cli: add --output-overwrite

cli/html2markdown/cmd/cmd_help.go

@@ -60,6 +60,9 @@ Use a HTML sanitizer before displaying the HTML in the browser!
     --output PATH
         Output file or directory (instead of stdout)
 
+    --output-overwrite
+        Replace existing files
+
     If --input is a directory or glob pattern, --output must be a directory.
 
 
@@ -95,7 +98,7 @@ func tmpl(w io.Writer, text string, data interface{}) error {
 func (cli *CLI) initUsageText() error {
 	var flags []*flag.Flag
 	cli.flags.VisitAll(func(f *flag.Flag) {
-		if f.Name == "v" || f.Name == "version" || f.Name == "input" || f.Name == "output" {
+		if f.Name == "v" || f.Name == "version" || f.Name == "input" || f.Name == "output" || f.Name == "output-overwrite" {
 			// We manually mention these in the usage
 			return
 		}

cli/html2markdown/cmd/exec.go

@@ -23,8 +23,9 @@ type Config struct {
 	// args are the positional (non-flag) command-line arguments.
 	args []string
 
-	inputFilepath  string
-	outputFilepath string
+	inputFilepath   string
+	outputFilepath  string
+	outputOverwrite bool
 
 	// - - - - - General - - - - - //
 	version bool

cli/html2markdown/cmd/files_test.go

@@ -253,6 +253,63 @@ func TestExecute_DirectoryOutput(t *testing.T) {
 	`)
 }
 
+func TestExecute_NotOverwrite(t *testing.T) {
+	directoryPath := newTestDir(t)
+	defer os.RemoveAll(directoryPath)
+
+	outputPath := filepath.Join(directoryPath, "output.md")
+
+	t.Run("the first run", func(t *testing.T) {
+		args := []string{"html2markdown", "--output", outputPath}
+
+		stdin := &FakeFile{mode: modePipe}
+		stdout := &FakeFile{mode: modePipe}
+		stderr := &FakeFile{mode: modePipe}
+		stdin.WriteString("<strong>file content A</strong>")
+
+		Run(stdin, stdout, stderr, args, testRelease)
+
+		stderrBytes := stderr.Bytes()
+		if len(stderrBytes) != 0 {
+			t.Fatalf("got error: %q", string(stderrBytes))
+		}
+		if len(stdout.Bytes()) != 0 {
+			t.Fatalf("expected no stdout content")
+		}
+
+		expectRepresentation(t, directoryPath, `
+.
+├─output.md "**file content A**"
+		`)
+	})
+
+	t.Run("the second run", func(t *testing.T) {
+		args := []string{"html2markdown", "--output", outputPath}
+
+		stdin := &FakeFile{mode: modePipe}
+		stdout := &FakeFile{mode: modePipe}
+		stderr := &FakeFile{mode: modePipe}
+		stdin.WriteString("<strong>file content B</strong>")
+
+		Run(stdin, stdout, stderr, args, testRelease)
+
+		actualStderr := stderr.String()
+		expectedStderr := fmt.Sprintf("\nerror: output path %q already exists. Use --output-overwrite to replace existing files\n\n", outputPath)
+		if actualStderr != expectedStderr {
+			t.Errorf("expected stderr %q but got %q", expectedStderr, actualStderr)
+		}
+		if len(stdout.Bytes()) != 0 {
+			t.Fatalf("expected no stdout content")
+		}
+
+		// The content should still be the same:
+		expectRepresentation(t, directoryPath, `
+.
+├─output.md "**file content A**"
+	`)
+	})
+}
+
 // - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - //
 
 func TestExecute_FilePattern(t *testing.T) {
@@ -307,7 +364,7 @@ func TestExecute_FilePattern(t *testing.T) {
 				input := filepath.Join(dir, "input", "website_a.html")
 				output := filepath.Join(dir, "input", "website_a.html")
 
-				return []string{"html2markdown", "--input", input, "--output", output}
+				return []string{"html2markdown", "--input", input, "--output", output, "--output-overwrite"}
 			},
 			// Note: The file content of "website_a.html" changed
 			expected: `
@@ -487,3 +544,50 @@ func TestExecute_FilePattern(t *testing.T) {
 		})
 	}
 }
+
+func TestWriteFile_OverrideFalse(t *testing.T) {
+	directoryPath := newTestDir(t)
+	defer os.RemoveAll(directoryPath)
+
+	filePath := filepath.Join(directoryPath, "test.txt")
+	override := false
+
+	expectRepresentation(t, directoryPath, ".")
+
+	// - - - file does not exist yet - - - //
+	err := WriteFile(filePath, []byte("A"), override)
+	if err != nil {
+		t.Error(err)
+	}
+	expectRepresentation(t, directoryPath, ".\n"+`├─test.txt "A"`)
+
+	// - - - file exists already - - - //
+	err = WriteFile(filePath, []byte("B"), override)
+	if err == nil {
+		t.Error("expected there to be an error but got nil")
+	}
+	expectRepresentation(t, directoryPath, ".\n"+`├─test.txt "A"`) // <-- still the old content
+}
+func TestWriteFile_OverrideTrue(t *testing.T) {
+	directoryPath := newTestDir(t)
+	defer os.RemoveAll(directoryPath)
+
+	filePath := filepath.Join(directoryPath, "test.txt")
+	override := true
+
+	expectRepresentation(t, directoryPath, ".")
+
+	// - - - file does not exist yet - - - //
+	err := WriteFile(filePath, []byte("A"), override)
+	if err != nil {
+		t.Error(err)
+	}
+	expectRepresentation(t, directoryPath, ".\n"+`├─test.txt "A"`)
+
+	// - - - file exists already - - - //
+	err = WriteFile(filePath, []byte("B"), override)
+	if err != nil {
+		t.Error(err)
+	}
+	expectRepresentation(t, directoryPath, ".\n"+`├─test.txt "B"`) // <-- the new content
+}

cli/html2markdown/cmd/flags.go

@@ -60,6 +60,7 @@ func (cli *CLI) initFlags(progname string) {
 		"output",
 		"Write output to FILE instead of stdout",
 	)
+	cli.flags.BoolVar(&cli.config.outputOverwrite, "output-overwrite", false, "replace existing files")
 
 	// TODO: --tag-type-block=script,style (and check that it is not a selector)
 	// TODO: --tag-type-inline=script,style (and check that it is not a selector)

cli/html2markdown/cmd/io_output.go

@@ -1,6 +1,7 @@
 package cmd
 
 import (
+	"errors"
 	"fmt"
 	"os"
 	"path/filepath"
@@ -92,17 +93,25 @@ func (cli *CLI) writeOutput(outputType outputType, filename string, markdown []b
 	switch outputType {
 	case outputTypeDirectory:
 		{
-			err := os.WriteFile(filepath.Join(cli.config.outputFilepath, filename), markdown, 0644)
+			err := WriteFile(filepath.Join(cli.config.outputFilepath, filename), markdown, cli.config.outputOverwrite)
 			if err != nil {
+				if errors.Is(err, os.ErrExist) {
+					return fmt.Errorf("output path %q already exists. Use --output-overwrite to replace existing files", cli.config.outputFilepath)
+				}
+
 				return fmt.Errorf("error while writing the file into the directory: %w", err)
 			}
 
 			return nil
 		}
 	case outputTypeFile:
 		{
-			err := os.WriteFile(cli.config.outputFilepath, markdown, 0644)
+			err := WriteFile(cli.config.outputFilepath, markdown, cli.config.outputOverwrite)
 			if err != nil {
+				if errors.Is(err, os.ErrExist) {
+					return fmt.Errorf("output path %q already exists. Use --output-overwrite to replace existing files", cli.config.outputFilepath)
+				}
+
 				return fmt.Errorf("error while writing the file: %w", err)
 			}
 
@@ -116,3 +125,33 @@ func (cli *CLI) writeOutput(outputType outputType, filename string, markdown []b
 		}
 	}
 }
+
+// WriteFile writes data to a file with override control
+// If override is false and file exists, returns an error
+// If override is true, truncates existing file or creates new one
+func WriteFile(filename string, data []byte, override bool) error {
+	// As the base flags we have:
+	//   O_WRONLY = write to the file, not read
+	//   O_CREATE = create the file if it doesn't exist
+	flag := os.O_WRONLY | os.O_CREATE
+
+	if override {
+		// We add this flag:
+		//   O_TRUNC = the existing contents are truncated to zero length
+		flag |= os.O_TRUNC
+	} else {
+		// We add this flag:
+		//   O_EXCL = if used with O_CREATE, causes error if file already exists
+		flag |= os.O_EXCL
+	}
+
+	f, err := os.OpenFile(filename, flag, 0644)
+	if err != nil {
+		return err
+	}
+	_, err = f.Write(data)
+	if err1 := f.Close(); err1 != nil && err == nil {
+		err = err1
+	}
+	return err
+}

cli/html2markdown/cmd/testdata/TestExecute/[general]_help_pipe/stdout.golden

@@ -49,6 +49,9 @@ Use a HTML sanitizer before displaying the HTML in the browser!
     --output PATH
         Output file or directory (instead of stdout)
 
+    --output-overwrite
+        Replace existing files
+
     If --input is a directory or glob pattern, --output must be a directory.
 
 

cli/html2markdown/cmd/testdata/TestExecute/[general]_help_terminal/stdout.golden

@@ -49,6 +49,9 @@ Use a HTML sanitizer before displaying the HTML in the browser!
     --output PATH
         Output file or directory (instead of stdout)
 
+    --output-overwrite
+        Replace existing files
+
     If --input is a directory or glob pattern, --output must be a directory.