Remove normalize and format packages and format testing

CLAUDE.md

@@ -5,19 +5,11 @@
 To find the next explain test to work on (fewest pending statements first), run:
 
 ```bash
-go run ./cmd/next-test -explain
+go run ./cmd/next-test
 ```
 
 This finds tests with `explain_todo` entries in their metadata.
 
-To find the next format roundtrip test to work on, run:
-
-```bash
-go run ./cmd/next-test -format
-```
-
-This finds tests with `todo_format: true` in their metadata.
-
 ## Running Tests
 
 Always run parser tests with a 5 second timeout:
@@ -28,16 +20,6 @@ go test ./parser/... -timeout 5s
 
 The tests are very fast. If a test is timing out, it indicates a bug (likely an infinite loop in the parser).
 
-## Checking for Newly Passing Format Tests
-
-After implementing format changes, run:
-
-```bash
-go test ./parser/... -check-format -v 2>&1 | grep "FORMAT PASSES NOW"
-```
-
-Tests that output `FORMAT PASSES NOW` can have their `todo_format` flag removed from `metadata.json`.
-
 ## Checking for Newly Passing Explain Tests
 
 After implementing parser/explain changes, run:
@@ -59,7 +41,6 @@ Each test in `parser/testdata/` contains:
 
 ### Metadata Options
 
-- `todo_format: true` - Format roundtrip test is pending implementation
 - `explain_todo: {"stmt2": true}` - Skip specific statement subtests
 - `skip: true` - Skip test entirely (e.g., causes infinite loop)
 - `explain: false` - Skip test (e.g., ClickHouse couldn't parse it)

cmd/next-test/main.go

@@ -2,18 +2,13 @@ package main
 
 import (
 	"encoding/json"
-	"flag"
 	"fmt"
 	"os"
 	"path/filepath"
 	"sort"
 )
 
-var formatFlag = flag.Bool("format", false, "Find tests with todo_format: true")
-var explainFlag = flag.Bool("explain", false, "Find tests with explain_todo entries (fewest first)")
-
 type testMetadata struct {
-	TodoFormat  bool            `json:"todo_format,omitempty"`
 	ExplainTodo map[string]bool `json:"explain_todo,omitempty"`
 	Explain     *bool           `json:"explain,omitempty"`
 	Skip        bool            `json:"skip,omitempty"`
@@ -27,15 +22,6 @@ type todoTest struct {
 }
 
 func main() {
-	flag.Parse()
-
-	if !*formatFlag && !*explainFlag {
-		fmt.Fprintf(os.Stderr, "Usage: go run ./cmd/next-test [-format | -explain]\n")
-		fmt.Fprintf(os.Stderr, "  -format   Find tests with todo_format: true\n")
-		fmt.Fprintf(os.Stderr, "  -explain  Find tests with explain_todo entries (fewest first)\n")
-		os.Exit(1)
-	}
-
 	testdataDir := "parser/testdata"
 	entries, err := os.ReadDir(testdataDir)
 	if err != nil {
@@ -69,15 +55,9 @@ func main() {
 			continue
 		}
 
-		// Check based on flag
-		if *formatFlag {
-			if !metadata.TodoFormat {
-				continue
-			}
-		} else if *explainFlag {
-			if len(metadata.ExplainTodo) == 0 {
-				continue
-			}
+		// Check for explain_todo entries
+		if len(metadata.ExplainTodo) == 0 {
+			continue
 		}
 
 		// Read query to get its size
@@ -94,41 +74,24 @@ func main() {
 		})
 	}
 
-	todoType := "todo_format"
-	if *explainFlag {
-		todoType = "explain_todo"
-	}
-
 	if len(todoTests) == 0 {
-		fmt.Printf("No %s tests found!\n", todoType)
+		fmt.Printf("No explain_todo tests found!\n")
 		return
 	}
 
-	// Sort based on mode
-	if *explainFlag {
-		// Sort by explain_todo count (fewest first), then by query size
-		sort.Slice(todoTests, func(i, j int) bool {
-			if todoTests[i].explainTodoLen != todoTests[j].explainTodoLen {
-				return todoTests[i].explainTodoLen < todoTests[j].explainTodoLen
-			}
-			return todoTests[i].querySize < todoTests[j].querySize
-		})
-	} else {
-		// Sort by query size (shortest first)
-		sort.Slice(todoTests, func(i, j int) bool {
-			return todoTests[i].querySize < todoTests[j].querySize
-		})
-	}
+	// Sort by explain_todo count (fewest first), then by query size
+	sort.Slice(todoTests, func(i, j int) bool {
+		if todoTests[i].explainTodoLen != todoTests[j].explainTodoLen {
+			return todoTests[i].explainTodoLen < todoTests[j].explainTodoLen
+		}
+		return todoTests[i].querySize < todoTests[j].querySize
+	})
 
 	// Print the best candidate
 	next := todoTests[0]
 	testDir := filepath.Join(testdataDir, next.name)
 
-	if *explainFlag {
-		fmt.Printf("Next %s test: %s (%d pending statements)\n\n", todoType, next.name, next.explainTodoLen)
-	} else {
-		fmt.Printf("Next %s test: %s\n\n", todoType, next.name)
-	}
+	fmt.Printf("Next explain_todo test: %s (%d pending statements)\n\n", next.name, next.explainTodoLen)
 
 	// Print query.sql contents
 	queryPath := filepath.Join(testDir, "query.sql")
@@ -141,19 +104,17 @@ func main() {
 		fmt.Printf("\nExpected EXPLAIN output:\n%s\n", string(explainBytes))
 	}
 
-	// Print explain_todo entries if in explain mode
-	if *explainFlag {
-		metadataPath := filepath.Join(testDir, "metadata.json")
-		if metadataBytes, err := os.ReadFile(metadataPath); err == nil {
-			var metadata testMetadata
-			if json.Unmarshal(metadataBytes, &metadata) == nil {
-				fmt.Printf("\nPending statements (explain_todo):\n")
-				for stmt := range metadata.ExplainTodo {
-					fmt.Printf("  - %s\n", stmt)
-				}
+	// Print explain_todo entries
+	metadataPath := filepath.Join(testDir, "metadata.json")
+	if metadataBytes, err := os.ReadFile(metadataPath); err == nil {
+		var metadata testMetadata
+		if json.Unmarshal(metadataBytes, &metadata) == nil {
+			fmt.Printf("\nPending statements (explain_todo):\n")
+			for stmt := range metadata.ExplainTodo {
+				fmt.Printf("  - %s\n", stmt)
 			}
 		}
 	}
 
-	fmt.Printf("\nRemaining %s tests: %d\n", todoType, len(todoTests))
+	fmt.Printf("\nRemaining explain_todo tests: %d\n", len(todoTests))
 }