From 35d1671715f86adc57545c7d938bf1caa35eb983 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Fran=C3=A7ois=20Daoust?= <fd@w3.org>
Date: Fri, 7 Mar 2025 10:51:17 +0100
Subject: [PATCH] Dfns extraction: expand logic that skips imported definitions
 (#1789)

The crawler expected "imported definitions" to use `<dfn>` as the wrapping
element with an inner `<a>` element. The Source map format spec used the
opposite pattern. This update expands the logic accordingly.

The Source map format spec actually swapped back the elements, but there
are a couple of other specs that use the same pattern (Audiobooks and the
JSON-LD API), and more specs may elect to do that in the future as well.

Close #1788.
---
 src/browserlib/extract-dfns.mjs | 4 +++-
 tests/extract-dfns.js           | 6 ++++++
 2 files changed, 9 insertions(+), 1 deletion(-)

diff --git a/src/browserlib/extract-dfns.mjs b/src/browserlib/extract-dfns.mjs
index 3c1efe1d..f5951969 100644
--- a/src/browserlib/extract-dfns.mjs
+++ b/src/browserlib/extract-dfns.mjs
@@ -341,7 +341,9 @@ export default function (spec, idToHeading = {}) {
     // (pending a proper dfns curation process, see:
     // https://github.com/w3c/webref/issues/789)
     .filter(node => {
-      const link = node.querySelector('a[href^="http"]');
+      const link =
+        node.querySelector('a[href^="http"]') ??
+        node.closest('a[href^="http"]');
       return !link ||
         (node.textContent.trim() !== link.textContent.trim()) ||
         (link.href === 'https://www.w3.org/TR/CSS2/syndata.html#vendor-keywords');
diff --git a/tests/extract-dfns.js b/tests/extract-dfns.js
index 72fd06f8..0acc1616 100644
--- a/tests/extract-dfns.js
+++ b/tests/extract-dfns.js
@@ -325,6 +325,12 @@ const tests = [
     changesToBaseDfn: [],
     spec: "html"
   },
+  {
+    title: "ignores definitions imported in the Source map format spec",
+    html: '<li>The <a href="https://infra.spec.whatwg.org/#byte-sequence"><dfn id="external-whatwg-infra-byte-sequence">byte sequence</dfn></a>',
+    changesToBaseDfn: [],
+    spec: "sourcemap"
+  },
   {
     "title": "extracts attribute definition from the SVG2 spec",
     html: `<table class="attrdef def"><tr>