fix: the definitions migration keeps and logs every url item

Migration 202610280002 converts the Phase 17 url items through
models.ConvertLegacyItems: a relative path gets its leading slash,
spaces and the bytes " < > \ are percent-encoded, and a malformed item
no longer drops the whole definition. Each url item that cannot be
copied is logged at warn level with its label, value and reason.

models/legacy_test.go

@@ -0,0 +1,60 @@
+package models_test
+
+import (
+	"strings"
+	"testing"
+
+	"git.golem15.com/golem15/sm-sitemap-plugin/models"
+)
+
+// TestConvertLegacyItems pins L2 WR-04 (D-32, P33-66): every url item the
+// Phase 17 sitemap served is copied, a relative path with a leading slash
+// and spaces and unsafe bytes percent-encoded; a duplicate is kept once
+// and not reported; an item that still is no allowed address, and a
+// malformed one, is dropped with its reason while the others are kept;
+// items of another type are neither kept nor reported.
+func TestConvertLegacyItems(t *testing.T) {
+	data := `[{"type":"url","url":"o-nas"},` +
+		`{"type":"url","url":"/blog/zażółć gęślą"},` +
+		`{"type":"url","url":"/o-nas"},` +
+		`{"type":"url","url":"https://site.test/x?a=1&b=2"},` +
+		`{"type":"url","url":["not", "text"]},` +
+		`{"type":"acme-post","reference":"7"},` +
+		`{"url":"/a\"b<c>d\\e"},` +
+		`{"type":"url","url":"https://site.test/kontakt"},` +
+		`{"type":"url","url":"javascript:alert(1)"}]`
+	rows, dropped := models.ConvertLegacyItems([]byte(data))
+	var got []string
+	for _, r := range rows {
+		got = append(got, r.URL)
+		if r.LastMod != "" {
+			t.Errorf("a copied address has a lastmod: %+v", r)
+		}
+	}
+	want := "/o-nas /blog/zażółć%20gęślą /a%22b%3Cc%3Ed%5Ce https://site.test/kontakt"
+	if strings.Join(got, " ") != want {
+		t.Fatalf("rows = %q, want %q", strings.Join(got, " "), want)
+	}
+	if len(dropped) != 3 {
+		t.Fatalf("dropped = %+v, want the & URL, the malformed item and javascript:", dropped)
+	}
+	if d := dropped[0]; d.Label != "3" || d.URL != "https://site.test/x?a=1&b=2" || !strings.Contains(d.Reason, "&") {
+		t.Errorf("dropped & URL = %+v", d)
+	}
+	if d := dropped[1]; d.Label != "4" || d.URL != "" || !strings.Contains(d.Reason, "url") {
+		t.Errorf("dropped malformed item = %+v", d)
+	}
+	if d := dropped[2]; d.Label != "8" || d.URL != "javascript:alert(1)" || d.Reason == "" {
+		t.Errorf("dropped javascript: = %+v", d)
+	}
+
+	if rows, dropped := models.ConvertLegacyItems([]byte(`{"a":{"url":"/x"},"b":{"url":"/x"}}`)); len(rows) != 1 || rows[0].URL != "/x" || len(dropped) != 0 {
+		t.Errorf("keyed object = %+v %+v", rows, dropped)
+	}
+	if rows, dropped := models.ConvertLegacyItems([]byte(`"not a list"`)); len(rows) != 0 || len(dropped) != 1 || dropped[0].Label != "" {
+		t.Errorf("not a list = %+v %+v", rows, dropped)
+	}
+	if rows, dropped := models.ConvertLegacyItems(nil); len(rows) != 0 || dropped != nil {
+		t.Errorf("empty = %+v %+v", rows, dropped)
+	}
+}