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

chunk.go

@@ -1,26 +0,0 @@
-package util
-
-// ChunkStrings splits input into consecutive slices of at most size elements.
-// The last slice holds whatever is left over and may be shorter than the
-// others. Each returned slice is a copy, so appending to one of them cannot
-// disturb another. A size below one, or an empty input, returns nil.
-func ChunkStrings(input []string, size int) [][]string {
-	if size < 1 || len(input) == 0 {
-		return nil
-	}
-
-	out := make([][]string, 0, (len(input)+size-1)/size)
-
-	for start := 0; start < len(input); start += size {
-		end := start + size
-		if end > len(input) {
-			end = len(input)
-		}
-
-		chunk := make([]string, end-start)
-		copy(chunk, input[start:end])
-		out = append(out, chunk)
-	}
-
-	return out
-}

chunk_test.go

@@ -1,43 +0,0 @@
-package util
-
-import (
-	"reflect"
-	"testing"
-)
-
-func TestChunkStrings(t *testing.T) {
-	tests := []struct {
-		name     string
-		input    []string
-		size     int
-		expected [][]string
-	}{
-		{"exact multiple", []string{"a", "b", "c", "d"}, 2, [][]string{{"a", "b"}, {"c", "d"}}},
-		{"with a remainder", []string{"a", "b", "c"}, 2, [][]string{{"a", "b"}, {"c"}}},
-		{"size larger than input", []string{"a", "b"}, 5, [][]string{{"a", "b"}}},
-		{"size of one", []string{"a", "b"}, 1, [][]string{{"a"}, {"b"}}},
-		{"size of zero", []string{"a", "b"}, 0, nil},
-		{"negative size", []string{"a", "b"}, -1, nil},
-		{"empty input", []string{}, 2, nil},
-		{"nil input", nil, 2, nil},
-	}
-
-	for _, test := range tests {
-		t.Run(test.name, func(t *testing.T) {
-			got := ChunkStrings(test.input, test.size)
-			if !reflect.DeepEqual(got, test.expected) {
-				t.Errorf("expected %#v got %#v", test.expected, got)
-			}
-		})
-	}
-}
-
-func TestChunkStringsReturnsCopies(t *testing.T) {
-	input := []string{"a", "b", "c", "d"}
-	chunks := ChunkStrings(input, 2)
-	chunks[0][0] = "changed"
-
-	if input[0] != "a" {
-		t.Errorf("expected the input to be unchanged, got %#v", input)
-	}
-}

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)
+			}
+		})
+	}
+}