@bettercms-ai/convert 0.17.1 → 0.17.2

package/CHANGELOG.md

@@ -1,5 +1,11 @@
 # @bettercms-ai/convert
 
+## 0.17.2
+
+### Patch Changes
+
+- d247026: `--xray` now measures a page exactly as the release pass does. `data-bcms-block` and `data-bcms-path` no longer own the text under them (only `data-bcms-field` and `data-bcms-layout-field` do); `<select>`, `<object>`, `<canvas>` and MathML are skipped; numbers such as `$1,200` and short copy such as `OK` count as copy; only a real in-page skip link is excluded. A page's unaddressable figure may rise: it now matches what the binding report says for the same build.
+
 ## 0.17.1
 
 ### Patch Changes

package/dist/cli.js

@@ -6801,7 +6801,7 @@ function isSourceCandidate(path) {
 }
 
 // src/xray.ts
-var SKIP_TAGS = /* @__PURE__ */ new Set(["svg", "iframe", "head", "title"]);
+var SKIP_TAGS = /* @__PURE__ */ new Set(["svg", "math", "iframe", "head", "title", "object", "canvas", "select"]);
 var CONTEXT_TAGS = /* @__PURE__ */ new Set([
   "nav",
   "header",
@@ -6819,10 +6819,16 @@ var CONTEXT_TAGS = /* @__PURE__ */ new Set([
   "figure",
   "blockquote"
 ]);
-var BOUND_ATTRS = ["data-bcms-field", "data-bcms-layout-field", "data-bcms-path", "data-bcms-block"];
-var isDecorativeGlyph = (text) => text.length <= 2 || /^\d{1,2}\s*[/.)\-–]?$/.test(text) || !new RegExp("\\p{L}", "u").test(text) || /^made with\b/i.test(text) || /^skip to\b/i.test(text);
+var BOUND_ATTRS = ["data-bcms-field", "data-bcms-layout-field"];
+var isDecorativeGlyph = (text) => {
+  const t = flat2(text);
+  return t === "" || !/[\p{L}\p{N}]/u.test(t) || new RegExp("^\\p{Nd}{1,2}\\s*[/.)\\-\u2013\u2014]?$", "u").test(t);
+};
+var PLATFORM_BADGE = /^made with\b/iu;
 var MAX_BUCKETS = 12;
 var flat2 = (value) => value.replace(/\s+/g, " ").trim();
+var textOf = (node) => node.type === "text" ? node.value : node.type === "element" ? node.children.map(textOf).join("") : "";
+var isSkipLink = (node, attrs) => node.tag === "a" && (attrs.get("href") ?? "").startsWith("#") && (/^skip\b/iu.test(flat2(textOf(node))) || /\bskip[-_]?(link|nav|to|content)\b/iu.test(attrs.get("class") ?? ""));
 function measureUnaddressable(html) {
   const parsed = parse2(html);
   let visible = 0;
@@ -6831,7 +6837,7 @@ function measureUnaddressable(html) {
   const visit = (node, tag, context, bound) => {
     if (node.type === "text") {
       const text = flat2(node.value);
-      if (!text || isDecorativeGlyph(text)) return;
+      if (isDecorativeGlyph(text) || PLATFORM_BADGE.test(text)) return;
       visible += text.length;
       if (bound) return;
       unaddressable += text.length;
@@ -6845,7 +6851,7 @@ function measureUnaddressable(html) {
     }
     if (node.type !== "element" || node.raw || SKIP_TAGS.has(node.tag)) return;
     const attrs = new Map(node.attrs.map((a) => [a.name, a.value]));
-    if (attrs.has("hidden") || attrs.get("aria-hidden") === "true" || /display\s*:\s*none/.test(attrs.get("style") ?? "")) {
+    if (attrs.has("hidden") || attrs.get("aria-hidden") === "true" || /display\s*:\s*none/i.test(attrs.get("style") ?? "") || isSkipLink(node, attrs)) {
       return;
     }
     const owned = bound || BOUND_ATTRS.some((name) => attrs.has(name));
@@ -6856,7 +6862,7 @@ function measureUnaddressable(html) {
   return {
     visible,
     unaddressable,
-    buckets: [...buckets.values()].sort((a, b) => b.chars - a.chars).slice(0, MAX_BUCKETS)
+    buckets: [...buckets.values()].sort((a, b) => b.chars - a.chars || `${a.tag} @ ${a.context}`.localeCompare(`${b.tag} @ ${b.context}`)).slice(0, MAX_BUCKETS)
   };
 }
 
@@ -9801,7 +9807,7 @@ function emitStubs2(dialects, slugs, contents, out) {
 // package.json
 var package_default = {
   name: "@bettercms-ai/convert",
-  version: "0.17.1",
+  version: "0.17.2",
   type: "module",
   description: "Deterministic codemod that turns an imported site's text-matched bindings into declared ones \u2014 parser offsets in, one magic-string pass out, plus a receipt of everything it could not do.",
   bin: {