}
+	if !strings.Contains(got, "SELECT 'ok', :body_extra;") {
+		t.Fatalf("placeholder boundary was not respected:\n%s", got)
+	}
+}