wip: refactor Golden to a interface, improve testability of main implementation

sanitize/filename.go

@@ -0,0 +1,34 @@
+package sanitize
+
+import (
+	"regexp"
+	"strings"
+)
+
+var (
+	whitespaceChars = regexp.MustCompile(`\s`)
+	illegalChars    = regexp.MustCompile(`[\/\?<>\\:\*\|"]`)
+	controlChars    = regexp.MustCompile(`[\x00-\x1f\x80-\x9f]`)
+	reservedNames   = regexp.MustCompile(`^\.+$`)
+	winReserved     = regexp.MustCompile(
+		`(?i)^(con|prn|aux|nul|com[0-9]|lpt[0-9])(\..*)?$`,
+	)
+)
+
+func Filename(name string) string {
+	if reservedNames.MatchString(name) || winReserved.MatchString(name) {
+		var b []byte
+		for i := 0; i < len(name); i++ {
+			b = append(b, byte('_'))
+		}
+
+		return string(b)
+	}
+
+	r := strings.TrimRight(name, ". ")
+	r = whitespaceChars.ReplaceAllString(r, "_")
+	r = illegalChars.ReplaceAllString(r, "_")
+	r = controlChars.ReplaceAllString(r, "_")
+
+	return r
+}

sanitize/filename_test.go

@@ -0,0 +1,126 @@
+package sanitize_test
+
+import (
+	"testing"
+
+	"github.com/jimeh/go-golden/sanitize"
+	"github.com/stretchr/testify/assert"
+)
+
+func TestFilename(t *testing.T) {
+	tests := []struct {
+		name     string
+		filename string
+		want     string
+	}{
+		{
+			name:     "empty",
+			filename: "",
+			want:     "",
+		},
+		{
+			name:     ".",
+			filename: ".",
+			want:     "_",
+		},
+		{
+			name:     "..",
+			filename: "..",
+			want:     "__",
+		},
+		{
+			name:     "...",
+			filename: "...",
+			want:     "___",
+		},
+		{
+			name:     "clean",
+			filename: "foo-bar-nope.golden",
+			want:     "foo-bar-nope.golden",
+		},
+		{
+			name:     "with spaces",
+			filename: "foo  bar nope.golden",
+			want:     "foo__bar_nope.golden",
+		},
+		{
+			name:     "illegal chars",
+			filename: `foo/?<>\:*|"bar.golden`,
+			want:     "foo_________bar.golden",
+		},
+		{
+			name: "control chars",
+			filename: "foo\x00\x01\x02\x03\x04\x05\x06\x07\x08\x09\x0a\x0b" +
+				"\x0c\x0d\x0e\x0f\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1a" +
+				"\x1b\x1c\x1d\x1e\x1fbar.golden",
+			want: "foo________________________________bar.golden",
+		},
+		{
+			name:     "trailing whitespace",
+			filename: "foobar.golden    ",
+			want:     "foobar.golden",
+		},
+		{
+			name:     "trailing dots",
+			filename: "foobar.golden......",
+			want:     "foobar.golden",
+		},
+		{
+			name:     "trailing whitespace and dots",
+			filename: "foobar.golden  ..  ..  ..  ",
+			want:     "foobar.golden",
+		},
+		// Protected Windows filenames.
+		{name: "con", filename: "con", want: "___"},
+		{name: "prn", filename: "prn", want: "___"},
+		{name: "aux", filename: "aux", want: "___"},
+		{name: "nul", filename: "nul", want: "___"},
+		{name: "com1", filename: "com1", want: "____"},
+		{name: "com2", filename: "com2", want: "____"},
+		{name: "com3", filename: "com3", want: "____"},
+		{name: "com4", filename: "com4", want: "____"},
+		{name: "com5", filename: "com5", want: "____"},
+		{name: "com6", filename: "com6", want: "____"},
+		{name: "com7", filename: "com7", want: "____"},
+		{name: "com8", filename: "com8", want: "____"},
+		{name: "com9", filename: "com9", want: "____"},
+		{name: "lpt1", filename: "lpt1", want: "____"},
+		{name: "lpt2", filename: "lpt2", want: "____"},
+		{name: "lpt3", filename: "lpt3", want: "____"},
+		{name: "lpt4", filename: "lpt4", want: "____"},
+		{name: "lpt5", filename: "lpt5", want: "____"},
+		{name: "lpt6", filename: "lpt6", want: "____"},
+		{name: "lpt7", filename: "lpt7", want: "____"},
+		{name: "lpt8", filename: "lpt8", want: "____"},
+		{name: "lpt9", filename: "lpt9", want: "____"},
+		{name: "CON", filename: "CON", want: "___"},
+		{name: "PRN", filename: "PRN", want: "___"},
+		{name: "AUX", filename: "AUX", want: "___"},
+		{name: "NUL", filename: "NUL", want: "___"},
+		{name: "COM1", filename: "COM1", want: "____"},
+		{name: "COM2", filename: "COM2", want: "____"},
+		{name: "COM3", filename: "COM3", want: "____"},
+		{name: "COM4", filename: "COM4", want: "____"},
+		{name: "COM5", filename: "COM5", want: "____"},
+		{name: "COM6", filename: "COM6", want: "____"},
+		{name: "COM7", filename: "COM7", want: "____"},
+		{name: "COM8", filename: "COM8", want: "____"},
+		{name: "COM9", filename: "COM9", want: "____"},
+		{name: "LPT1", filename: "LPT1", want: "____"},
+		{name: "LPT2", filename: "LPT2", want: "____"},
+		{name: "LPT3", filename: "LPT3", want: "____"},
+		{name: "LPT4", filename: "LPT4", want: "____"},
+		{name: "LPT5", filename: "LPT5", want: "____"},
+		{name: "LPT6", filename: "LPT6", want: "____"},
+		{name: "LPT7", filename: "LPT7", want: "____"},
+		{name: "LPT8", filename: "LPT8", want: "____"},
+		{name: "LPT9", filename: "LPT9", want: "____"},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := sanitize.Filename(tt.filename)
+
+			assert.Equal(t, tt.want, got)
+		})
+	}
+}

sanitize/line_breaks.go

@@ -0,0 +1,21 @@
+package sanitize
+
+import "bytes"
+
+var (
+	lf   = []byte{10}
+	cr   = []byte{13}
+	crlf = []byte{13, 10}
+)
+
+// LineBreaks replaces Windows CRLF (\r\n) and MacOS Classic CR (\r)
+// line-breaks with Unix LF (\n) line breaks.
+func LineBreaks(data []byte) []byte {
+	// Replace Windows CRLF (\r\n) with Unix LF (\n)
+	result := bytes.ReplaceAll(data, crlf, lf)
+
+	// Replace Classic MacOS CR (\r) with Unix LF (\n)
+	result = bytes.ReplaceAll(result, cr, lf)
+
+	return result
+}

sanitize/line_breaks_test.go

@@ -0,0 +1,67 @@
+package sanitize_test
+
+import (
+	"testing"
+
+	"github.com/jimeh/go-golden/sanitize"
+	"github.com/stretchr/testify/assert"
+)
+
+func TestLineBreaks(t *testing.T) {
+	type args struct {
+		data []byte
+	}
+	tests := []struct {
+		name string
+		args args
+		want []byte
+	}{
+		{
+			name: "nil",
+			args: args{data: nil},
+			want: nil,
+		},
+		{
+			name: "empty",
+			args: args{data: []byte{}},
+			want: nil,
+		},
+		{
+			name: "no line breaks",
+			args: args{data: []byte("hello world")},
+			want: []byte("hello world"),
+		},
+		{
+			name: "UNIX line breaks",
+			args: args{data: []byte("hello\nworld\nhow are you?")},
+			want: []byte("hello\nworld\nhow are you?"),
+		},
+		{
+			name: "Windows line breaks",
+			args: args{data: []byte("hello\r\nworld\r\nhow are you?")},
+			want: []byte("hello\nworld\nhow are you?"),
+		},
+		{
+			name: "MacOS Classic line breaks",
+			args: args{data: []byte("hello\rworld\rhow are you?")},
+			want: []byte("hello\nworld\nhow are you?"),
+		},
+		{
+			name: "Windows and MacOS Classic line breaks",
+			args: args{data: []byte("hello\r\nworld\rhow are you?")},
+			want: []byte("hello\nworld\nhow are you?"),
+		},
+		{
+			name: "Windows, MacOS Classic, and UNIX line breaks",
+			args: args{data: []byte("hello\r\nworld\rhow are you?\nGood!")},
+			want: []byte("hello\nworld\nhow are you?\nGood!"),
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := sanitize.LineBreaks(tt.args.data)
+
+			assert.Equal(t, tt.want, got)
+		})
+	}
+}