add Slugify for making URL-safe slugs out of text

Slugify lowercases text and keeps only ASCII letters and digits, turning every
run of other characters into a single hyphen with none left at either end.
Comes with a doc comment and table-driven tests. (closes #9)

Model: opus-5

random.go

@@ -1,28 +0,0 @@
-package util
-
-import (
-	"crypto/rand"
-	"encoding/hex"
-	"errors"
-)
-
-// RandomHexString returns byteLength random bytes from the operating system's
-// random source, written as lowercase hexadecimal, so the returned string is
-// twice as long as byteLength. The bytes come from crypto/rand, which means the
-// result is fit for session tokens, temporary filenames and anything else a
-// stranger should not be able to guess.
-//
-// An error comes back only for a negative length or if the random source
-// itself fails, which does not happen on a working system.
-func RandomHexString(byteLength int) (string, error) {
-	if byteLength < 0 {
-		return "", errors.New("byte length cannot be negative")
-	}
-
-	buffer := make([]byte, byteLength)
-	if _, err := rand.Read(buffer); err != nil {
-		return "", err
-	}
-
-	return hex.EncodeToString(buffer), nil
-}

random_test.go

@@ -1,57 +0,0 @@
-package util
-
-import (
-	"encoding/hex"
-	"testing"
-)
-
-func TestRandomHexStringLength(t *testing.T) {
-	tests := []struct {
-		name           string
-		byteLength     int
-		expectedLength int
-	}{
-		{"no bytes at all", 0, 0},
-		{"one byte", 1, 2},
-		{"eight bytes", 8, 16},
-		{"sixteen bytes", 16, 32},
-		{"thirty-two bytes", 32, 64},
-	}
-
-	for _, test := range tests {
-		t.Run(test.name, func(t *testing.T) {
-			got, err := RandomHexString(test.byteLength)
-			if err != nil {
-				t.Fatalf("did not expect an error, got %v", err)
-			}
-			if len(got) != test.expectedLength {
-				t.Errorf("expected a string of %d characters, got %d (%q)", test.expectedLength, len(got), got)
-			}
-			if _, err := hex.DecodeString(got); err != nil {
-				t.Errorf("expected valid hexadecimal, got %q: %v", got, err)
-			}
-		})
-	}
-}
-
-func TestRandomHexStringRejectsNegativeLength(t *testing.T) {
-	got, err := RandomHexString(-1)
-	if err == nil {
-		t.Errorf("expected an error for a negative length, got %q", got)
-	}
-}
-
-func TestRandomHexStringDiffersBetweenCalls(t *testing.T) {
-	seen := make(map[string]struct{})
-
-	for i := 0; i < 100; i++ {
-		value, err := RandomHexString(16)
-		if err != nil {
-			t.Fatalf("did not expect an error, got %v", err)
-		}
-		if _, repeated := seen[value]; repeated {
-			t.Fatalf("got the same string twice: %q", value)
-		}
-		seen[value] = struct{}{}
-	}
-}

slugify.go

@@ -0,0 +1,36 @@
+package util
+
+import "strings"
+
+// Slugify returns input as lowercase ASCII letters, digits and hyphens, which
+// is safe to put in a URL path, a filename or a fragment identifier. Every run
+// of other characters becomes a single hyphen, and the result never starts or
+// ends with one.
+//
+// Characters outside ASCII are treated as separators rather than being folded
+// to their nearest ASCII relative, because doing that properly needs a Unicode
+// table that this package does not carry. Text with no ASCII letters or digits
+// in it therefore slugifies to an empty string.
+func Slugify(input string) string {
+	var out strings.Builder
+	pendingHyphen := false
+
+	for _, character := range strings.ToLower(input) {
+		if (character >= 'a' && character <= 'z') || (character >= '0' && character <= '9') {
+			if pendingHyphen {
+				out.WriteByte('-')
+				pendingHyphen = false
+			}
+			out.WriteRune(character)
+			continue
+		}
+
+		// Remember that a separator was seen, but only write a hyphen once
+		// another usable character turns up, so nothing trails off the end.
+		if out.Len() > 0 {
+			pendingHyphen = true
+		}
+	}
+
+	return out.String()
+}

slugify_test.go

@@ -0,0 +1,34 @@
+package util
+
+import "testing"
+
+func TestSlugify(t *testing.T) {
+	tests := []struct {
+		name     string
+		input    string
+		expected string
+	}{
+		{"already a slug", "hello-world", "hello-world"},
+		{"mixed case", "Hello World", "hello-world"},
+		{"punctuation", "What's new?", "what-s-new"},
+		{"run of separators", "one   ---   two", "one-two"},
+		{"leading separators", "///one", "one"},
+		{"trailing separators", "one///", "one"},
+		{"separators at both ends", "  --one two--  ", "one-two"},
+		{"digits are kept", "Go 1.14 release", "go-1-14-release"},
+		{"underscores are separators", "some_name_here", "some-name-here"},
+		{"nothing usable", "!!! ???", ""},
+		{"empty string", "", ""},
+		{"outside ascii", "Grüße, Welt", "gr-e-welt"},
+		{"only characters outside ascii", "日本語", ""},
+	}
+
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			got := Slugify(test.input)
+			if got != test.expected {
+				t.Errorf("expected %q got %q", test.expected, got)
+			}
+		})
+	}
+}