krino: matching — scan, ignore, conditions, extraction, duplicates, explain, dry run

1 files changed, 73 insertions, 0 deletions

diff --git a/internal/norm/norm_test.go b/internal/norm/norm_test.go
new file mode 100644
index 0000000..2fc35b6
--- /dev/null
+++ b/internal/norm/norm_test.go
@@ -0,0 +1,73 @@
+// SPDX-License-Identifier: GPL-3.0-or-later
+
+package norm
+
+import "testing"
+
+func TestFold(t *testing.T) {
+	tests := map[string]string{
+		"":                "",
+		"plain ASCII 123": "plain ASCII 123",
+		"spółka z ograniczoną odpowiedzialnością": "spolka z ograniczona odpowiedzialnoscia",
+		"Łódź":              "Lodz",
+		"ZAŻÓŁĆ GĘŚLĄ JAŹŃ": "ZAZOLC GESLA JAZN",
+		"Straße":            "Strasse",
+		"Øresund":           "Oresund",
+		"Ærø":               "AEro",
+		"œuvre":             "oeuvre",
+		"déjà vu":           "deja vu",
+		"Đakovo":            "Dakovo",
+		"ħelp":              "help",
+	}
+	for in, want := range tests {
+		if got := Fold(in); got != want {
+			t.Errorf("Fold(%q) = %q, want %q", in, got, want)
+		}
+	}
+}
+
+func TestText(t *testing.T) {
+	tests := []struct {
+		in               string
+		ignoreCase, fold bool
+		want             string
+	}{
+		{"  Hello\n\tWORLD  ", true, false, "hello world"},
+		{"A  B\r\nC", false, false, "A B C"},
+		{"ZAŻÓŁĆ   gęślą", true, true, "zazolc gesla"},
+		{"Spółka", false, true, "Spolka"},
+		{"Faktura\u00A0VAT", true, false, "faktura vat"},
+		{"", true, true, ""},
+		{"   ", true, true, ""},
+	}
+	for _, tt := range tests {
+		if got := Text(tt.in, tt.ignoreCase, tt.fold); got != tt.want {
+			t.Errorf("Text(%q, %v, %v) = %q, want %q", tt.in, tt.ignoreCase, tt.fold, got, tt.want)
+		}
+	}
+}
+
+func TestTextIdempotent(t *testing.T) {
+	for _, s := range []string{"  Łódź\n\nMIASTO ", "a\tb", "Ærø  Œuvre"} {
+		once := Text(s, true, true)
+		if twice := Text(once, true, true); twice != once {
+			t.Errorf("Text not idempotent on %q: %q then %q", s, once, twice)
+		}
+	}
+}
+
+func TestName(t *testing.T) {
+	if got := Name("Spółka.PDF", true); got != "Spolka.PDF" {
+		t.Errorf("Name fold = %q", got)
+	}
+	if got := Name("Spółka.PDF", false); got != "Spółka.PDF" {
+		t.Errorf("Name no fold = %q", got)
+	}
+}
+
+func TestFoldASCIINoAlloc(t *testing.T) {
+	s := "already plain ascii text"
+	if n := testing.AllocsPerRun(100, func() { _ = Fold(s) }); n != 0 {
+		t.Errorf("Fold allocates %v times on ASCII input", n)
+	}
+}