Build: more thorough processing of spec links

Deals with duplicate anchors in the one-page spec.

astro.config.mjs

@@ -12,6 +12,7 @@ import myAutoLinkHeadings from './integrations/my_autolink_headings';
 import mySvgInline from './integrations/my_svg_inline';
 import mySearchIndex from './integrations/my_search_index';
 import myAddTooltips from './integrations/my_add_tooltips';
+import mySpecLinks from './integrations/my_spec_links';
 import myFixupLinks from './integrations/my_fixup_links';
 import myCleanupHtml from './integrations/my_cleanup_html';
 import myHtaccess from './integrations/my_htaccess';
@@ -27,6 +28,7 @@ export default defineConfig({
     mySvgInline({ filePath: 'src/', cachePath: './.svg_cache/' }),
     mySearchIndex(searchOptions),
     myAddTooltips({ constantsFile: 'src/include/constants.json' }),
+    mySpecLinks(),
     myFixupLinks(),
     myCleanupHtml(),
     myHtaccess(),

integrations/my_spec_links.js

@@ -0,0 +1,108 @@
+import { visit, CONTINUE, SKIP } from 'unist-util-visit';
+import GithubSlugger from 'github-slugger';
+
+// Fix up internal links in the one-page annotated spec.
+// Must be configured to run after myAutoLinkHeadings, and before myFixupLinks and myCleanupHtml.
+// The one-page spec is excluded from search index processing, so no need to worry about that.
+
+// Ignore SVGs and anything to do with footnotes (which should be fine without help)
+function isIgnoredElement(node) {
+  return (
+    node.tagName === 'svg' ||
+    (node.tagName === 'a' && node.properties.dataFootnoteRef !== undefined) ||
+    (node.tagName === 'section' && node.properties.dataFootnotes !== undefined)
+  );
+}
+
+// Only headings and <a id="..."> are of interest
+function isTargetElement(node) {
+  return (
+    node.properties?.id !== undefined &&
+    ['a', 'h1', 'h2', 'h3', 'h4', 'h5', 'h6'].includes(node.tagName)
+  );
+}
+
+// We look at all links, but '/part3/' should point to the main book, not the one-page spec
+function isLinkElement(node) {
+  return (
+    node.tagName === 'a' &&
+    node.properties?.href !== undefined &&
+    node.properties.href !== '/part3/'
+  );
+}
+
+// New pages are indicated by comments attached to certain headings in the Markdown
+function isNewPage(node) {
+  return (
+    ['h1', 'h2', 'h3'].includes(node.tagName) &&
+    node.children[node.children.length - 1].type === 'comment'
+  );
+}
+
+function specLinks({ logger }) {
+  return function (tree, file) {
+    if (file.data.astro.frontmatter.path !== '/annotated-spec/') return;
+
+    // We re-slug the slug to handle duplicates
+    const slugger = new GithubSlugger();
+
+    // Pass 1: Build a map of pages and ids/slugs
+    const map = {};
+    let page = '';
+    visit(tree, 'element', (node) => {
+      if (isIgnoredElement(node)) return SKIP;
+
+      if (isTargetElement(node)) {
+        const oldSlug = node.properties.id;
+        const newSlug = slugger.slug(oldSlug);
+        node.properties.id = newSlug;
+        if (isNewPage(node)) {
+          page = node.children[node.children.length - 1].value.trim();
+          map[page] = newSlug;
+        }
+        page || logger.warn('Page is not set when processing ' + oldSlug);
+        map[page + '#' + oldSlug] = newSlug;
+      }
+    });
+
+    // Pass 2: Adjust hrefs - we need two passes in case any references are forward-looking
+    page = '';
+    visit(tree, 'element', (node) => {
+      if (isIgnoredElement(node)) return SKIP;
+
+      if (isNewPage(node)) {
+        page = node.children[node.children.length - 1].value.trim();
+        return CONTINUE;
+      }
+
+      if (isLinkElement(node)) {
+        const oldHref = node.properties.href;
+        let newHref;
+        if (oldHref.startsWith('#')) {
+          newHref = map[page + oldHref];
+        } else if (oldHref.startsWith('/part3/')) {
+          newHref = map[oldHref];
+        } else {
+          return CONTINUE;
+        }
+        newHref || logger.warn('Failed to fix spec link: ' + oldHref);
+        node.properties.href = '#' + newHref;
+      }
+    });
+  };
+}
+
+export default function () {
+  return {
+    name: 'mySpecLinks',
+    hooks: {
+      'astro:config:setup': ({ updateConfig, logger }) => {
+        updateConfig({
+          markdown: {
+            rehypePlugins: [[specLinks, { logger: logger }]],
+          },
+        });
+      },
+    },
+  };
+}

src/loaders/spec-loader.js

@@ -8,7 +8,7 @@ const reEnd = /^# .*<!-- \/part4\/ -->$/dm;
 const preamble =
   '# One Page Annotated Spec\n\n' +
   '**Note:** This page is automatically generated from the chapters ' +
-  'in [Part 3](/part3/). You may find that some internal links are broken.';
+  'in [Part 3](/part3/).';
 
 export function specLoader(fileName) {
   return {
@@ -37,13 +37,10 @@ export function specLoader(fileName) {
 
       let markdown = preamble;
       if (startMatch && endMatch) {
-        // Remove the title - we will replace it
+        // Extract the markdown, removing the title and replacing it with the preamble
         const start = startMatch.indices[0][1] + 1;
         const end = endMatch.indices[0][0];
-        // Extract the spec, add the preamble, and rewrite internal links
-        markdown += allMarkdown
-          .substring(start, end)
-          .replace(/]\(\/part3\/[^#)]*/g, '](');
+        markdown += allMarkdown.substring(start, end);
       } else {
         logger.warn('Creating empty annotated spec');
       }