Apply overlapping styling drectives instead of duplicating content

2 files changed, 61 insertions, 8 deletions

diff --git a/src/scraper.test.ts b/src/scraper.test.ts
index a843561..4432231 100644
--- a/src/scraper.test.ts
+++ b/src/scraper.test.ts
@@ -95,7 +95,7 @@ describe("GitHub scraper", () => {
       assert.strictEqual(data.textLines?.[0], "# deploy-pages 🚀");
       assert.strictEqual(
         data.htmlLines?.[0],
-        '<span class="pl-mh"># deploy-pages 🚀</span>deploy-pages 🚀<span class="pl-en">deploy-pages 🚀</span>',
+        '<span class="pl-mh"># <span class="pl-en">deploy-pages 🚀</span></span>',
       );
       const firstLine =
         /^<article class="markdown-body entry-content container-lg" itemprop="text"><div class="markdown-heading" dir="auto"><h1 tabindex="-1" class="heading-element" dir="auto">deploy-pages 🚀<\/h1>/;
@@ -197,6 +197,28 @@ describe("GitHub scraper", () => {
     });
   });
 
+  describe("dustinlyons/nixos-config repository", () => {
+    it("should fetch nix blob with combined overlapping styling directives", async () => {
+      const data = await getGitHubBlob("dustinlyons", "nixos-config", "main", "flake.nix");
+
+      assert.strictEqual(data.repo.owner, "dustinlyons");
+      assert.strictEqual(data.repo.name, "nixos-config");
+      assert.strictEqual(data.path, "flake.nix");
+      assert.strictEqual(data.image, false);
+      assert.strictEqual(data.htmlContent, null);
+      assert.ok(data.htmlLines !== null && data.htmlLines.length > 0);
+
+      // Line 2 has directives [2,13,"pl-c1"] and [2,13,"pl-s1"] with identical ranges.
+      // They should be combined into one span, not duplicated.
+      const line2 = data.htmlLines[1];
+      assert.ok(line2.includes('class="pl-c1 pl-s1"'), `expected combined classes in: ${line2}`);
+      assert.ok(
+        (line2.match(/description/g) ?? []).length === 1,
+        `expected "description" to appear once in: ${line2}`,
+      );
+    });
+  });
+
   describe("rich file types", () => {
     it("should support PNG", async () => {
       const data = await getGitHubBlob("github", "docs", "main", "assets/images/site/logo.png");
diff --git a/src/scraper.ts b/src/scraper.ts
index 5d17e3f..2dcd1d0 100644
--- a/src/scraper.ts
+++ b/src/scraper.ts
@@ -153,15 +153,46 @@ function escapeHtml(text: string): string {
     .replace(/'/g, "&#039;");
 }
 
-function applyStyling(line: string, directives: StylingDirective[]): string {
+// Renders directives[i..] within [rangeStart, rangeEnd], producing nested spans.
+// Returns the rendered content and the index of the first directive not consumed.
+function renderRange(
+  line: string,
+  directives: StylingDirective[],
+  i: number,
+  rangeStart: number,
+  rangeEnd: number,
+): [string, number] {
   let result = "";
-  let lastPos = 0;
-  for (const [start, end, className] of directives) {
-    result += escapeHtml(line.substring(lastPos, start));
-    result += `<span class="${className}">${escapeHtml(line.substring(start, end))}</span>`;
-    lastPos = end;
+  let pos = rangeStart;
+
+  while (i < directives.length) {
+    const [start, end, className] = directives[i];
+    // Stop at directives that start outside or extend past this range.
+    if (start >= rangeEnd || end > rangeEnd) break;
+
+    // Combine classes for directives with identical ranges.
+    let classes = className;
+    let j = i + 1;
+    while (j < directives.length && directives[j][0] === start && directives[j][1] === end) {
+      classes += " " + directives[j][2];
+      j++;
+    }
+
+    result += escapeHtml(line.substring(pos, start));
+    const [innerContent, nextI] = renderRange(line, directives, j, start, end);
+    result += `<span class="${classes}">${innerContent}</span>`;
+    pos = end;
+    i = nextI;
   }
-  result += escapeHtml(line.substring(lastPos));
+
+  result += escapeHtml(line.substring(pos, rangeEnd));
+  return [result, i];
+}
+
+function applyStyling(line: string, directives: StylingDirective[]): string {
+  // Sort by start ascending, then end descending so outer spans precede nested ones.
+  const sorted = [...directives].sort((a, b) => (a[0] !== b[0] ? a[0] - b[0] : b[1] - a[1]));
+  const [result] = renderRange(line, sorted, 0, 0, line.length);
   return result;
 }