add TruncateString for shortening a string safely

TruncateString returns at most the requested number of characters from the
start of a string, counting characters rather than bytes so a multi-byte
character is never cut in half. Comes with a doc comment and table-driven
tests. (closes #5)

Model: opus-5

slugify.go

@@ -1,36 +0,0 @@
-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

@@ -1,34 +0,0 @@
-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)
-			}
-		})
-	}
-}

truncate.go

@@ -0,0 +1,21 @@
+package util
+
+// TruncateString returns at most max characters from the start of s. It counts
+// characters rather than bytes, so a multi-byte character is never cut in half
+// and the result is always valid text. A max of zero or less returns an empty
+// string, and a string already at or below the limit is returned unchanged.
+func TruncateString(s string, max int) string {
+	if max <= 0 {
+		return ""
+	}
+
+	count := 0
+	for offset := range s {
+		if count == max {
+			return s[:offset]
+		}
+		count++
+	}
+
+	return s
+}

truncate_test.go

@@ -0,0 +1,41 @@
+package util
+
+import "testing"
+
+func TestTruncateString(t *testing.T) {
+	tests := []struct {
+		name     string
+		input    string
+		max      int
+		expected string
+	}{
+		{"shorter than the limit", "abc", 5, "abc"},
+		{"exactly at the limit", "abc", 3, "abc"},
+		{"longer than the limit", "abcdef", 3, "abc"},
+		{"max of one", "abc", 1, "a"},
+		{"max of zero", "abc", 0, ""},
+		{"negative max", "abc", -1, ""},
+		{"empty string", "", 3, ""},
+		{"multi-byte characters kept whole", "héllo", 3, "hél"},
+		{"multi-byte characters below the limit", "héllo", 99, "héllo"},
+		{"characters outside the basic plane", "a😀b", 2, "a😀"},
+	}
+
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			got := TruncateString(test.input, test.max)
+			if got != test.expected {
+				t.Errorf("expected %q got %q", test.expected, got)
+			}
+		})
+	}
+}
+
+func TestTruncateStringCountsCharactersNotBytes(t *testing.T) {
+	// Each of these characters is two bytes long, so a byte-based cut at
+	// three would leave half a character behind.
+	got := TruncateString("äöü", 3)
+	if got != "äöü" {
+		t.Errorf("expected %q got %q", "äöü", got)
+	}
+}