@reddoorla/maintenance 0.94.0 → 0.95.0

package/dist/{chunk-VELWHWOW.js → chunk-WJMPCVOR.js}

@@ -883,7 +883,7 @@ function decodeXmlText(s) {
   return s.replace(/<!\[CDATA\[([\s\S]*?)\]\]>/g, "$1").replace(/&#x([0-9a-f]+);/gi, (_, hex) => String.fromCodePoint(parseInt(hex, 16))).replace(/&#(\d+);/g, (_, dec) => String.fromCodePoint(Number(dec))).replace(/&lt;/g, "<").replace(/&gt;/g, ">").replace(/&quot;/g, '"').replace(/&apos;/g, "'").replace(/&amp;/g, "&");
 }
 function parseSitemapLocs(xml) {
-  return [...xml.matchAll(/<(?:[\w-]+:)?loc\b[^>]*>([\s\S]*?)<\/(?:[\w-]+:)?loc>/gi)].map((m) => decodeXmlText(m[1] ?? "").trim()).filter(Boolean);
+  return [...xml.matchAll(/<(?!image:)(?:[\w-]+:)?loc\b[^>]*>([\s\S]*?)<\/(?:[\w-]+:)?loc>/gi)].map((m) => decodeXmlText(m[1] ?? "").trim()).filter(Boolean);
 }
 function isSafeNestedSitemap(child, origin) {
   let url;
@@ -2983,7 +2983,7 @@ function headerChecks(headers, measured) {
   }
   return out;
 }
-function sidecarChecks(crawl, linkedPages) {
+function sidecarChecks(crawl, linkedPages, aliasOf) {
   const discoveredLinks = linkedPages.size;
   const out = [];
   const robotsMeasured = crawl.sidecarErrors.robots === null;
@@ -3069,7 +3069,12 @@ function sidecarChecks(crawl, linkedPages) {
     const ok = crawl.sitemap.urlCount >= discoveredLinks;
     const sample = crawl.sitemap.sample;
     const whole = sample !== void 0 && crawl.sitemap.urlCount <= sample.length;
-    const listed = new Set(sample?.map((u) => norm2(u) ?? u));
+    const listed = new Set(
+      sample?.map((u) => {
+        const at = norm2(u) ?? u;
+        return aliasOf.get(at) ?? at;
+      })
+    );
     const unlisted = whole ? [...linkedPages].filter((a) => !listed.has(a)) : [];
     out.push(
       check(
@@ -3207,7 +3212,7 @@ function metaChecks(pages) {
   );
   const TITLE_MIN = 10;
   const TITLE_MAX = 70;
-  const titled = pages.filter((p) => p.extract.title);
+  const titled = contentPages(pages).filter((p) => p.extract.title);
   if (titled.length === 0) {
     out.push(
       skip(
@@ -3509,6 +3514,23 @@ function byDeclaredAddress(pages) {
   }
   return [...seen.values()];
 }
+function isPageAddress(address) {
+  const last = address.split("/").pop() ?? "";
+  const ext = /\.([A-Za-z0-9]{1,5})$/.exec(last);
+  return ext === null || PAGE_EXTENSIONS.has(ext[1].toLowerCase());
+}
+var PAGE_EXTENSIONS = /* @__PURE__ */ new Set([
+  "html",
+  "htm",
+  "xhtml",
+  "shtml",
+  "php",
+  "asp",
+  "aspx",
+  "jsp",
+  "jspx",
+  "cfm"
+]);
 function hiddenFromSearch(extract) {
   return /\bnoindex\b/i.test(extract.metas?.["robots"] ?? "");
 }
@@ -4137,11 +4159,17 @@ function runSiteChecks(crawl, checks, businessName = null, dns = null, http = nu
       }
       const at = norm2(abs);
       if (!at) continue;
+      if (!isPageAddress(at)) continue;
       const target = read.get(at);
       if (target && hiddenFromSearch(target.extract)) continue;
       linked.add(target ? aliasAddress(target, read) ?? at : at);
     }
   }
+  const aliasOf = /* @__PURE__ */ new Map();
+  for (const [at, p] of read) {
+    const to = aliasAddress(p, read);
+    if (to) aliasOf.set(at, to);
+  }
   const phones = (checks?.consistency?.phones ?? []).map((p) => p.normalized);
   const headersMeasured = Object.keys(crawl.homeHeaders ?? {}).length > 0;
   return [
@@ -4164,7 +4192,7 @@ function runSiteChecks(crawl, checks, businessName = null, dns = null, http = nu
     ...httpChecks(http),
     ...formInteractionChecks(crawl.pages.map((p) => p.formProbe)),
     ...headerChecks(crawl.homeHeaders ?? {}, headersMeasured),
-    ...sidecarChecks(crawl, linked),
+    ...sidecarChecks(crawl, linked, aliasOf),
     analyticsCheck(pages)
   ];
 }
@@ -4191,4 +4219,4 @@ export {
   tally,
   runSiteChecks
 };
-//# sourceMappingURL=chunk-VELWHWOW.js.map
+//# sourceMappingURL=chunk-WJMPCVOR.js.map