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

dedupe.go

@@ -1,23 +0,0 @@
-package util
-
-// DedupeStrings returns a new slice holding each value of input once, in the
-// order in which each value first appears. The input slice is not modified.
-// A nil input returns nil.
-func DedupeStrings(input []string) []string {
-	if input == nil {
-		return nil
-	}
-
-	seen := make(map[string]struct{}, len(input))
-	out := make([]string, 0, len(input))
-
-	for _, value := range input {
-		if _, ok := seen[value]; ok {
-			continue
-		}
-		seen[value] = struct{}{}
-		out = append(out, value)
-	}
-
-	return out
-}

dedupe_test.go

@@ -1,40 +0,0 @@
-package util
-
-import (
-	"reflect"
-	"testing"
-)
-
-func TestDedupeStrings(t *testing.T) {
-	tests := []struct {
-		name     string
-		input    []string
-		expected []string
-	}{
-		{"nil stays nil", nil, nil},
-		{"empty slice", []string{}, []string{}},
-		{"nothing repeated", []string{"a", "b", "c"}, []string{"a", "b", "c"}},
-		{"all one value", []string{"a", "a", "a"}, []string{"a"}},
-		{"repeats next to each other", []string{"a", "a", "b", "b"}, []string{"a", "b"}},
-		{"repeats far apart", []string{"a", "b", "a", "c", "b"}, []string{"a", "b", "c"}},
-		{"empty string is a value", []string{"", "a", ""}, []string{"", "a"}},
-	}
-
-	for _, test := range tests {
-		t.Run(test.name, func(t *testing.T) {
-			got := DedupeStrings(test.input)
-			if !reflect.DeepEqual(got, test.expected) {
-				t.Errorf("expected %#v got %#v", test.expected, got)
-			}
-		})
-	}
-}
-
-func TestDedupeStringsLeavesInputAlone(t *testing.T) {
-	input := []string{"b", "a", "b"}
-	DedupeStrings(input)
-
-	if !reflect.DeepEqual(input, []string{"b", "a", "b"}) {
-		t.Errorf("expected the input to be unchanged, got %#v", input)
-	}
-}

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)
+	}
+}