From d0dccc7a910a579d8c172c3d5a5a02ea3abd964d Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Thu, 22 Jan 2026 13:27:36 -0500
Subject: Simple web server for viewing trees

---
 cgithub/src/scraper.ts | 91 ++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 91 insertions(+)
 create mode 100644 cgithub/src/scraper.ts

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
new file mode 100644
index 0000000..add03b2
--- /dev/null
+++ b/cgithub/src/scraper.ts
@@ -0,0 +1,91 @@
+export interface GitHubTreeItem {
+  name: string;
+  path: string;
+  contentType: "directory" | "file";
+}
+
+export interface GitHubTree {
+  path: string;
+  branch: string;
+  items: GitHubTreeItem[];
+  repo: {
+    name: string;
+    owner: string;
+    isPublic: boolean;
+  };
+}
+
+export async function getGitHubTree(
+  owner: string,
+  repo: string,
+  branch: string,
+  path: string,
+): Promise<GitHubTree> {
+  const url = `https://github.com/${owner}/${repo}/tree/${branch}/${path}`;
+
+  // GitHub throttles/blocks requests without realistic browser headers.
+  // These headers make the request appear as a standard browser visit.
+  const response = await fetch(url, {
+    headers: {
+      "User-Agent":
+        "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36",
+      Accept:
+        "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
+      "Accept-Language": "en-US,en;q=0.5",
+    },
+  });
+  const html = await response.text();
+
+  // GitHub uses different formats for root vs subdirectories.
+  let payload;
+
+  // Try root format first.
+  const rootFormatMatches = [
+    ...html.matchAll(
+      /<script type="application\/json" data-target="react-partial\.embeddedData">([^<]+)<\/script>/g,
+    ),
+  ];
+  for (const match of rootFormatMatches) {
+    try {
+      const data = JSON.parse(match[1]);
+      if (data.props?.initialPayload?.tree) {
+        payload = data.props.initialPayload;
+        break;
+      }
+    } catch (e) {
+      continue;
+    }
+  }
+
+  // Try subdirectory format.
+  if (!payload) {
+    const dirFormatMatch = html.match(
+      /<script type="application\/json" data-target="react-app\.embeddedData">([^<]+)<\/script>/,
+    );
+    if (dirFormatMatch) {
+      try {
+        const data = JSON.parse(dirFormatMatch[1]);
+        if (data.payload?.tree) {
+          payload = data.payload;
+        }
+      } catch (e) {
+        // Skip invalid JSON
+      }
+    }
+  }
+
+  if (!payload) {
+    throw new Error("Could not find tree data in embedded JSON");
+  }
+
+  return {
+    path: payload.path,
+    branch: payload.refInfo.name,
+    items: payload.tree.items,
+    repo: {
+      name: payload.repo.name,
+      owner: payload.repo.ownerLogin,
+      isPublic: payload.repo.public,
+    },
+  };
+}
-- 
cgit v1.3.1


From eff2e1e7b31f046c974b88b409296af81b9f5615 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Thu, 22 Jan 2026 18:35:27 -0500
Subject: Add blob view for displaying file contents

---
 cgithub/src/index.ts     | 23 ++++++++++++++-
 cgithub/src/scraper.ts   | 77 ++++++++++++++++++++++++++++++++++++++++++++++++
 cgithub/static/style.css | 14 +++++++++
 cgithub/views/blob.eta   | 37 +++++++++++++++++++++++
 cgithub/views/tree.eta   |  2 +-
 5 files changed, 151 insertions(+), 2 deletions(-)
 create mode 100644 cgithub/views/blob.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/index.ts b/cgithub/src/index.ts
index df0a116..a546f13 100644
--- a/cgithub/src/index.ts
+++ b/cgithub/src/index.ts
@@ -2,7 +2,7 @@ import { Hono } from "hono";
 import { serve } from "@hono/node-server";
 import { serveStatic } from "@hono/node-server/serve-static";
 import { Eta } from "eta";
-import { getGitHubTree } from "./scraper.js";
+import { getGitHubTree, getGitHubBlob } from "./scraper.js";
 
 const app = new Hono();
 const eta = new Eta({ views: "./views" });
@@ -30,6 +30,27 @@ app.get("/:owner/:repo/tree/:branch/:path{.*}?", async (c) => {
   }
 });
 
+app.get("/:owner/:repo/blob/:branch/:path{.*}", async (c) => {
+  try {
+    const owner = c.req.param("owner");
+    const repo = c.req.param("repo");
+    const branch = c.req.param("branch");
+    const path = c.req.param("path");
+
+    const data = await getGitHubBlob(owner, repo, branch, path);
+    const html = await eta.renderAsync("blob.eta", data);
+
+    return c.html(html);
+  } catch (error) {
+    const errorMessage =
+      error instanceof Error ? error.message : "Unknown error";
+    return c.html(
+      `<html><body><h1>Error</h1><p>${errorMessage}</p></body></html>`,
+      500,
+    );
+  }
+});
+
 const port = parseInt(process.env.PORT || "3000", 10);
 serve({ fetch: app.fetch, port }, (info) => {
   console.log(`Server running on http://localhost:${info.port}`);
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index add03b2..fa8d655 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -15,6 +15,20 @@ export interface GitHubTree {
   };
 }
 
+export interface GitHubBlob {
+  path: string;
+  branch: string;
+  content: string;
+  rawContent: string;
+  language: string | null;
+  size: number;
+  repo: {
+    name: string;
+    owner: string;
+    isPublic: boolean;
+  };
+}
+
 export async function getGitHubTree(
   owner: string,
   repo: string,
@@ -89,3 +103,66 @@ export async function getGitHubTree(
     },
   };
 }
+
+export async function getGitHubBlob(
+  owner: string,
+  repo: string,
+  branch: string,
+  path: string,
+): Promise<GitHubBlob> {
+  const url = `https://github.com/${owner}/${repo}/blob/${branch}/${path}`;
+
+  const response = await fetch(url, {
+    headers: {
+      "User-Agent":
+        "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36",
+      Accept:
+        "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
+      "Accept-Language": "en-US,en;q=0.5",
+    },
+  });
+  const html = await response.text();
+
+  // GitHub embeds file data in a JSON script tag
+  let payload;
+
+  const match = html.match(
+    /<script type="application\/json" data-target="react-app\.embeddedData">([^<]+)<\/script>/,
+  );
+  if (match) {
+    try {
+      const data = JSON.parse(match[1]);
+      if (data.payload?.blob) {
+        payload = data.payload;
+      }
+    } catch (e) {
+      // Skip invalid JSON
+    }
+  }
+
+  if (!payload) {
+    throw new Error("Could not find blob data in embedded JSON");
+  }
+
+  const blob = payload.blob;
+  const rawLines = blob.rawLines || [];
+  const rawContent = rawLines.join("\n");
+  const highlightedLines = blob.highlightedLines || [];
+  const content = highlightedLines
+    .map((line: any) => line.text || "")
+    .join("\n");
+
+  return {
+    path: payload.path,
+    branch: payload.refInfo.name,
+    content: content || rawContent,
+    rawContent,
+    language: blob.language || null,
+    size: blob.size,
+    repo: {
+      name: payload.repo.name,
+      owner: payload.repo.ownerLogin,
+      isPublic: payload.repo.public,
+    },
+  };
+}
diff --git a/cgithub/static/style.css b/cgithub/static/style.css
index 3e19563..bd36dce 100644
--- a/cgithub/static/style.css
+++ b/cgithub/static/style.css
@@ -1,6 +1,16 @@
 body {
   margin: 0;
   padding: 1rem;
+  font-family: Helvetica, Arial, sans-serif;
+}
+
+pre.blob {
+  margin: 1rem 0;
+  padding: 1rem;
+  border: 1px solid #d0d7de;
+  border-radius: 6px;
+  overflow-x: auto;
+  font-family: monospace;
 }
 
 @media (prefers-color-scheme: dark) {
@@ -11,4 +21,8 @@ body {
   a {
     color: #4493f8;
   }
+  .blob {
+    border-color: #30363d;
+    background: #161b22;
+  }
 }
diff --git a/cgithub/views/blob.eta b/cgithub/views/blob.eta
new file mode 100644
index 0000000..d3c6a56
--- /dev/null
+++ b/cgithub/views/blob.eta
@@ -0,0 +1,37 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+  <meta charset="UTF-8">
+  <title><%= it.repo.owner %>/<%= it.repo.name %> - <%= it.path %></title>
+  <link rel="stylesheet" href="/static/style.css">
+</head>
+<body>
+  <header>
+    <h1><%= it.repo.owner %>/<%= it.repo.name %></h1>
+    <div>branch: <%= it.branch %></div>
+    <div>
+      path: <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/tree/<%= it.branch %>">root</a><%
+      const pathParts = it.path.split('/').filter(p => p);
+      let currentPath = '';
+      for (let i = 0; i < pathParts.length; i++) {
+        const part = pathParts[i];
+        currentPath += (currentPath ? '/' : '') + part;
+        const isLast = i === pathParts.length - 1;
+        if (isLast) {
+          %>/<%= part %><%
+        } else {
+          %>/<a href="/<%= it.repo.owner %>/<%= it.repo.name %>/tree/<%= it.branch %>/<%= currentPath %>"><%= part %></a><%
+        }
+      } %>
+    </div>
+    <% if (it.language) { %>
+      <div>language: <%= it.language %></div>
+    <% } %>
+    <div>size: <%= it.size %> bytes</div>
+  </header>
+
+  <pre class="blob"><code><%= it.rawContent %></code></pre>
+
+  <footer>Generated by cgithub</footer>
+</body>
+</html>
diff --git a/cgithub/views/tree.eta b/cgithub/views/tree.eta
index 4194216..b49668d 100644
--- a/cgithub/views/tree.eta
+++ b/cgithub/views/tree.eta
@@ -30,7 +30,7 @@
           <% if (item.contentType === 'directory') { %>
             <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/tree/<%= it.branch %>/<%= item.path %>"><%= item.name %></a>
           <% } else { %>
-            <%= item.name %>
+            <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/blob/<%= it.branch %>/<%= item.path %>"><%= item.name %></a>
           <% } %>
         </td>
       </tr>
-- 
cgit v1.3.1


From 940c186cd8dd1d2bd12fa9405400f717fe32440e Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Sun, 25 Jan 2026 16:27:48 -0500
Subject: Refactor

---
 cgithub/src/index.ts   |  42 ++--------
 cgithub/src/scraper.ts | 214 ++++++++++++++++++++++---------------------------
 2 files changed, 104 insertions(+), 152 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/index.ts b/cgithub/src/index.ts
index 525ac74..fcd0069 100644
--- a/cgithub/src/index.ts
+++ b/cgithub/src/index.ts
@@ -14,45 +14,15 @@ const eta = new Eta({ views: path.join(__dirname, "..", "views") });
 app.use("/static/*", serveStatic({ root: path.join(__dirname, "..") }));
 
 app.get("/:owner/:repo/tree/:branch/:path{.*}?", async (c) => {
-  try {
-    const owner = c.req.param("owner");
-    const repo = c.req.param("repo");
-    const branch = c.req.param("branch");
-    const path = c.req.param("path") || "";
-
-    const data = await getGitHubTree(owner, repo, branch, path);
-    const html = await eta.renderAsync("tree.eta", data);
-
-    return c.html(html);
-  } catch (error) {
-    const errorMessage =
-      error instanceof Error ? error.message : "Unknown error";
-    return c.html(
-      `<html><body><h1>Error</h1><p>${errorMessage}</p></body></html>`,
-      500,
-    );
-  }
+  const { owner, repo, branch, path = "" } = c.req.param();
+  const data = await getGitHubTree(owner, repo, branch, path);
+  return c.html(await eta.renderAsync("tree.eta", data));
 });
 
 app.get("/:owner/:repo/blob/:branch/:path{.*}", async (c) => {
-  try {
-    const owner = c.req.param("owner");
-    const repo = c.req.param("repo");
-    const branch = c.req.param("branch");
-    const path = c.req.param("path");
-
-    const data = await getGitHubBlob(owner, repo, branch, path);
-    const html = await eta.renderAsync("blob.eta", data);
-
-    return c.html(html);
-  } catch (error) {
-    const errorMessage =
-      error instanceof Error ? error.message : "Unknown error";
-    return c.html(
-      `<html><body><h1>Error</h1><p>${errorMessage}</p></body></html>`,
-      500,
-    );
-  }
+  const { owner, repo, branch, path } = c.req.param();
+  const data = await getGitHubBlob(owner, repo, branch, path);
+  return c.html(await eta.renderAsync("blob.eta", data));
 });
 
 if (process.env.LISTEN_FDS === "1") {
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index fa8d655..d6702cc 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -1,146 +1,137 @@
-export interface GitHubTreeItem {
+interface TreeItem {
+  contentType: "directory" | "file";
   name: string;
   path: string;
-  contentType: "directory" | "file";
 }
 
-export interface GitHubTree {
-  path: string;
-  branch: string;
-  items: GitHubTreeItem[];
+interface RawPayload {
   repo: {
+    ownerLogin: string;
     name: string;
-    owner: string;
-    isPublic: boolean;
+  };
+  refInfo: { name: string };
+  path: string;
+  tree?: {
+    items: TreeItem[];
+  };
+  blob?: {
+    language?: string;
+    rawLines?: string[];
+    highlightedLines?: {
+      text?: string;
+    }[];
+    size: number;
   };
 }
 
-export interface GitHubBlob {
-  path: string;
-  branch: string;
-  content: string;
-  rawContent: string;
-  language: string | null;
-  size: number;
+interface GitHubCommon {
   repo: {
-    name: string;
     owner: string;
-    isPublic: boolean;
+    name: string;
   };
+  branch: string;
+  path: string;
 }
 
-export async function getGitHubTree(
-  owner: string,
-  repo: string,
-  branch: string,
-  path: string,
-): Promise<GitHubTree> {
-  const url = `https://github.com/${owner}/${repo}/tree/${branch}/${path}`;
+export interface GitHubTree extends GitHubCommon {
+  items: TreeItem[];
+}
+
+export interface GitHubBlob extends GitHubCommon {
+  language: string | null;
+  size: number;
+  content: string;
+  rawContent: string;
+}
 
+async function fetchGitHubPage(path: string): Promise<string> {
   // GitHub throttles/blocks requests without realistic browser headers.
   // These headers make the request appear as a standard browser visit.
-  const response = await fetch(url, {
-    headers: {
-      "User-Agent":
-        "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36",
-      Accept:
-        "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
-      "Accept-Language": "en-US,en;q=0.5",
-    },
-  });
-  const html = await response.text();
+  const headers = {
+    "User-Agent":
+      "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36",
+    Accept:
+      "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
+    "Accept-Language": "en-US,en;q=0.5",
+  };
+  const response = await fetch(`https://github.com/${path}`, { headers });
+  return response.text();
+}
 
-  // GitHub uses different formats for root vs subdirectories.
-  let payload;
-
-  // Try root format first.
-  const rootFormatMatches = [
-    ...html.matchAll(
-      /<script type="application\/json" data-target="react-partial\.embeddedData">([^<]+)<\/script>/g,
-    ),
-  ];
-  for (const match of rootFormatMatches) {
+function parsePayload(
+  html: string,
+  dataTarget: string,
+  prop: string[],
+): RawPayload | null {
+  const regex = new RegExp(
+    `<script type="application/json" data-target="${dataTarget}">([^<]+)</script>`,
+    "g",
+  );
+
+  match: for (const match of html.matchAll(regex)) {
+    let data;
     try {
-      const data = JSON.parse(match[1]);
-      if (data.props?.initialPayload?.tree) {
-        payload = data.props.initialPayload;
-        break;
-      }
-    } catch (e) {
+      data = JSON.parse(match[1]);
+    } catch {
       continue;
     }
-  }
-
-  // Try subdirectory format.
-  if (!payload) {
-    const dirFormatMatch = html.match(
-      /<script type="application\/json" data-target="react-app\.embeddedData">([^<]+)<\/script>/,
-    );
-    if (dirFormatMatch) {
-      try {
-        const data = JSON.parse(dirFormatMatch[1]);
-        if (data.payload?.tree) {
-          payload = data.payload;
-        }
-      } catch (e) {
-        // Skip invalid JSON
-      }
+    for (const key of prop) {
+      data = data[key];
+      if (data === undefined) continue match;
     }
+    return data;
   }
+  return null;
+}
 
-  if (!payload) {
-    throw new Error("Could not find tree data in embedded JSON");
-  }
-
+function extractGitHub<T>(payload: RawPayload, extra: T): GitHubCommon & T {
   return {
-    path: payload.path,
-    branch: payload.refInfo.name,
-    items: payload.tree.items,
     repo: {
-      name: payload.repo.name,
       owner: payload.repo.ownerLogin,
-      isPublic: payload.repo.public,
+      name: payload.repo.name,
     },
+    branch: payload.refInfo.name,
+    path: payload.path,
+    ...extra,
   };
 }
 
+export async function getGitHubTree(
+  owner: string,
+  repo: string,
+  branch: string,
+  path: string,
+): Promise<GitHubTree> {
+  const html = await fetchGitHubPage(`${owner}/${repo}/tree/${branch}/${path}`);
+
+  // GitHub uses different formats for root vs subdirectories.
+  const payload =
+    // Root directory format.
+    parsePayload(html, "react-partial.embeddedData", [
+      "props",
+      "initialPayload",
+    ]) ||
+    // Subdirectory format.
+    parsePayload(html, "react-app.embeddedData", ["payload"]);
+
+  if (payload?.tree === undefined) {
+    throw new Error("Could not find tree data in embedded JSON");
+  }
+
+  return extractGitHub(payload, { items: payload.tree.items });
+}
+
 export async function getGitHubBlob(
   owner: string,
   repo: string,
   branch: string,
   path: string,
 ): Promise<GitHubBlob> {
-  const url = `https://github.com/${owner}/${repo}/blob/${branch}/${path}`;
-
-  const response = await fetch(url, {
-    headers: {
-      "User-Agent":
-        "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36",
-      Accept:
-        "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
-      "Accept-Language": "en-US,en;q=0.5",
-    },
-  });
-  const html = await response.text();
+  const html = await fetchGitHubPage(`${owner}/${repo}/blob/${branch}/${path}`);
 
-  // GitHub embeds file data in a JSON script tag
-  let payload;
-
-  const match = html.match(
-    /<script type="application\/json" data-target="react-app\.embeddedData">([^<]+)<\/script>/,
-  );
-  if (match) {
-    try {
-      const data = JSON.parse(match[1]);
-      if (data.payload?.blob) {
-        payload = data.payload;
-      }
-    } catch (e) {
-      // Skip invalid JSON
-    }
-  }
+  const payload = parsePayload(html, "react-app.embeddedData", ["payload"]);
 
-  if (!payload) {
+  if (payload?.blob === undefined) {
     throw new Error("Could not find blob data in embedded JSON");
   }
 
@@ -148,21 +139,12 @@ export async function getGitHubBlob(
   const rawLines = blob.rawLines || [];
   const rawContent = rawLines.join("\n");
   const highlightedLines = blob.highlightedLines || [];
-  const content = highlightedLines
-    .map((line: any) => line.text || "")
-    .join("\n");
+  const content = highlightedLines.map((line) => line.text || "").join("\n");
 
-  return {
-    path: payload.path,
-    branch: payload.refInfo.name,
+  return extractGitHub(payload, {
     content: content || rawContent,
     rawContent,
     language: blob.language || null,
     size: blob.size,
-    repo: {
-      name: payload.repo.name,
-      owner: payload.repo.ownerLogin,
-      isPublic: payload.repo.public,
-    },
-  };
+  });
 }
-- 
cgit v1.3.1


From 0391e279f80f6406502b4069bf2e13d3f13ed56e Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Sun, 25 Jan 2026 16:27:48 -0500
Subject: Repository page with info

---
 cgithub/src/index.ts    |  8 +++++++-
 cgithub/src/scraper.ts  | 51 +++++++++++++++++++++++++++++++++++++++++++++++++
 cgithub/views/_tree.eta | 14 ++++++++++++++
 cgithub/views/repo.eta  | 23 ++++++++++++++++++++++
 cgithub/views/tree.eta  | 15 +--------------
 5 files changed, 96 insertions(+), 15 deletions(-)
 create mode 100644 cgithub/views/_tree.eta
 create mode 100644 cgithub/views/repo.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/index.ts b/cgithub/src/index.ts
index fcd0069..393c5ec 100644
--- a/cgithub/src/index.ts
+++ b/cgithub/src/index.ts
@@ -4,7 +4,7 @@ import { Eta } from "eta";
 import { Hono } from "hono";
 import * as path from "node:path";
 import * as url from "node:url";
-import { getGitHubBlob, getGitHubTree } from "./scraper.js";
+import { getGitHubBlob, getGitHubRepo, getGitHubTree } from "./scraper.js";
 
 const __dirname = path.dirname(url.fileURLToPath(import.meta.url));
 
@@ -13,6 +13,12 @@ const eta = new Eta({ views: path.join(__dirname, "..", "views") });
 
 app.use("/static/*", serveStatic({ root: path.join(__dirname, "..") }));
 
+app.get("/:owner/:repo", async (c) => {
+  const { owner, repo } = c.req.param();
+  const data = await getGitHubRepo(owner, repo);
+  return c.html(await eta.renderAsync("repo.eta", data));
+});
+
 app.get("/:owner/:repo/tree/:branch/:path{.*}?", async (c) => {
   const { owner, repo, branch, path = "" } = c.req.param();
   const data = await getGitHubTree(owner, repo, branch, path);
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index d6702cc..a405139 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -37,6 +37,18 @@ export interface GitHubTree extends GitHubCommon {
   items: TreeItem[];
 }
 
+export interface RepoInfo {
+  description: string | null;
+  website: string | null;
+  stars: string | null;
+  watchers: string | null;
+  forks: string | null;
+}
+
+export interface GitHubRepo extends GitHubTree {
+  info: RepoInfo;
+}
+
 export interface GitHubBlob extends GitHubCommon {
   language: string | null;
   size: number;
@@ -84,6 +96,25 @@ function parsePayload(
   return null;
 }
 
+function parseRepoInfo(html: string): RepoInfo {
+  const descMatch = html.match(/<p[^>]*class="[^"]*f4[^"]*"[^>]*>([\s\S]*?)<\/p>/);
+  const description = descMatch?.[1]?.replace(/<[^>]+>/g, "").trim() || null;
+  const websiteMatch = html.match(
+    /<a[^>]*rel="noopener noreferrer nofollow"[^>]*href="([^"]+)"/,
+  );
+  const starsMatch = html.match(/id="repo-stars-counter-star"[^>]*>([^<]+)/);
+  const watchersMatch = html.match(/<strong>(\d+)<\/strong>\s*watching/);
+  const forksMatch = html.match(/id="repo-network-counter"[^>]*>([^<]+)/);
+
+  return {
+    description,
+    website: websiteMatch?.[1] || null,
+    stars: starsMatch?.[1]?.trim() || null,
+    watchers: watchersMatch?.[1] || null,
+    forks: forksMatch?.[1]?.trim() || null,
+  };
+}
+
 function extractGitHub<T>(payload: RawPayload, extra: T): GitHubCommon & T {
   return {
     repo: {
@@ -96,6 +127,26 @@ function extractGitHub<T>(payload: RawPayload, extra: T): GitHubCommon & T {
   };
 }
 
+export async function getGitHubRepo(
+  owner: string,
+  repo: string,
+): Promise<GitHubRepo> {
+  const html = await fetchGitHubPage(`${owner}/${repo}`);
+
+  const payload = parsePayload(html, "react-partial.embeddedData", [
+    "props",
+    "initialPayload",
+  ]);
+
+  if (payload?.tree === undefined) {
+    throw new Error("Could not find tree data in embedded JSON");
+  }
+
+  const info = parseRepoInfo(html);
+
+  return extractGitHub(payload, { items: payload.tree.items, info });
+}
+
 export async function getGitHubTree(
   owner: string,
   repo: string,
diff --git a/cgithub/views/_tree.eta b/cgithub/views/_tree.eta
new file mode 100644
index 0000000..ac7b2b1
--- /dev/null
+++ b/cgithub/views/_tree.eta
@@ -0,0 +1,14 @@
+<table>
+  <% for (const item of it.items) { %>
+    <tr>
+      <td>
+        <span class="icon"><%= item.contentType === 'directory' ? '📁' : '📄' %></span>
+        <% if (item.contentType === 'directory') { %>
+          <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/tree/<%= it.branch %>/<%= item.path %>"><%= item.name %></a>
+        <% } else { %>
+          <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/blob/<%= it.branch %>/<%= item.path %>"><%= item.name %></a>
+        <% } %>
+      </td>
+    </tr>
+  <% } %>
+</table>
diff --git a/cgithub/views/repo.eta b/cgithub/views/repo.eta
new file mode 100644
index 0000000..0b345df
--- /dev/null
+++ b/cgithub/views/repo.eta
@@ -0,0 +1,23 @@
+<% layout("./layout") %>
+
+<% if (it.info) { %>
+  <div class="repo-info">
+    <% if (it.info.description) { %>
+      <div class="description"><%= it.info.description %></div>
+    <% } %>
+    <% if (it.info.website) { %>
+      <div><a href="<%= it.info.website %>" rel="noopener noreferrer"><%= it.info.website %></a></div>
+    <% } %>
+    <% if (it.info.stars) { %>
+      <span>★ <%= it.info.stars %></span>
+    <% } %>
+    <% if (it.info.watchers) { %>
+      <span>👁 <%= it.info.watchers %></span>
+    <% } %>
+    <% if (it.info.forks) { %>
+      <span>🔀 <%= it.info.forks %></span>
+    <% } %>
+  </div>
+<% } %>
+
+<%~ include("./_tree", it) %>
diff --git a/cgithub/views/tree.eta b/cgithub/views/tree.eta
index 911b7ee..7068754 100644
--- a/cgithub/views/tree.eta
+++ b/cgithub/views/tree.eta
@@ -1,16 +1,3 @@
 <% layout("./layout") %>
 
-<table>
-  <% for (const item of it.items) { %>
-    <tr>
-      <td>
-        <span class="icon"><%= item.contentType === 'directory' ? '📁' : '📄' %></span>
-        <% if (item.contentType === 'directory') { %>
-          <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/tree/<%= it.branch %>/<%= item.path %>"><%= item.name %></a>
-        <% } else { %>
-          <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/blob/<%= it.branch %>/<%= item.path %>"><%= item.name %></a>
-        <% } %>
-      </td>
-    </tr>
-  <% } %>
-</table>
+<%~ include("./_tree", it) %>
-- 
cgit v1.3.1


From 1631e0f74a488a8cf42e3fe699e96e9389bbb596 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Mon, 26 Jan 2026 11:47:06 -0500
Subject: Overview (README) in trees

---
 cgithub/.prettierrc      |  3 +++
 cgithub/src/scraper.ts   | 56 +++++++++++++++++++++++++++---------------------
 cgithub/static/style.css | 43 +++++++++++++++++++++++++++++++++++--
 cgithub/views/_tree.eta  |  9 ++++++++
 cgithub/views/blob.eta   |  3 +++
 5 files changed, 88 insertions(+), 26 deletions(-)
 create mode 100644 cgithub/.prettierrc

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/.prettierrc b/cgithub/.prettierrc
new file mode 100644
index 0000000..de753c5
--- /dev/null
+++ b/cgithub/.prettierrc
@@ -0,0 +1,3 @@
+{
+  "printWidth": 100
+}
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index a405139..c1bdef5 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -4,6 +4,13 @@ interface TreeItem {
   path: string;
 }
 
+interface OverviewFile {
+  displayName: string;
+  preferredFileType: string;
+  richText?: string;
+  loaded: boolean;
+}
+
 interface RawPayload {
   repo: {
     ownerLogin: string;
@@ -22,6 +29,9 @@ interface RawPayload {
     }[];
     size: number;
   };
+  overview?: {
+    overviewFiles?: OverviewFile[];
+  };
 }
 
 interface GitHubCommon {
@@ -35,6 +45,7 @@ interface GitHubCommon {
 
 export interface GitHubTree extends GitHubCommon {
   items: TreeItem[];
+  overviewHtml?: Record<string, string>;
 }
 
 export interface RepoInfo {
@@ -62,19 +73,14 @@ async function fetchGitHubPage(path: string): Promise<string> {
   const headers = {
     "User-Agent":
       "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36",
-    Accept:
-      "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
+    Accept: "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
     "Accept-Language": "en-US,en;q=0.5",
   };
   const response = await fetch(`https://github.com/${path}`, { headers });
   return response.text();
 }
 
-function parsePayload(
-  html: string,
-  dataTarget: string,
-  prop: string[],
-): RawPayload | null {
+function parsePayload(html: string, dataTarget: string, prop: string[]): RawPayload | null {
   const regex = new RegExp(
     `<script type="application/json" data-target="${dataTarget}">([^<]+)</script>`,
     "g",
@@ -99,9 +105,7 @@ function parsePayload(
 function parseRepoInfo(html: string): RepoInfo {
   const descMatch = html.match(/<p[^>]*class="[^"]*f4[^"]*"[^>]*>([\s\S]*?)<\/p>/);
   const description = descMatch?.[1]?.replace(/<[^>]+>/g, "").trim() || null;
-  const websiteMatch = html.match(
-    /<a[^>]*rel="noopener noreferrer nofollow"[^>]*href="([^"]+)"/,
-  );
+  const websiteMatch = html.match(/<a[^>]*rel="noopener noreferrer nofollow"[^>]*href="([^"]+)"/);
   const starsMatch = html.match(/id="repo-stars-counter-star"[^>]*>([^<]+)/);
   const watchersMatch = html.match(/<strong>(\d+)<\/strong>\s*watching/);
   const forksMatch = html.match(/id="repo-network-counter"[^>]*>([^<]+)/);
@@ -115,6 +119,16 @@ function parseRepoInfo(html: string): RepoInfo {
   };
 }
 
+function extractOverviewHtml(payload: RawPayload): Record<string, string> | undefined {
+  const result: Record<string, string> = {};
+  for (const file of payload.overview?.overviewFiles ?? []) {
+    if (file.loaded && file.richText !== undefined) {
+      result[file.displayName] = file.richText;
+    }
+  }
+  return Object.keys(result).length > 0 ? result : undefined;
+}
+
 function extractGitHub<T>(payload: RawPayload, extra: T): GitHubCommon & T {
   return {
     repo: {
@@ -127,24 +141,19 @@ function extractGitHub<T>(payload: RawPayload, extra: T): GitHubCommon & T {
   };
 }
 
-export async function getGitHubRepo(
-  owner: string,
-  repo: string,
-): Promise<GitHubRepo> {
+export async function getGitHubRepo(owner: string, repo: string): Promise<GitHubRepo> {
   const html = await fetchGitHubPage(`${owner}/${repo}`);
 
-  const payload = parsePayload(html, "react-partial.embeddedData", [
-    "props",
-    "initialPayload",
-  ]);
+  const payload = parsePayload(html, "react-partial.embeddedData", ["props", "initialPayload"]);
 
   if (payload?.tree === undefined) {
     throw new Error("Could not find tree data in embedded JSON");
   }
 
   const info = parseRepoInfo(html);
+  const overviewHtml = extractOverviewHtml(payload);
 
-  return extractGitHub(payload, { items: payload.tree.items, info });
+  return extractGitHub(payload, { items: payload.tree.items, info, overviewHtml });
 }
 
 export async function getGitHubTree(
@@ -158,10 +167,7 @@ export async function getGitHubTree(
   // GitHub uses different formats for root vs subdirectories.
   const payload =
     // Root directory format.
-    parsePayload(html, "react-partial.embeddedData", [
-      "props",
-      "initialPayload",
-    ]) ||
+    parsePayload(html, "react-partial.embeddedData", ["props", "initialPayload"]) ||
     // Subdirectory format.
     parsePayload(html, "react-app.embeddedData", ["payload"]);
 
@@ -169,7 +175,9 @@ export async function getGitHubTree(
     throw new Error("Could not find tree data in embedded JSON");
   }
 
-  return extractGitHub(payload, { items: payload.tree.items });
+  const overviewHtml = extractOverviewHtml(payload);
+
+  return extractGitHub(payload, { items: payload.tree.items, overviewHtml });
 }
 
 export async function getGitHubBlob(
diff --git a/cgithub/static/style.css b/cgithub/static/style.css
index 1229fce..ff787d6 100644
--- a/cgithub/static/style.css
+++ b/cgithub/static/style.css
@@ -3,6 +3,7 @@ body {
   font-family: Helvetica, Arial, sans-serif;
 
   --border-color: #d0d7de;
+  --shade-background: #eeeeee;
 }
 
 @media (prefers-color-scheme: dark) {
@@ -23,10 +24,20 @@ a {
   text-decoration: none;
 }
 
-header, .box {
+body > header, .box {
+  --box-padding: .5rem;
+
   margin: 1rem 0;
   border: 1px solid var(--border-color);
-  padding: .5rem;
+  padding: var(--box-padding);
+
+  > header:first-child {
+    margin: calc(-1 * var(--box-padding));
+    margin-bottom: 0;
+    padding: .5rem;
+    border-bottom: 1px solid var(--border-color);
+    font-weight: bold;
+  }
 }
 
 footer {
@@ -64,3 +75,31 @@ pre.box {
   /* Instead of scrolling overflow for long lines, grow the box. */
   width: fit-content;
 }
+
+.overview {
+  overflow-wrap: break-word;
+  &.box {
+    --box-padding: 1rem;
+  }
+
+  a {
+    text-decoration: underline;
+  }
+
+  h1, h2 {
+    border-bottom: 1px solid var(--border-color);
+  }
+
+  img {
+    max-width: 100%;
+  }
+
+  code {
+    background: var(--shade-background);
+  }
+
+  /* TODO Strip permalinks? */
+  svg.octicon-link {
+    display: none;
+  }
+}
diff --git a/cgithub/views/_tree.eta b/cgithub/views/_tree.eta
index 2549bb8..f287c03 100644
--- a/cgithub/views/_tree.eta
+++ b/cgithub/views/_tree.eta
@@ -13,3 +13,12 @@
     </tr>
   <% } %>
 </table>
+
+<% if (it.overviewHtml) { %>
+  <% for (const [filename, html] of Object.entries(it.overviewHtml)) { %>
+    <div class="overview box">
+      <header><%= filename %></header>
+      <%~ html %>
+    </div>
+  <% } %>
+<% } %>
diff --git a/cgithub/views/blob.eta b/cgithub/views/blob.eta
index 9b478b5..7328a62 100644
--- a/cgithub/views/blob.eta
+++ b/cgithub/views/blob.eta
@@ -7,6 +7,9 @@
   <% if (it.size !== undefined) { %>
     <div>size: <%= it.size %> bytes</div>
   <% } %>
+  <div>
+    <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/raw/<%= it.branch %>/<%= it.path %>">Raw</a>
+  </div>
 </div>
 
 <pre class="box"><code><%= it.rawContent %></code></pre>
-- 
cgit v1.3.1


From 744f42a2ae7968881e4d8f1ce0ec0d5209c8d7a4 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Mon, 26 Jan 2026 13:27:28 -0500
Subject: Show releases

---
 cgithub/src/scraper.ts | 10 ++++++++++
 cgithub/views/repo.eta |  3 +++
 2 files changed, 13 insertions(+)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index c1bdef5..4ffdc43 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -54,6 +54,7 @@ export interface RepoInfo {
   stars: string | null;
   watchers: string | null;
   forks: string | null;
+  numReleases?: number;
 }
 
 export interface GitHubRepo extends GitHubTree {
@@ -110,12 +111,21 @@ function parseRepoInfo(html: string): RepoInfo {
   const watchersMatch = html.match(/<strong>(\d+)<\/strong>\s*watching/);
   const forksMatch = html.match(/id="repo-network-counter"[^>]*>([^<]+)/);
 
+  let numReleases: number | undefined;
+  const releaseCountMatch = html.match(
+    /<a[^>]*href="[^"]*\/releases"[^>]*>[\s\S]*?<span[^>]*title="(\d+)"[^>]*(?:class="[^"]*Counter|data-view-component="true")[^>]*>/,
+  );
+  if (releaseCountMatch) {
+    numReleases = parseInt(releaseCountMatch[1], 10);
+  }
+
   return {
     description,
     website: websiteMatch?.[1] || null,
     stars: starsMatch?.[1]?.trim() || null,
     watchers: watchersMatch?.[1] || null,
     forks: forksMatch?.[1]?.trim() || null,
+    numReleases,
   };
 }
 
diff --git a/cgithub/views/repo.eta b/cgithub/views/repo.eta
index 4965e75..5bebb95 100644
--- a/cgithub/views/repo.eta
+++ b/cgithub/views/repo.eta
@@ -17,6 +17,9 @@
     <% if (it.info.forks) { %>
       <span class="count"><img src="/static/icons/git-fork.svg" alt="forks" class="icon"> <%= it.info.forks %></span>
     <% } %>
+    <% if (it.info.numReleases) { %>
+      <div><a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases" rel="noreferrer"><%= it.info.numReleases %> releases</a></div>
+    <% } %>
   </div>
 <% } %>
 
-- 
cgit v1.3.1


From dac0a13f247fa17667ac5983ec65df0756e0829f Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Mon, 26 Jan 2026 15:36:23 -0500
Subject: GitHub scraper test

---
 cgithub/package.json        |  3 +-
 cgithub/src/scraper.test.ts | 70 +++++++++++++++++++++++++++++++++++++++++++++
 cgithub/src/scraper.ts      |  7 +++--
 3 files changed, 77 insertions(+), 3 deletions(-)
 create mode 100644 cgithub/src/scraper.test.ts

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/package.json b/cgithub/package.json
index b3d778b..71dfb4d 100644
--- a/cgithub/package.json
+++ b/cgithub/package.json
@@ -5,7 +5,8 @@
   "type": "module",
   "scripts": {
     "dev": "node --watch --watch-preserve-output src/index.ts",
-    "build": "rm -rf dist/* && esbuild --bundle --format=esm --platform=node --outdir=dist/build --sourcemap src/index.ts && cp -R static views dist/"
+    "build": "rm -rf dist/* && esbuild --bundle --format=esm --platform=node --outdir=dist/build --sourcemap src/index.ts && cp -R static views dist/",
+    "test": "node --test src/scraper.test.ts"
   },
   "dependencies": {
     "@hono/node-server": "^1.13.7",
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
new file mode 100644
index 0000000..3784190
--- /dev/null
+++ b/cgithub/src/scraper.test.ts
@@ -0,0 +1,70 @@
+import assert from "node:assert";
+import { describe, it } from "node:test";
+import { getGitHubBlob, getGitHubRepo, getGitHubTree } from "./scraper.ts";
+
+describe("GitHub scraper", () => {
+  describe("actions/deploy-pages repository", () => {
+    it("should fetch repository", async () => {
+      const data = await getGitHubRepo("actions", "deploy-pages");
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+      assert.strictEqual(data.branch, "main");
+      assert.strictEqual(data.path, "/");
+
+      assert.strictEqual(
+        data.info.description,
+        "GitHub Action to publish artifacts to GitHub Pages for deployments",
+      );
+      assert.strictEqual(data.info.website, "https://pages.github.com");
+      assert.strictEqual(data.info.stars?.length, 3);
+      assert.strictEqual(data.info.watchers?.length, 2);
+      assert.strictEqual(data.info.forks?.length, 3);
+      assert.strictEqual(data.info.numReleases, 36);
+
+      assert.ok(data.items.length > 0);
+      assert.ok(Object.keys(data.overviewHtml ?? {}).includes("README.md"));
+    });
+
+    it("should fetch tree", async () => {
+      const data = await getGitHubTree("actions", "deploy-pages", "main", "src");
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+      assert.strictEqual(data.branch, "main");
+      assert.strictEqual(data.path, "src");
+
+      assert.deepStrictEqual(data.items, [
+        {
+          name: "__tests__",
+          path: "src/__tests__",
+          contentType: "directory",
+        },
+        {
+          name: "internal",
+          path: "src/internal",
+          contentType: "directory",
+        },
+        {
+          name: "index.js",
+          path: "src/index.js",
+          contentType: "file",
+        },
+      ]);
+    });
+
+    it("should fetch blob", async () => {
+      const data = await getGitHubBlob("actions", "deploy-pages", "main", "LICENSE");
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+      assert.strictEqual(data.branch, "main");
+      assert.strictEqual(data.path, "LICENSE");
+
+      assert.strictEqual(data.size, 1068);
+      assert.strictEqual(data.language, "Text");
+      assert.match(data.rawContent, /^MIT License\n\n/);
+      assert.match(data.content, /^MIT License\n\n/);
+    });
+  });
+});
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 4ffdc43..4feda73 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -63,7 +63,7 @@ export interface GitHubRepo extends GitHubTree {
 
 export interface GitHubBlob extends GitHubCommon {
   language: string | null;
-  size: number;
+  size: number | null;
   content: string;
   rawContent: string;
 }
@@ -210,10 +210,13 @@ export async function getGitHubBlob(
   const highlightedLines = blob.highlightedLines || [];
   const content = highlightedLines.map((line) => line.text || "").join("\n");
 
+  // Calculate size from raw content if not provided
+  const size = blob.size ?? (rawContent.length > 0 ? Buffer.byteLength(rawContent, "utf8") : null);
+
   return extractGitHub(payload, {
     content: content || rawContent,
     rawContent,
     language: blob.language || null,
-    size: blob.size,
+    size,
   });
 }
-- 
cgit v1.3.1


From 04bbba0a7b71f1fb090d3e3a73110a00c36bf765 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Tue, 27 Jan 2026 11:23:11 -0500
Subject: Propagate GitHub HTTP errors

Render `error.eta` on HTTP error from GitHub.

Implemented with Gemini.
---
 cgithub/GEMINI.md           |  1 +
 cgithub/package.json        |  2 +-
 cgithub/src/index.ts        | 31 +++++++++++++++++++++++--------
 cgithub/src/scraper.test.ts | 12 +++++++++++-
 cgithub/src/scraper.ts      | 11 +++++++++++
 cgithub/views/error.eta     | 11 +++++++++++
 6 files changed, 58 insertions(+), 10 deletions(-)
 create mode 120000 cgithub/GEMINI.md
 create mode 100644 cgithub/views/error.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/GEMINI.md b/cgithub/GEMINI.md
new file mode 120000
index 0000000..681311e
--- /dev/null
+++ b/cgithub/GEMINI.md
@@ -0,0 +1 @@
+CLAUDE.md
\ No newline at end of file
diff --git a/cgithub/package.json b/cgithub/package.json
index 6faa90b..3d339a4 100644
--- a/cgithub/package.json
+++ b/cgithub/package.json
@@ -7,7 +7,7 @@
   "scripts": {
     "dev": "node --watch --watch-preserve-output src/index.ts",
     "build": "rm -rf dist/* && esbuild --bundle --format=esm --platform=node --outdir=dist/build --sourcemap src/index.ts && cp -R static views dist/",
-    "test": "node --test src/scraper.test.ts"
+    "test": "node --test --test-concurrency=4 src/scraper.test.ts"
   },
   "dependencies": {
     "@hono/node-server": "^1.13.7",
diff --git a/cgithub/src/index.ts b/cgithub/src/index.ts
index 388b1e0..51e8cf2 100644
--- a/cgithub/src/index.ts
+++ b/cgithub/src/index.ts
@@ -1,10 +1,11 @@
 import { createAdaptorServer, serve } from "@hono/node-server";
 import { serveStatic } from "@hono/node-server/serve-static";
 import { Eta } from "eta";
-import { Hono } from "hono";
+import { type Context, Hono } from "hono";
+import type { StatusCode } from "hono/utils/http-status";
 import path from "node:path";
 import url from "node:url";
-import { getGitHubBlob, getGitHubRepo, getGitHubTree } from "./scraper.ts";
+import { getGitHubBlob, getGitHubRepo, getGitHubTree, GitHubHTTPError } from "./scraper.ts";
 
 const __dirname = path.dirname(url.fileURLToPath(import.meta.url));
 
@@ -17,22 +18,36 @@ app.get("/", async (c) => {
   return c.html(await eta.renderAsync("home.eta", {}));
 });
 
+async function tryRender<T extends object>(c: Context, template: string, promise: Promise<T>) {
+  let data: T;
+  try {
+    data = await promise;
+  } catch (e) {
+    if (e instanceof GitHubHTTPError) {
+      c.status(e.status as StatusCode);
+      const message = `GitHub responded with HTTP ${e.status} ${e.message}`;
+      return c.html(await eta.renderAsync("error.eta", { title: e.message, message }));
+    } else {
+      c.status(500);
+      return c.html(await eta.renderAsync("error.eta", { message: "" + e }));
+    }
+  }
+  return c.html(await eta.renderAsync(template, data));
+}
+
 app.get("/:owner/:repo", async (c) => {
   const { owner, repo } = c.req.param();
-  const data = await getGitHubRepo(owner, repo);
-  return c.html(await eta.renderAsync("repo.eta", data));
+  return tryRender(c, "repo.eta", getGitHubRepo(owner, repo));
 });
 
 app.get("/:owner/:repo/tree/:branch/:path{.*}?", async (c) => {
   const { owner, repo, branch, path = "" } = c.req.param();
-  const data = await getGitHubTree(owner, repo, branch, path);
-  return c.html(await eta.renderAsync("tree.eta", data));
+  return tryRender(c, "tree.eta", getGitHubTree(owner, repo, branch, path));
 });
 
 app.get("/:owner/:repo/blob/:branch/:path{.*}", async (c) => {
   const { owner, repo, branch, path } = c.req.param();
-  const data = await getGitHubBlob(owner, repo, branch, path);
-  return c.html(await eta.renderAsync("blob.eta", data));
+  return tryRender(c, "blob.eta", getGitHubBlob(owner, repo, branch, path));
 });
 
 app.get("/:owner/:repo/raw/:branch/:path{.*}", async (c) => {
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 3784190..adf70ae 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -1,6 +1,6 @@
 import assert from "node:assert";
 import { describe, it } from "node:test";
-import { getGitHubBlob, getGitHubRepo, getGitHubTree } from "./scraper.ts";
+import { getGitHubBlob, getGitHubRepo, getGitHubTree, GitHubHTTPError } from "./scraper.ts";
 
 describe("GitHub scraper", () => {
   describe("actions/deploy-pages repository", () => {
@@ -67,4 +67,14 @@ describe("GitHub scraper", () => {
       assert.match(data.content, /^MIT License\n\n/);
     });
   });
+
+  describe("error handling", () => {
+    it("should throw for non-existent repository", async () => {
+      await assert.rejects(getGitHubRepo("nosuchowner", "nosuchrepo"), (err) => {
+        assert(err instanceof GitHubHTTPError, "error should be an HTTPError");
+        assert.strictEqual(err.status, 404);
+        return true;
+      });
+    });
+  });
 });
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 4feda73..b76f8c8 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -1,3 +1,13 @@
+export class GitHubHTTPError extends Error {
+  status: number;
+
+  constructor(status: number, message: string) {
+    super(message);
+    this.name = "GitHubHTTPError";
+    this.status = status;
+  }
+}
+
 interface TreeItem {
   contentType: "directory" | "file";
   name: string;
@@ -78,6 +88,7 @@ async function fetchGitHubPage(path: string): Promise<string> {
     "Accept-Language": "en-US,en;q=0.5",
   };
   const response = await fetch(`https://github.com/${path}`, { headers });
+  if (!response.ok) throw new GitHubHTTPError(response.status, response.statusText);
   return response.text();
 }
 
diff --git a/cgithub/views/error.eta b/cgithub/views/error.eta
new file mode 100644
index 0000000..784a843
--- /dev/null
+++ b/cgithub/views/error.eta
@@ -0,0 +1,11 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+  <meta charset="UTF-8">
+  <title><%= it.title ?? it.message %></title>
+  <link rel="stylesheet" href="/static/style.css">
+</head>
+<body>
+  <%= it.message %>
+</body>
+</html>
-- 
cgit v1.3.1


From d29bdad2f5c8326d111901373f0a56b3a09a2cd8 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Tue, 27 Jan 2026 15:05:36 -0500
Subject: List issues

---
 cgithub/CLAUDE.md           | 18 ++++------
 cgithub/src/index.ts        | 13 ++++++-
 cgithub/src/scraper.test.ts | 21 ++++++++++-
 cgithub/src/scraper.ts      | 85 +++++++++++++++++++++++++++++++++++++++++++--
 cgithub/static/style.css    | 40 +++++++++++++++++++++
 cgithub/views/issues.eta    | 20 +++++++++++
 cgithub/views/layout.eta    |  4 +++
 cgithub/views/repo.eta      |  2 +-
 8 files changed, 185 insertions(+), 18 deletions(-)
 create mode 100644 cgithub/views/issues.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/CLAUDE.md b/cgithub/CLAUDE.md
index e258e55..8feb529 100644
--- a/cgithub/CLAUDE.md
+++ b/cgithub/CLAUDE.md
@@ -1,10 +1,6 @@
-# CLAUDE.md
-
-This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
-
 ## Project Overview
 
-cgithub is a lightweight GitHub alternative frontend that scrapes GitHub's server-rendered HTML instead of using the API. This approach avoids API rate limits entirely by extracting embedded JSON data from `<script type="application/json">` tags that GitHub includes for hydration.
+cgithub is a lightweight GitHub alternative frontend that scrapes GitHub's server-rendered HTML instead of using the API. Data is usually extracted from embedded JSON data from `<script type="application/json">` tags that GitHub includes for hydration.
 
 ## Commands
 
@@ -15,8 +11,8 @@ pnpm run dev
 # Production build (bundles to dist/)
 pnpm run build
 
-# Deploy to production server
-make deploy
+# Run tests
+pnpm test
 ```
 
 ## Architecture
@@ -30,13 +26,11 @@ make deploy
 **Templates:** `views/*.eta` - Eta templates with layout inheritance (`layout.eta` as base)
 
 **Routes:**
-- `GET /` - Home page
 - `GET /:owner/:repo` - Repository info with directory listing
 - `GET /:owner/:repo/tree/:branch/:path*` - Directory listing
 - `GET /:owner/:repo/blob/:branch/:path` - File content view
 
-**Scraper pattern:** GitHub embeds tree/blob data in two formats depending on page type:
-- `react-partial.embeddedData` (props.initialPayload) for root directories
-- `react-app.embeddedData` (payload) for subdirectories and blobs
+## Instructions
 
-The scraper uses browser-like headers to avoid throttling.
+- Use `for (const ... of ...)` instead of `.forEach()`
+- Update `scraper.test.ts`
\ No newline at end of file
diff --git a/cgithub/src/index.ts b/cgithub/src/index.ts
index 51e8cf2..d4229fd 100644
--- a/cgithub/src/index.ts
+++ b/cgithub/src/index.ts
@@ -5,7 +5,13 @@ import { type Context, Hono } from "hono";
 import type { StatusCode } from "hono/utils/http-status";
 import path from "node:path";
 import url from "node:url";
-import { getGitHubBlob, getGitHubRepo, getGitHubTree, GitHubHTTPError } from "./scraper.ts";
+import {
+  getGitHubBlob,
+  getGitHubIssues,
+  getGitHubRepo,
+  getGitHubTree,
+  GitHubHTTPError,
+} from "./scraper.ts";
 
 const __dirname = path.dirname(url.fileURLToPath(import.meta.url));
 
@@ -56,6 +62,11 @@ app.get("/:owner/:repo/raw/:branch/:path{.*}", async (c) => {
   return c.redirect(`https://raw.githubusercontent.com/${owner}/${repo}/${branch}/${path}`);
 });
 
+app.get("/:owner/:repo/issues", async (c) => {
+  const { owner, repo } = c.req.param();
+  return tryRender(c, "issues.eta", getGitHubIssues(owner, repo));
+});
+
 app.all("*", async (c) => {
   const path = c.req.path;
   c.header("Referrer-Policy", "no-referrer");
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index adf70ae..e7e65ba 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -1,6 +1,12 @@
 import assert from "node:assert";
 import { describe, it } from "node:test";
-import { getGitHubBlob, getGitHubRepo, getGitHubTree, GitHubHTTPError } from "./scraper.ts";
+import {
+  getGitHubBlob,
+  getGitHubIssues,
+  getGitHubRepo,
+  getGitHubTree,
+  GitHubHTTPError,
+} from "./scraper.ts";
 
 describe("GitHub scraper", () => {
   describe("actions/deploy-pages repository", () => {
@@ -66,6 +72,19 @@ describe("GitHub scraper", () => {
       assert.match(data.rawContent, /^MIT License\n\n/);
       assert.match(data.content, /^MIT License\n\n/);
     });
+
+    it("should fetch issues", async () => {
+      const data = await getGitHubIssues("actions", "deploy-pages");
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+
+      const issue402 = data.issues.find((issue) => issue.number === 402);
+      assert.ok(issue402);
+      assert.strictEqual(issue402.title, "Dry Run");
+      assert.strictEqual(issue402.state, "OPEN");
+      assert.strictEqual(issue402.createdAt, "2025-07-03T19:07:08Z");
+    });
   });
 
   describe("error handling", () => {
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index b76f8c8..bbf7511 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -21,6 +21,38 @@ interface OverviewFile {
   loaded: boolean;
 }
 
+interface IssueNode {
+  number: number;
+  title: string;
+  author?: {
+    login: string;
+  };
+  createdAt: string;
+  state: string;
+  labels?: {
+    edges: {
+      node: {
+        name: string;
+      };
+    }[];
+  };
+}
+
+interface IssueIndexPageQuery {
+  queryName: "IssueIndexPageQuery";
+  result: {
+    data: {
+      repository: {
+        search: {
+          edges: {
+            node: IssueNode;
+          }[];
+        };
+      };
+    };
+  };
+}
+
 interface RawPayload {
   repo: {
     ownerLogin: string;
@@ -42,6 +74,8 @@ interface RawPayload {
   overview?: {
     overviewFiles?: OverviewFile[];
   };
+  // Actually there may be other query types.
+  preloadedQueries?: IssueIndexPageQuery[];
 }
 
 interface GitHubCommon {
@@ -49,11 +83,14 @@ interface GitHubCommon {
     owner: string;
     name: string;
   };
+}
+
+interface GitHubNav extends GitHubCommon {
   branch: string;
   path: string;
 }
 
-export interface GitHubTree extends GitHubCommon {
+export interface GitHubTree extends GitHubNav {
   items: TreeItem[];
   overviewHtml?: Record<string, string>;
 }
@@ -71,13 +108,26 @@ export interface GitHubRepo extends GitHubTree {
   info: RepoInfo;
 }
 
-export interface GitHubBlob extends GitHubCommon {
+export interface GitHubBlob extends GitHubNav {
   language: string | null;
   size: number | null;
   content: string;
   rawContent: string;
 }
 
+interface Issue {
+  number: number;
+  title: string;
+  state: string;
+  createdAt: string;
+  numReplies?: number;
+  numPRs?: number;
+}
+
+export interface GitHubIssues extends GitHubCommon {
+  issues: Issue[];
+}
+
 async function fetchGitHubPage(path: string): Promise<string> {
   // GitHub throttles/blocks requests without realistic browser headers.
   // These headers make the request appear as a standard browser visit.
@@ -150,7 +200,7 @@ function extractOverviewHtml(payload: RawPayload): Record<string, string> | unde
   return Object.keys(result).length > 0 ? result : undefined;
 }
 
-function extractGitHub<T>(payload: RawPayload, extra: T): GitHubCommon & T {
+function extractGitHub<T>(payload: RawPayload, extra: T): GitHubNav & T {
   return {
     repo: {
       owner: payload.repo.ownerLogin,
@@ -231,3 +281,32 @@ export async function getGitHubBlob(
     size,
   });
 }
+
+export async function getGitHubIssues(owner: string, repo: string): Promise<GitHubIssues> {
+  const html = await fetchGitHubPage(`${owner}/${repo}/issues`);
+
+  const payload = parsePayload(html, "react-app.embeddedData", ["payload"]);
+
+  let issuesQuery: IssueIndexPageQuery | undefined;
+  for (const query of payload?.preloadedQueries ?? []) {
+    if (query.queryName === "IssueIndexPageQuery") {
+      issuesQuery = query;
+    }
+  }
+  if (issuesQuery === undefined) {
+    throw new Error("Could not find IssueIndexPageQuery in embedded JSON");
+  }
+
+  const issues: Issue[] = issuesQuery.result.data.repository.search.edges.map((edge) => {
+    const { number, title, state, createdAt } = edge.node;
+    return { number, title, state, createdAt };
+  });
+
+  return {
+    repo: {
+      owner,
+      name: repo,
+    },
+    issues,
+  };
+}
diff --git a/cgithub/static/style.css b/cgithub/static/style.css
index cccb1ba..7618937 100644
--- a/cgithub/static/style.css
+++ b/cgithub/static/style.css
@@ -39,6 +39,41 @@ a {
   color: var(--link-color);
 }
 
+table {
+  width: 100%;
+  border-collapse: collapse;
+}
+
+tr {
+  border: 1px solid var(--border-color);
+}
+
+td {
+  padding: .25rem;
+  vertical-align: top;
+  &:first-child {
+    padding-left: .5rem;
+  }
+  &:last-child {
+    padding-right: .5rem;
+  }
+}
+
+nav {
+  display: flex;
+  > a {
+    display: block;
+    padding-right: .5rem;
+    + a {
+      padding-left: .5rem;
+      border-left: 1px solid var(--border-color);
+    }
+    &:last-child {
+      padding-right: 0;
+    }
+  }
+}
+
 section {
   margin: 1rem 0;
 }
@@ -134,3 +169,8 @@ pre.box {
     display: none;
   }
 }
+
+.timestamp {
+  white-space: nowrap;
+  font-size: .8rem;
+}
diff --git a/cgithub/views/issues.eta b/cgithub/views/issues.eta
new file mode 100644
index 0000000..6e4c9ba
--- /dev/null
+++ b/cgithub/views/issues.eta
@@ -0,0 +1,20 @@
+<% layout("./layout") %>
+
+<table>
+  <% for (const issue of it.issues) { %>
+    <tr>
+      <td>#<%= issue.number %></td>
+      <td>
+        <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/issues/<%= issue.number %>">
+          <%= issue.title %>
+        </a>
+      </td>
+      <td>
+        <% if (issue.commentCount > 0) { %>
+          <img src="/static/icons/link.svg" alt="comments" class="icon"> <%= issue.commentCount %>
+        <% } %>
+      </td>
+      <td class="timestamp"><time title="<%= issue.createdAt %>"><%= issue.createdAt.split("T")[0] %></time></td>
+    </tr>
+  <% } %>
+</table>
diff --git a/cgithub/views/layout.eta b/cgithub/views/layout.eta
index f053b84..f4535df 100644
--- a/cgithub/views/layout.eta
+++ b/cgithub/views/layout.eta
@@ -10,6 +10,10 @@
     <div class="repo"><%= it.repo.owner %>/<%= it.repo.name %></div>
     <pre><code>git clone https://github.com/<%= it.repo.owner %>/<%= it.repo.name %>.git
 git clone git@github.com:<%= it.repo.owner %>/<%= it.repo.name %>.git</code></pre>
+    <nav>
+      <a href="/<%= it.repo.owner %>/<%= it.repo.name %>">Code</a>
+      <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/issues">Issues</a>
+    </nav>
   </div>
 
   <%~ it.body %>
diff --git a/cgithub/views/repo.eta b/cgithub/views/repo.eta
index 9122775..fbca299 100644
--- a/cgithub/views/repo.eta
+++ b/cgithub/views/repo.eta
@@ -18,7 +18,7 @@
       <span class="count"><img src="/static/icons/git-fork.svg" alt="forks" class="icon"> <%= it.info.forks %></span>
     <% } %>
     <% if (it.info.numReleases) { %>
-      <div><a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases" rel="noreferrer"><%= it.info.numReleases %> releases</a></div>
+      <div><a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases"><%= it.info.numReleases %> releases</a></div>
     <% } %>
   </div>
 <% } %>
-- 
cgit v1.3.1


From 89a771538a6a673de7cfe658b1db028acdfa0e44 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Wed, 28 Jan 2026 13:05:02 -0500
Subject: Render Markdown blobs

---
 cgithub/src/scraper.test.ts | 38 ++++++++++++++++++++++++++++++++++----
 cgithub/src/scraper.ts      | 39 +++++++++++++++++++--------------------
 cgithub/static/style.css    | 10 +++++-----
 cgithub/views/blob.eta      | 10 ++++++----
 4 files changed, 64 insertions(+), 33 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index e7e65ba..f3a472d 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -59,7 +59,7 @@ describe("GitHub scraper", () => {
       ]);
     });
 
-    it("should fetch blob", async () => {
+    it("should fetch text blob", async () => {
       const data = await getGitHubBlob("actions", "deploy-pages", "main", "LICENSE");
 
       assert.strictEqual(data.repo.owner, "actions");
@@ -67,10 +67,40 @@ describe("GitHub scraper", () => {
       assert.strictEqual(data.branch, "main");
       assert.strictEqual(data.path, "LICENSE");
 
-      assert.strictEqual(data.size, 1068);
+      assert.strictEqual(data.size, "1.04 KB / 21 lines / 17 loc");
       assert.strictEqual(data.language, "Text");
-      assert.match(data.rawContent, /^MIT License\n\n/);
-      assert.match(data.content, /^MIT License\n\n/);
+      assert.strictEqual(data.rawLines?.[0], "MIT License");
+      assert.strictEqual(data.htmlContent, null);
+    });
+
+    it("should fetch Markdown blob", async () => {
+      const data = await getGitHubBlob("actions", "deploy-pages", "main", "README.md");
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+      assert.strictEqual(data.branch, "main");
+      assert.strictEqual(data.path, "README.md");
+
+      assert.strictEqual(data.size, "8.95 KB / 133 lines / 94 loc");
+      assert.strictEqual(data.language, "Markdown");
+      assert.strictEqual(data.rawLines, null);
+      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>/;
+      assert.match(data.htmlContent ?? "", firstLine);
+    });
+
+    it("should fetch code blob", async () => {
+      const data = await getGitHubBlob("actions", "deploy-pages", "main", ".gitattributes");
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+      assert.strictEqual(data.branch, "main");
+      assert.strictEqual(data.path, ".gitattributes");
+
+      assert.strictEqual(data.size, "39 Bytes / 1 lines / 1 loc");
+      assert.strictEqual(data.language, "Git Attributes");
+      assert.deepStrictEqual(data.rawLines, ["dist/** -diff linguist-generated=true "]);
+      assert.strictEqual(data.htmlContent, null);
     });
 
     it("should fetch issues", async () => {
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index bbf7511..c2983c9 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -16,7 +16,6 @@ interface TreeItem {
 
 interface OverviewFile {
   displayName: string;
-  preferredFileType: string;
   richText?: string;
   loaded: boolean;
 }
@@ -64,12 +63,17 @@ interface RawPayload {
     items: TreeItem[];
   };
   blob?: {
+    headerInfo: {
+      blobSize: string;
+      lineInfo: {
+        truncatedLoc: string;
+        truncatedSloc: string;
+      };
+    };
     language?: string;
-    rawLines?: string[];
-    highlightedLines?: {
-      text?: string;
-    }[];
-    size: number;
+    rawLines: string[] | null;
+    colorizedLines: string[] | null;
+    richText: string | null;
   };
   overview?: {
     overviewFiles?: OverviewFile[];
@@ -110,9 +114,9 @@ export interface GitHubRepo extends GitHubTree {
 
 export interface GitHubBlob extends GitHubNav {
   language: string | null;
-  size: number | null;
-  content: string;
-  rawContent: string;
+  size: string;
+  rawLines: string[] | null;
+  htmlContent: string | null;
 }
 
 interface Issue {
@@ -265,20 +269,15 @@ export async function getGitHubBlob(
     throw new Error("Could not find blob data in embedded JSON");
   }
 
-  const blob = payload.blob;
-  const rawLines = blob.rawLines || [];
-  const rawContent = rawLines.join("\n");
-  const highlightedLines = blob.highlightedLines || [];
-  const content = highlightedLines.map((line) => line.text || "").join("\n");
-
-  // Calculate size from raw content if not provided
-  const size = blob.size ?? (rawContent.length > 0 ? Buffer.byteLength(rawContent, "utf8") : null);
+  const { headerInfo, language, rawLines, richText } = payload.blob;
+  // TODO fine tune
+  const size = `${headerInfo.blobSize} / ${headerInfo.lineInfo.truncatedLoc} lines / ${headerInfo.lineInfo.truncatedSloc} loc`;
 
   return extractGitHub(payload, {
-    content: content || rawContent,
-    rawContent,
-    language: blob.language || null,
+    language: language || null,
     size,
+    rawLines,
+    htmlContent: richText,
   });
 }
 
diff --git a/cgithub/static/style.css b/cgithub/static/style.css
index 7618937..1efbb12 100644
--- a/cgithub/static/style.css
+++ b/cgithub/static/style.css
@@ -163,14 +163,14 @@ pre.box {
   code {
     background: var(--shade-background);
   }
-
-  /* TODO Strip permalinks? */
-  svg.octicon-link {
-    display: none;
-  }
 }
 
 .timestamp {
   white-space: nowrap;
   font-size: .8rem;
 }
+
+/* TODO Strip permalinks? */
+svg.octicon-link {
+  display: none;
+}
diff --git a/cgithub/views/blob.eta b/cgithub/views/blob.eta
index 277b46d..346afab 100644
--- a/cgithub/views/blob.eta
+++ b/cgithub/views/blob.eta
@@ -6,12 +6,14 @@
   <% if (it.language) { %>
     <div>language: <%= it.language %></div>
   <% } %>
-  <% if (it.size !== undefined) { %>
-    <div>size: <%= it.size %> bytes</div>
-  <% } %>
+  <div><%= it.size %></div>
   <div>
     <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/raw/<%= it.branch %>/<%= it.path %>">Raw</a>
   </div>
 </div>
 
-<pre class="box"><code><%= it.rawContent %></code></pre>
+<% if (it.htmlContent) { %>
+  <div class="box"><%~ it.htmlContent %></div>
+<% } else { %>
+  <pre class="box"><code><%= it.rawLines.join("\n") %></code></pre>
+<% } %>
-- 
cgit v1.3.1


From 71442a856d2180c68a51a1db480514d611137bd0 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Wed, 28 Jan 2026 14:40:22 -0500
Subject: Syntax highlighting

Sometimes GitHub returns colorizedLines and sometimes it returns a
terser stylingDirectives, which needs to be applied to rawLines.
---
 cgithub/src/scraper.test.ts | 13 ++++++++++---
 cgithub/src/scraper.ts      | 43 ++++++++++++++++++++++++++++++++++++++-----
 cgithub/static/style.css    | 22 ++++++++++++++++++++++
 cgithub/views/blob.eta      |  4 +++-
 4 files changed, 73 insertions(+), 9 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index f3a472d..9326747 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -69,7 +69,10 @@ describe("GitHub scraper", () => {
 
       assert.strictEqual(data.size, "1.04 KB / 21 lines / 17 loc");
       assert.strictEqual(data.language, "Text");
-      assert.strictEqual(data.rawLines?.[0], "MIT License");
+      assert.strictEqual(data.textLines?.length, 21);
+      assert.strictEqual(data.textLines[0], "MIT License");
+      assert.strictEqual(data.htmlLines?.length, 21);
+      assert.strictEqual(data.htmlLines[0], "MIT License");
       assert.strictEqual(data.htmlContent, null);
     });
 
@@ -83,7 +86,8 @@ describe("GitHub scraper", () => {
 
       assert.strictEqual(data.size, "8.95 KB / 133 lines / 94 loc");
       assert.strictEqual(data.language, "Markdown");
-      assert.strictEqual(data.rawLines, null);
+      assert.strictEqual(data.textLines, null);
+      assert.strictEqual(data.htmlLines, null);
       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>/;
       assert.match(data.htmlContent ?? "", firstLine);
@@ -99,7 +103,10 @@ describe("GitHub scraper", () => {
 
       assert.strictEqual(data.size, "39 Bytes / 1 lines / 1 loc");
       assert.strictEqual(data.language, "Git Attributes");
-      assert.deepStrictEqual(data.rawLines, ["dist/** -diff linguist-generated=true "]);
+      assert.deepStrictEqual(data.textLines, ["dist/** -diff linguist-generated=true "]);
+      assert.deepStrictEqual(data.htmlLines, [
+        '\u003cspan class="pl-e"\u003edist\u003c/span\u003e/\u003cspan class="pl-k"\u003e**\u003c/span\u003e \u003cspan class="pl-k"\u003e-\u003c/span\u003e\u003cspan class="pl-v"\u003ediff\u003c/span\u003e \u003cspan class="pl-v"\u003elinguist-generated\u003c/span\u003e=\u003cspan class="pl-c1"\u003etrue\u003c/span\u003e ',
+      ]);
       assert.strictEqual(data.htmlContent, null);
     });
 
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index c2983c9..05afc33 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -52,6 +52,8 @@ interface IssueIndexPageQuery {
   };
 }
 
+type StylingDirective = [number, number, string];
+
 interface RawPayload {
   repo: {
     ownerLogin: string;
@@ -74,6 +76,7 @@ interface RawPayload {
     rawLines: string[] | null;
     colorizedLines: string[] | null;
     richText: string | null;
+    stylingDirectives: StylingDirective[][] | null;
   };
   overview?: {
     overviewFiles?: OverviewFile[];
@@ -82,6 +85,27 @@ interface RawPayload {
   preloadedQueries?: IssueIndexPageQuery[];
 }
 
+function escapeHtml(text: string): string {
+  return text
+    .replace(/&/g, "&amp;")
+    .replace(/</g, "&lt;")
+    .replace(/>/g, "&gt;")
+    .replace(/"/g, "&quot;")
+    .replace(/'/g, "&#039;");
+}
+
+function applyStyling(line: string, directives: StylingDirective[]): string {
+  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;
+  }
+  result += escapeHtml(line.substring(lastPos));
+  return result;
+}
+
 interface GitHubCommon {
   repo: {
     owner: string;
@@ -115,7 +139,8 @@ export interface GitHubRepo extends GitHubTree {
 export interface GitHubBlob extends GitHubNav {
   language: string | null;
   size: string;
-  rawLines: string[] | null;
+  textLines: string[] | null;
+  htmlLines: string[] | null;
   htmlContent: string | null;
 }
 
@@ -269,15 +294,23 @@ export async function getGitHubBlob(
     throw new Error("Could not find blob data in embedded JSON");
   }
 
-  const { headerInfo, language, rawLines, richText } = payload.blob;
+  const blob = payload.blob;
+  const { headerInfo, stylingDirectives } = blob;
+
+  let htmlLines = blob.colorizedLines;
+  if (!htmlLines && stylingDirectives && blob.rawLines) {
+    htmlLines = blob.rawLines.map((line, i) => applyStyling(line, stylingDirectives[i]));
+  }
+
   // TODO fine tune
   const size = `${headerInfo.blobSize} / ${headerInfo.lineInfo.truncatedLoc} lines / ${headerInfo.lineInfo.truncatedSloc} loc`;
 
   return extractGitHub(payload, {
-    language: language || null,
+    language: blob.language || null,
     size,
-    rawLines,
-    htmlContent: richText,
+    textLines: blob.rawLines,
+    htmlLines,
+    htmlContent: blob.richText,
   });
 }
 
diff --git a/cgithub/static/style.css b/cgithub/static/style.css
index 1efbb12..2020dc2 100644
--- a/cgithub/static/style.css
+++ b/cgithub/static/style.css
@@ -25,6 +25,14 @@
     oklch(50% 0.15 250),
     oklch(60% 0.15 250)
   );
+
+  --pl-c: light-dark(#6e7781, #8b949e);
+  --pl-c1: light-dark(#0550ae, #79c0ff);
+  --pl-e: light-dark(#6639ba, #d2a8ff);
+  --pl-en: light-dark(#0550ae, #d2a8ff);
+  --pl-k: light-dark(#cf222e, #ff7b72);
+  --pl-s: light-dark(#0a3069, #a5d6ff);
+  --pl-v: light-dark(#953800, #ffa657);
 }
 
 body {
@@ -174,3 +182,17 @@ pre.box {
 svg.octicon-link {
   display: none;
 }
+
+.pl-c { color: var(--pl-c); }
+.pl-c1 { color: var(--pl-c1); }
+.pl-e { color: var(--pl-e); }
+.pl-en { color: var(--pl-en); }
+.pl-k { color: var(--pl-k); }
+.pl-s { color: var(--pl-s); }
+.pl-v { color: var(--pl-v); }
+.pl-ent { color: var(--pl-k); }
+.pl-mh { font-weight: bold; color: var(--pl-en); }
+.pl-bu { color: var(--pl-k); }
+.pl-ii { color: var(--text-color); background-color: var(--pl-k); }
+.pl-pds { color: var(--pl-s); }
+.pl-smi { color: var(--text-color); }
diff --git a/cgithub/views/blob.eta b/cgithub/views/blob.eta
index 346afab..4972209 100644
--- a/cgithub/views/blob.eta
+++ b/cgithub/views/blob.eta
@@ -14,6 +14,8 @@
 
 <% if (it.htmlContent) { %>
   <div class="box"><%~ it.htmlContent %></div>
+<% } else if (it.htmlLines) { %>
+  <pre class="box"><code><%~ it.htmlLines.join("\n") %></code></pre>
 <% } else { %>
-  <pre class="box"><code><%= it.rawLines.join("\n") %></code></pre>
+  <pre class="box"><code><%= it.textLines.join("\n") %></code></pre>
 <% } %>
-- 
cgit v1.3.1


From 631bc70d94cd12d0432cca1a220a3642eea6976a Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Wed, 28 Jan 2026 15:10:00 -0500
Subject: Image support

---
 cgithub/src/scraper.test.ts | 24 ++++++++++++++++++++++++
 cgithub/src/scraper.ts      | 18 +++++++++++++-----
 cgithub/views/blob.eta      |  4 +++-
 3 files changed, 40 insertions(+), 6 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 9326747..ae38c44 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -69,6 +69,8 @@ describe("GitHub scraper", () => {
 
       assert.strictEqual(data.size, "1.04 KB / 21 lines / 17 loc");
       assert.strictEqual(data.language, "Text");
+      assert.strictEqual(data.image, false);
+
       assert.strictEqual(data.textLines?.length, 21);
       assert.strictEqual(data.textLines[0], "MIT License");
       assert.strictEqual(data.htmlLines?.length, 21);
@@ -86,6 +88,8 @@ describe("GitHub scraper", () => {
 
       assert.strictEqual(data.size, "8.95 KB / 133 lines / 94 loc");
       assert.strictEqual(data.language, "Markdown");
+      assert.strictEqual(data.image, false);
+
       assert.strictEqual(data.textLines, null);
       assert.strictEqual(data.htmlLines, null);
       const firstLine =
@@ -103,6 +107,8 @@ describe("GitHub scraper", () => {
 
       assert.strictEqual(data.size, "39 Bytes / 1 lines / 1 loc");
       assert.strictEqual(data.language, "Git Attributes");
+      assert.strictEqual(data.image, false);
+
       assert.deepStrictEqual(data.textLines, ["dist/** -diff linguist-generated=true "]);
       assert.deepStrictEqual(data.htmlLines, [
         '\u003cspan class="pl-e"\u003edist\u003c/span\u003e/\u003cspan class="pl-k"\u003e**\u003c/span\u003e \u003cspan class="pl-k"\u003e-\u003c/span\u003e\u003cspan class="pl-v"\u003ediff\u003c/span\u003e \u003cspan class="pl-v"\u003elinguist-generated\u003c/span\u003e=\u003cspan class="pl-c1"\u003etrue\u003c/span\u003e ',
@@ -124,6 +130,24 @@ describe("GitHub scraper", () => {
     });
   });
 
+  describe("rich file types", () => {
+    it("should support PNG", async () => {
+      const data = await getGitHubBlob("github", "docs", "main", "assets/images/site/logo.png");
+
+      assert.strictEqual(data.repo.owner, "github");
+      assert.strictEqual(data.repo.name, "docs");
+      assert.strictEqual(data.branch, "main");
+      assert.strictEqual(data.path, "assets/images/site/logo.png");
+
+      assert.strictEqual(data.size, "1.26 KB");
+      assert.strictEqual(data.language, null);
+      assert.strictEqual(data.image, true);
+      assert.strictEqual(data.textLines, null);
+      assert.strictEqual(data.htmlLines, null);
+      assert.strictEqual(data.htmlContent, null);
+    });
+  });
+
   describe("error handling", () => {
     it("should throw for non-existent repository", async () => {
       await assert.rejects(getGitHubRepo("nosuchowner", "nosuchrepo"), (err) => {
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 05afc33..e58402a 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -68,11 +68,12 @@ interface RawPayload {
     headerInfo: {
       blobSize: string;
       lineInfo: {
-        truncatedLoc: string;
-        truncatedSloc: string;
+        truncatedLoc: string | null;
+        truncatedSloc: string | null;
       };
     };
-    language?: string;
+    language: string | null;
+    image: boolean;
     rawLines: string[] | null;
     colorizedLines: string[] | null;
     richText: string | null;
@@ -139,6 +140,7 @@ export interface GitHubRepo extends GitHubTree {
 export interface GitHubBlob extends GitHubNav {
   language: string | null;
   size: string;
+  image: boolean;
   textLines: string[] | null;
   htmlLines: string[] | null;
   htmlContent: string | null;
@@ -302,12 +304,18 @@ export async function getGitHubBlob(
     htmlLines = blob.rawLines.map((line, i) => applyStyling(line, stylingDirectives[i]));
   }
 
-  // TODO fine tune
-  const size = `${headerInfo.blobSize} / ${headerInfo.lineInfo.truncatedLoc} lines / ${headerInfo.lineInfo.truncatedSloc} loc`;
+  let size = headerInfo.blobSize;
+  if (headerInfo.lineInfo.truncatedLoc) {
+    size += ` / ${headerInfo.lineInfo.truncatedLoc} lines`;
+  }
+  if (headerInfo.lineInfo.truncatedSloc) {
+    size += ` / ${headerInfo.lineInfo.truncatedSloc} loc`;
+  }
 
   return extractGitHub(payload, {
     language: blob.language || null,
     size,
+    image: blob.image,
     textLines: blob.rawLines,
     htmlLines,
     htmlContent: blob.richText,
diff --git a/cgithub/views/blob.eta b/cgithub/views/blob.eta
index 4972209..ea07059 100644
--- a/cgithub/views/blob.eta
+++ b/cgithub/views/blob.eta
@@ -12,7 +12,9 @@
   </div>
 </div>
 
-<% if (it.htmlContent) { %>
+<% if (it.image) { %>
+  <div class="box"><img src="/<%= it.repo.owner %>/<%= it.repo.name %>/raw/<%= it.branch %>/<%= it.path %>"></div>
+<% } else if (it.htmlContent) { %>
   <div class="box"><%~ it.htmlContent %></div>
 <% } else if (it.htmlLines) { %>
   <pre class="box"><code><%~ it.htmlLines.join("\n") %></code></pre>
-- 
cgit v1.3.1


From 94fcda59a064b5e8952d1bccb6b9a153c39acda1 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Wed, 28 Jan 2026 22:41:00 -0500
Subject: Commits list

---
 cgithub/src/index.ts        |  6 +++++
 cgithub/src/scraper.test.ts | 31 ++++++++++++++++++++++++++
 cgithub/src/scraper.ts      | 53 +++++++++++++++++++++++++++++++++++++++++++++
 cgithub/static/style.css    | 20 +++++++++++++++++
 cgithub/views/commits.eta   | 25 +++++++++++++++++++++
 cgithub/views/layout.eta    |  3 +++
 6 files changed, 138 insertions(+)
 create mode 100644 cgithub/views/commits.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/index.ts b/cgithub/src/index.ts
index d4229fd..4486dff 100644
--- a/cgithub/src/index.ts
+++ b/cgithub/src/index.ts
@@ -7,6 +7,7 @@ import path from "node:path";
 import url from "node:url";
 import {
   getGitHubBlob,
+  getGitHubCommits,
   getGitHubIssues,
   getGitHubRepo,
   getGitHubTree,
@@ -67,6 +68,11 @@ app.get("/:owner/:repo/issues", async (c) => {
   return tryRender(c, "issues.eta", getGitHubIssues(owner, repo));
 });
 
+app.get("/:owner/:repo/commits/:branch/:path{.*}?", async (c) => {
+  const { owner, repo, branch, path = "" } = c.req.param();
+  return tryRender(c, "commits.eta", getGitHubCommits(owner, repo, branch, path));
+});
+
 app.all("*", async (c) => {
   const path = c.req.path;
   c.header("Referrer-Policy", "no-referrer");
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index ae38c44..7cafd9c 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -2,6 +2,7 @@ import assert from "node:assert";
 import { describe, it } from "node:test";
 import {
   getGitHubBlob,
+  getGitHubCommits,
   getGitHubIssues,
   getGitHubRepo,
   getGitHubTree,
@@ -128,6 +129,36 @@ describe("GitHub scraper", () => {
       assert.strictEqual(issue402.state, "OPEN");
       assert.strictEqual(issue402.createdAt, "2025-07-03T19:07:08Z");
     });
+
+    it("should fetch commits", async () => {
+      const data = await getGitHubCommits("actions", "deploy-pages", "main", "");
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+      assert.strictEqual(data.branch, "main");
+      assert.strictEqual(data.path, "");
+
+      assert.ok(data.commitGroups.length > 0);
+      const firstGroup = data.commitGroups[0];
+      assert.ok(firstGroup.title.length > 0);
+      assert.ok(firstGroup.commits.length > 0);
+
+      const firstCommit = firstGroup.commits[0];
+      assert.ok(firstCommit.oid.length === 40);
+      assert.ok(firstCommit.shortMessage.length > 0);
+      assert.ok(firstCommit.authors.length > 0);
+    });
+
+    it("should fetch commits with path", async () => {
+      const data = await getGitHubCommits("actions", "deploy-pages", "main", "src");
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+      assert.strictEqual(data.branch, "main");
+      assert.strictEqual(data.path, "src");
+
+      assert.ok(data.commitGroups.length > 0);
+    });
   });
 
   describe("rich file types", () => {
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index e58402a..71d0e3e 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -54,6 +54,26 @@ interface IssueIndexPageQuery {
 
 type StylingDirective = [number, number, string];
 
+interface Commit {
+  oid: string;
+  url: string;
+  authoredDate: string;
+  committedDate: string;
+  shortMessage: string;
+  bodyMessageHtml: string;
+  authors: {
+    login: string;
+    displayName: string;
+    avatarUrl: string;
+    path: string;
+  }[];
+}
+
+export interface CommitGroup {
+  title: string;
+  commits: Commit[];
+}
+
 interface RawPayload {
   repo: {
     ownerLogin: string;
@@ -82,6 +102,7 @@ interface RawPayload {
   overview?: {
     overviewFiles?: OverviewFile[];
   };
+  commitGroups?: CommitGroup[];
   // Actually there may be other query types.
   preloadedQueries?: IssueIndexPageQuery[];
 }
@@ -159,6 +180,10 @@ export interface GitHubIssues extends GitHubCommon {
   issues: Issue[];
 }
 
+export interface GitHubCommits extends GitHubNav {
+  commitGroups: CommitGroup[];
+}
+
 async function fetchGitHubPage(path: string): Promise<string> {
   // GitHub throttles/blocks requests without realistic browser headers.
   // These headers make the request appear as a standard browser visit.
@@ -350,3 +375,31 @@ export async function getGitHubIssues(owner: string, repo: string): Promise<GitH
     issues,
   };
 }
+
+export async function getGitHubCommits(
+  owner: string,
+  repo: string,
+  branch: string,
+  path: string,
+): Promise<GitHubCommits> {
+  const urlPath = path
+    ? `${owner}/${repo}/commits/${branch}/${path}`
+    : `${owner}/${repo}/commits/${branch}`;
+  const html = await fetchGitHubPage(urlPath);
+
+  const payload = parsePayload(html, "react-app.embeddedData", ["payload"]);
+
+  if (payload?.commitGroups === undefined) {
+    throw new Error("Could not find commit history in embedded JSON");
+  }
+
+  return {
+    repo: {
+      owner,
+      name: repo,
+    },
+    branch,
+    path,
+    commitGroups: payload.commitGroups,
+  };
+}
diff --git a/cgithub/static/style.css b/cgithub/static/style.css
index 2020dc2..5673aa7 100644
--- a/cgithub/static/style.css
+++ b/cgithub/static/style.css
@@ -178,6 +178,26 @@ pre.box {
   font-size: .8rem;
 }
 
+.commit {
+  display: flex;
+  justify-content: space-between;
+  padding: 0.5rem;
+  border: 1px solid var(--border-color);
+  + .commit {
+    border-top: none;
+  }
+}
+
+.commit-author {
+  font-size: 0.8rem;
+  margin-top: 0.2rem;
+}
+
+.commit-sha {
+  align-self: center;
+  font-size: 0.8rem;
+}
+
 /* TODO Strip permalinks? */
 svg.octicon-link {
   display: none;
diff --git a/cgithub/views/commits.eta b/cgithub/views/commits.eta
new file mode 100644
index 0000000..2ddcba3
--- /dev/null
+++ b/cgithub/views/commits.eta
@@ -0,0 +1,25 @@
+<% layout("./layout.eta") %>
+
+<%~ include("./_nav.eta", {...it, isTree: false}) %>
+
+<% for (const group of it.commitGroups) { %>
+  <h3><%= group.title %></h3>
+  <% for (const commit of group.commits) { %>
+    <div class="commit">
+      <div>
+        <div>
+          <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/commit/<%= commit.oid %>"><%= commit.shortMessage %></a>
+        </div>
+        <div class="commit-author">
+          by
+          <% for (const author of commit.authors) { %>
+            <a href="/<%= author.login %>"><%= author.login %></a>
+          <% } %>
+        </div>
+      </div>
+      <div class="commit-sha">
+        <code><%= commit.oid.substring(0, 7) %></code>
+      </div>
+    </div>
+  <% } %>
+<% } %>
diff --git a/cgithub/views/layout.eta b/cgithub/views/layout.eta
index f4535df..a404437 100644
--- a/cgithub/views/layout.eta
+++ b/cgithub/views/layout.eta
@@ -13,6 +13,9 @@ git clone git@github.com:<%= it.repo.owner %>/<%= it.repo.name %>.git</code></pr
     <nav>
       <a href="/<%= it.repo.owner %>/<%= it.repo.name %>">Code</a>
       <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/issues">Issues</a>
+      <% if (it.branch) { %>
+        <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/commits/<%= it.branch %><%= it.path && it.path !== '/' ? '/' + it.path : '' %>">Commits</a>
+      <% } %>
     </nav>
   </div>
 
-- 
cgit v1.3.1


From 34eea157db19c276fa4a5578707d0f8af2eeffd3 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Wed, 11 Feb 2026 01:56:37 -0500
Subject: Add GitHub releases support

Scrapes releases page HTML to display version tags, publish dates, and changelog content.

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
---
 cgithub/src/index.ts        |  6 +++++
 cgithub/src/scraper.test.ts | 14 ++++++++++
 cgithub/src/scraper.ts      | 63 +++++++++++++++++++++++++++++++++++++++++++++
 cgithub/static/style.css    | 42 ++++++++++++++++++++++++++++++
 cgithub/views/releases.eta  | 25 ++++++++++++++++++
 5 files changed, 150 insertions(+)
 create mode 100644 cgithub/views/releases.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/index.ts b/cgithub/src/index.ts
index 4486dff..2f8efef 100644
--- a/cgithub/src/index.ts
+++ b/cgithub/src/index.ts
@@ -9,6 +9,7 @@ import {
   getGitHubBlob,
   getGitHubCommits,
   getGitHubIssues,
+  getGitHubReleases,
   getGitHubRepo,
   getGitHubTree,
   GitHubHTTPError,
@@ -73,6 +74,11 @@ app.get("/:owner/:repo/commits/:branch/:path{.*}?", async (c) => {
   return tryRender(c, "commits.eta", getGitHubCommits(owner, repo, branch, path));
 });
 
+app.get("/:owner/:repo/releases", async (c) => {
+  const { owner, repo } = c.req.param();
+  return tryRender(c, "releases.eta", getGitHubReleases(owner, repo));
+});
+
 app.all("*", async (c) => {
   const path = c.req.path;
   c.header("Referrer-Policy", "no-referrer");
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 7cafd9c..8c11317 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -4,6 +4,7 @@ import {
   getGitHubBlob,
   getGitHubCommits,
   getGitHubIssues,
+  getGitHubReleases,
   getGitHubRepo,
   getGitHubTree,
   GitHubHTTPError,
@@ -159,6 +160,19 @@ describe("GitHub scraper", () => {
 
       assert.ok(data.commitGroups.length > 0);
     });
+
+    it("should fetch releases", async () => {
+      const data = await getGitHubReleases("actions", "deploy-pages");
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+
+      assert.ok(data.releases.length > 0);
+      const firstRelease = data.releases[0];
+      assert.ok(firstRelease.tagName.length > 0);
+      assert.ok(firstRelease.title.match(/^v\d/));
+      assert.ok(firstRelease.publishedAt.length > 0);
+    });
   });
 
   describe("rich file types", () => {
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 71d0e3e..e52ddac 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -184,6 +184,17 @@ export interface GitHubCommits extends GitHubNav {
   commitGroups: CommitGroup[];
 }
 
+interface Release {
+  tagName: string;
+  title: string;
+  publishedAt: string;
+  bodyHtml: string;
+}
+
+export interface GitHubReleases extends GitHubCommon {
+  releases: Release[];
+}
+
 async function fetchGitHubPage(path: string): Promise<string> {
   // GitHub throttles/blocks requests without realistic browser headers.
   // These headers make the request appear as a standard browser visit.
@@ -403,3 +414,55 @@ export async function getGitHubCommits(
     commitGroups: payload.commitGroups,
   };
 }
+
+export async function getGitHubReleases(owner: string, repo: string): Promise<GitHubReleases> {
+  const html = await fetchGitHubPage(`${owner}/${repo}/releases`);
+
+  // Releases page doesn't have embedded JSON, so we parse HTML directly
+  const releases: Release[] = [];
+
+  // Find all h2 tags with version numbers
+  const h2Regex = /<h2[^>]*id="([^"]*)"[^>]*>(v[^<]*)<\/h2>/g;
+
+  for (const h2Match of html.matchAll(h2Regex)) {
+    const id = h2Match[1];
+    const title = h2Match[2].trim();
+
+    // Find the section containing this h2
+    const sectionRegex = new RegExp(
+      `<section[^>]*>\\s*<h2[^>]*id="${id}"[^>]*>[^<]*<\\/h2>([\\s\\S]*?)<\\/section>`,
+      "s",
+    );
+    const sectionMatch = html.match(sectionRegex);
+    if (!sectionMatch) continue;
+
+    const sectionHtml = sectionMatch[1];
+
+    // Extract tag name from link
+    const tagMatch = sectionHtml.match(/<a[^>]*href="[^"]*\/tree\/([^"]+)"[^>]*>/);
+    const tagName = tagMatch?.[1] || title;
+
+    // Extract published date
+    const dateMatch = sectionHtml.match(/<relative-time[^>]*datetime="([^"]*)"[^>]*>/);
+    const publishedAt = dateMatch?.[1] || "";
+
+    // Extract body HTML (markdown content)
+    const bodyMatch = sectionHtml.match(/<div[^>]*class="[^"]*markdown-body[^"]*"[^>]*>([\s\S]*?)<\/div>/);
+    const bodyHtml = bodyMatch?.[1]?.trim() || "";
+
+    releases.push({
+      tagName,
+      title,
+      publishedAt,
+      bodyHtml,
+    });
+  }
+
+  return {
+    repo: {
+      owner,
+      name: repo,
+    },
+    releases,
+  };
+}
diff --git a/cgithub/static/style.css b/cgithub/static/style.css
index 8709dc2..fdf3285 100644
--- a/cgithub/static/style.css
+++ b/cgithub/static/style.css
@@ -232,6 +232,48 @@ pre.box {
   font-size: 0.8rem;
 }
 
+.release {
+  margin-bottom: 2rem;
+}
+
+.release-header {
+  display: flex;
+  justify-content: space-between;
+  align-items: baseline;
+  margin-bottom: 1rem;
+  padding-bottom: .5rem;
+  border-bottom: 1px solid var(--border-color);
+
+  h3 {
+    margin: 0;
+  }
+}
+
+.release-body {
+  overflow-wrap: break-word;
+
+  a {
+    text-decoration: underline;
+  }
+
+  h1, h2 {
+    border-bottom: 1px solid var(--border-color);
+  }
+
+  img {
+    max-width: 100%;
+  }
+
+  code, pre {
+    background: var(--shade-background);
+  }
+
+  pre {
+    overflow-x: auto;
+    scrollbar-color: var(--border-color) var(--shade-background);
+  }
+}
+
 /* TODO Strip permalinks? */
 svg.octicon-link {
   display: none;
diff --git a/cgithub/views/releases.eta b/cgithub/views/releases.eta
new file mode 100644
index 0000000..39850d8
--- /dev/null
+++ b/cgithub/views/releases.eta
@@ -0,0 +1,25 @@
+<% layout("./layout.eta") %>
+
+<h2>Releases</h2>
+
+<% for (const release of it.releases) { %>
+  <div class="box release">
+    <div class="release-header">
+      <h3>
+        <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases/tag/<%= release.tagName %>">
+          <%= release.title %>
+        </a>
+      </h3>
+      <% if (release.publishedAt) { %>
+        <time class="timestamp" title="<%= release.publishedAt %>">
+          <%= release.publishedAt.split("T")[0] %>
+        </time>
+      <% } %>
+    </div>
+    <% if (release.bodyHtml) { %>
+      <div class="release-body">
+        <%~ release.bodyHtml %>
+      </div>
+    <% } %>
+  </div>
+<% } %>
-- 
cgit v1.3.1


From 1435dea09882f110448b034d7cffa32d03e19a8f Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Sat, 21 Feb 2026 16:21:35 +1100
Subject: Parse release page HTML

Fixes reading complex release descriptions that had unbalanced tags.

Written with Gemini.
---
 cgithub/package.json        |   6 +-
 cgithub/pnpm-lock.yaml      | 240 +++++++++++++-------------------------------
 cgithub/src/scraper.test.ts |   1 +
 cgithub/src/scraper.ts      |  46 +++++----
 4 files changed, 102 insertions(+), 191 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/package.json b/cgithub/package.json
index 3d339a4..06e2876 100644
--- a/cgithub/package.json
+++ b/cgithub/package.json
@@ -11,8 +11,12 @@
   },
   "dependencies": {
     "@hono/node-server": "^1.13.7",
+    "css-select": "^6.0.0",
+    "domhandler": "^5.0.3",
+    "domutils": "^3.2.2",
     "eta": "^3.5.0",
-    "hono": "^4.6.14"
+    "hono": "^4.6.14",
+    "htmlparser2": "^10.1.0"
   },
   "devDependencies": {
     "@types/node": "^22.10.2",
diff --git a/cgithub/pnpm-lock.yaml b/cgithub/pnpm-lock.yaml
index 2cd9be3..4fd01e4 100644
--- a/cgithub/pnpm-lock.yaml
+++ b/cgithub/pnpm-lock.yaml
@@ -11,19 +11,28 @@ importers:
       '@hono/node-server':
         specifier: ^1.13.7
         version: 1.19.9(hono@4.11.5)
+      css-select:
+        specifier: ^6.0.0
+        version: 6.0.0
+      domhandler:
+        specifier: ^5.0.3
+        version: 5.0.3
+      domutils:
+        specifier: ^3.2.2
+        version: 3.2.2
       eta:
         specifier: ^3.5.0
         version: 3.5.0
       hono:
         specifier: ^4.6.14
         version: 4.11.5
+      htmlparser2:
+        specifier: ^10.1.0
+        version: 10.1.0
     devDependencies:
       '@types/node':
         specifier: ^22.10.2
         version: 22.19.7
-      concurrently:
-        specifier: ^9.1.2
-        version: 9.2.1
       esbuild:
         specifier: ^0.27.2
         version: 0.27.2
@@ -201,105 +210,61 @@ packages:
   '@types/node@22.19.7':
     resolution: {integrity: sha512-MciR4AKGHWl7xwxkBa6xUGxQJ4VBOmPTF7sL+iGzuahOFaO0jHCsuEfS80pan1ef4gWId1oWOweIhrDEYLuaOw==}
 
-  ansi-regex@5.0.1:
-    resolution: {integrity: sha512-quJQXlTSUGL2LH9SUXo8VwsY4soanhgo6LNSm84E1LBcE8s3O0wpdiRzyR9z/ZZJMlMWv37qOOb9pdJlMUEKFQ==}
-    engines: {node: '>=8'}
+  boolbase@1.0.0:
+    resolution: {integrity: sha512-JZOSA7Mo9sNGB8+UjSgzdLtokWAky1zbztM3WRLCbZ70/3cTANmQmOdR7y2g+J0e2WXywy1yS468tY+IruqEww==}
 
-  ansi-styles@4.3.0:
-    resolution: {integrity: sha512-zbB9rCJAT1rbjiVDb2hqKFHNYLxgtk8NURxZ3IZwD3F6NtxbXZQCnnSi1Lkx+IDohdPlFp222wVALIheZJQSEg==}
-    engines: {node: '>=8'}
+  css-select@6.0.0:
+    resolution: {integrity: sha512-rZZVSLle8v0+EY8QAkDWrKhpgt6SA5OtHsgBnsj6ZaLb5dmDVOWUDtQitd9ydxxvEjhewNudS6eTVU7uOyzvXw==}
 
-  chalk@4.1.2:
-    resolution: {integrity: sha512-oKnbhFyRIXpUuez8iBMmyEa4nbj4IOQyuhc/wy9kY7/WVPcwIO9VA668Pu8RkO7+0G76SLROeyw9CpQ061i4mA==}
-    engines: {node: '>=10'}
+  css-what@7.0.0:
+    resolution: {integrity: sha512-wD5oz5xibMOPHzy13CyGmogB3phdvcDaB5t0W/Nr5Z2O/agcB8YwOz6e2Lsp10pNDzBoDO9nVa3RGs/2BttpHQ==}
+    engines: {node: '>= 6'}
 
-  cliui@8.0.1:
-    resolution: {integrity: sha512-BSeNnyus75C4//NQ9gQt1/csTXyo/8Sb+afLAkzAptFuMsod9HFokGNudZpi/oQV73hnVK+sR+5PVRMd+Dr7YQ==}
-    engines: {node: '>=12'}
+  dom-serializer@2.0.0:
+    resolution: {integrity: sha512-wIkAryiqt/nV5EQKqQpo3SToSOV9J0DnbJqwK7Wv/Trc92zIAYZ4FlMu+JPFW1DfGFt81ZTCGgDEabffXeLyJg==}
 
-  color-convert@2.0.1:
-    resolution: {integrity: sha512-RRECPsj7iu/xb5oKYcsFHSppFNnsj/52OVTRKb4zP5onXwVF3zVmmToNcOfGC+CRDpfK/U584fMg38ZHCaElKQ==}
-    engines: {node: '>=7.0.0'}
+  domelementtype@2.3.0:
+    resolution: {integrity: sha512-OLETBj6w0OsagBwdXnPdN0cnMfF9opN69co+7ZrbfPGrdpPVNBUj02spi6B1N7wChLQiPn4CSH/zJvXw56gmHw==}
 
-  color-name@1.1.4:
-    resolution: {integrity: sha512-dOy+3AuW3a2wNbZHIuMZpTcgjGuLU/uBL/ubcZF9OXbDo8ff4O8yVp5Bf0efS8uEoYo5q4Fx7dY9OgQGXgAsQA==}
+  domhandler@5.0.3:
+    resolution: {integrity: sha512-cgwlv/1iFQiFnU96XXgROh8xTeetsnJiDsTc7TYCLFd9+/WNkIqPTxiM/8pSd8VIrhXGTf1Ny1q1hquVqDJB5w==}
+    engines: {node: '>= 4'}
 
-  concurrently@9.2.1:
-    resolution: {integrity: sha512-fsfrO0MxV64Znoy8/l1vVIjjHa29SZyyqPgQBwhiDcaW8wJc2W3XWVOGx4M3oJBnv/zdUZIIp1gDeS98GzP8Ng==}
-    engines: {node: '>=18'}
-    hasBin: true
+  domutils@3.2.2:
+    resolution: {integrity: sha512-6kZKyUajlDuqlHKVX1w7gyslj9MPIXzIFiz/rGu35uC1wMi+kMhQwGhl4lt9unC9Vb9INnY9Z3/ZA3+FhASLaw==}
+
+  entities@4.5.0:
+    resolution: {integrity: sha512-V0hjH4dGPh9Ao5p0MoRY6BVqtwCjhz6vI5LT8AJ55H+4g9/4vbHx1I54fS0XuclLhDHArPQCiMjDxjaL8fPxhw==}
+    engines: {node: '>=0.12'}
 
-  emoji-regex@8.0.0:
-    resolution: {integrity: sha512-MSjYzcWNOA0ewAHpz0MxpYFvwg6yjy1NG3xteoqz644VCo/RPgnr1/GGt+ic3iJTzQ8Eu3TdM14SawnVUmGE6A==}
+  entities@7.0.1:
+    resolution: {integrity: sha512-TWrgLOFUQTH994YUyl1yT4uyavY5nNB5muff+RtWaqNVCAK408b5ZnnbNAUEWLTCpum9w6arT70i1XdQ4UeOPA==}
+    engines: {node: '>=0.12'}
 
   esbuild@0.27.2:
     resolution: {integrity: sha512-HyNQImnsOC7X9PMNaCIeAm4ISCQXs5a5YasTXVliKv4uuBo1dKrG0A+uQS8M5eXjVMnLg3WgXaKvprHlFJQffw==}
     engines: {node: '>=18'}
     hasBin: true
 
-  escalade@3.2.0:
-    resolution: {integrity: sha512-WUj2qlxaQtO4g6Pq5c29GTcWGDyd8itL8zTlipgECz3JesAiiOKotd8JU6otB3PACgG6xkJUyVhboMS+bje/jA==}
-    engines: {node: '>=6'}
-
   eta@3.5.0:
     resolution: {integrity: sha512-e3x3FBvGzeCIHhF+zhK8FZA2vC5uFn6b4HJjegUbIWrDb4mJ7JjTGMJY9VGIbRVpmSwHopNiaJibhjIr+HfLug==}
     engines: {node: '>=6.0.0'}
 
-  get-caller-file@2.0.5:
-    resolution: {integrity: sha512-DyFP3BM/3YHTQOCUL/w0OZHR0lpKeGrxotcHWcqNEdnltqFwXVfhEBQ94eIo34AfQpo0rGki4cyIiftY06h2Fg==}
-    engines: {node: 6.* || 8.* || >= 10.*}
-
-  has-flag@4.0.0:
-    resolution: {integrity: sha512-EykJT/Q1KjTWctppgIAgfSO0tKVuZUjhgMr17kqTumMl6Afv3EISleU7qZUzoXDFTAHTDC4NOoG/ZxU3EvlMPQ==}
-    engines: {node: '>=8'}
-
   hono@4.11.5:
     resolution: {integrity: sha512-WemPi9/WfyMwZs+ZUXdiwcCh9Y+m7L+8vki9MzDw3jJ+W9Lc+12HGsd368Qc1vZi1xwW8BWMMsnK5efYKPdt4g==}
     engines: {node: '>=16.9.0'}
 
-  is-fullwidth-code-point@3.0.0:
-    resolution: {integrity: sha512-zymm5+u+sCsSWyD9qNaejV3DFvhCKclKdizYaJUuHA83RLjb7nSuGnddCHGv0hk+KY7BMAlsWeK4Ueg6EV6XQg==}
-    engines: {node: '>=8'}
+  htmlparser2@10.1.0:
+    resolution: {integrity: sha512-VTZkM9GWRAtEpveh7MSF6SjjrpNVNNVJfFup7xTY3UpFtm67foy9HDVXneLtFVt4pMz5kZtgNcvCniNFb1hlEQ==}
+
+  nth-check@2.1.1:
+    resolution: {integrity: sha512-lqjrjmaOoAnWfMmBPL+XNnynZh2+swxiX3WUE0s4yEHI6m+AwrK2UZOimIRl3X/4QctVqS8AiZjFqyOGrMXb/w==}
 
   prettier@3.8.1:
     resolution: {integrity: sha512-UOnG6LftzbdaHZcKoPFtOcCKztrQ57WkHDeRD9t/PTQtmT0NHSeWWepj6pS0z/N7+08BHFDQVUrfmfMRcZwbMg==}
     engines: {node: '>=14'}
     hasBin: true
 
-  require-directory@2.1.1:
-    resolution: {integrity: sha512-fGxEI7+wsG9xrvdjsrlmL22OMTTiHRwAMroiEeMgq8gzoLC/PQr7RsRDSTLUg/bZAZtF+TVIkHc6/4RIKrui+Q==}
-    engines: {node: '>=0.10.0'}
-
-  rxjs@7.8.2:
-    resolution: {integrity: sha512-dhKf903U/PQZY6boNNtAGdWbG85WAbjT/1xYoZIC7FAY0yWapOBQVsVrDl58W86//e1VpMNBtRV4MaXfdMySFA==}
-
-  shell-quote@1.8.3:
-    resolution: {integrity: sha512-ObmnIF4hXNg1BqhnHmgbDETF8dLPCggZWBjkQfhZpbszZnYur5DUljTcCHii5LC3J5E0yeO/1LIMyH+UvHQgyw==}
-    engines: {node: '>= 0.4'}
-
-  string-width@4.2.3:
-    resolution: {integrity: sha512-wKyQRQpjJ0sIp62ErSZdGsjMJWsap5oRNihHhu6G7JVO/9jIB6UyevL+tXuOqrng8j/cxKTWyWUwvSTriiZz/g==}
-    engines: {node: '>=8'}
-
-  strip-ansi@6.0.1:
-    resolution: {integrity: sha512-Y38VPSHcqkFrCpFnQ9vuSXmquuv5oXOKpGeT6aGrr3o3Gc9AlVa6JBfUSOCnbxGGZF+/0ooI7KrPuUSztUdU5A==}
-    engines: {node: '>=8'}
-
-  supports-color@7.2.0:
-    resolution: {integrity: sha512-qpCAvRl9stuOHveKsn7HncJRvv501qIacKzQlO/+Lwxc9+0q2wLyv4Dfvt80/DPn2pqOBsJdDiogXGR9+OvwRw==}
-    engines: {node: '>=8'}
-
-  supports-color@8.1.1:
-    resolution: {integrity: sha512-MpUEN2OodtUzxvKQl72cUF7RQ5EiHsGvSsVG0ia9c5RbWGL2CI4C7EpPS8UTBIplnlzZiNuV56w+FuNxy3ty2Q==}
-    engines: {node: '>=10'}
-
-  tree-kill@1.2.2:
-    resolution: {integrity: sha512-L0Orpi8qGpRG//Nd+H90vFB+3iHnue1zSSGmNOOCh1GLJ7rUKVwV2HvijphGQS2UmhUZewS9VgvxYIdgr+fG1A==}
-    hasBin: true
-
-  tslib@2.8.1:
-    resolution: {integrity: sha512-oJFu94HQb+KVduSUQL7wnpmqnfmLsOA/nAh6b6EH0wCEoK0/mPeXU6c3wKDV83MkOuHPRHtSXKKU99IBazS/2w==}
-
   typescript@5.9.3:
     resolution: {integrity: sha512-jl1vZzPDinLr9eUt3J/t7V6FgNEw9QjvBPdysz9KfQDD41fQrC2Y4vKQdiaUpFT4bXlb1RHhLpp8wtm6M5TgSw==}
     engines: {node: '>=14.17'}
@@ -308,22 +273,6 @@ packages:
   undici-types@6.21.0:
     resolution: {integrity: sha512-iwDZqg0QAGrg9Rav5H4n0M64c3mkR59cJ6wQp+7C4nI0gsmExaedaYLNO44eT4AtBBwjbTiGPMlt2Md0T9H9JQ==}
 
-  wrap-ansi@7.0.0:
-    resolution: {integrity: sha512-YVGIj2kamLSTxw6NsZjoBxfSwsn0ycdesmc4p+Q21c5zPuZ1pl+NfxVdxPtdHvmNVOQ6XSYG4AUtyt/Fi7D16Q==}
-    engines: {node: '>=10'}
-
-  y18n@5.0.8:
-    resolution: {integrity: sha512-0pfFzegeDWJHJIAmTLRP2DwHjdF5s7jo9tuztdQxAhINCdvS+3nGINqPd00AphqJR/0LhANUS6/+7SCb98YOfA==}
-    engines: {node: '>=10'}
-
-  yargs-parser@21.1.1:
-    resolution: {integrity: sha512-tVpsJW7DdjecAiFpbIB1e3qxIQsE6NoPc5/eTdrbbIC4h0LVsWhnoa3g+m2HclBIujHzsxZ4VJVA+GUuc2/LBw==}
-    engines: {node: '>=12'}
-
-  yargs@17.7.2:
-    resolution: {integrity: sha512-7dSzzRQ++CKnNI/krKnYRV7JKKPUXMEh61soaHKg9mrWEhzFWhFnxPxGl+69cD1Ou63C13NUPCnmIcrvqCuM6w==}
-    engines: {node: '>=12'}
-
 snapshots:
 
   '@esbuild/aix-ppc64@0.27.2':
@@ -412,39 +361,39 @@ snapshots:
     dependencies:
       undici-types: 6.21.0
 
-  ansi-regex@5.0.1: {}
+  boolbase@1.0.0: {}
 
-  ansi-styles@4.3.0:
+  css-select@6.0.0:
     dependencies:
-      color-convert: 2.0.1
+      boolbase: 1.0.0
+      css-what: 7.0.0
+      domhandler: 5.0.3
+      domutils: 3.2.2
+      nth-check: 2.1.1
 
-  chalk@4.1.2:
-    dependencies:
-      ansi-styles: 4.3.0
-      supports-color: 7.2.0
+  css-what@7.0.0: {}
 
-  cliui@8.0.1:
+  dom-serializer@2.0.0:
     dependencies:
-      string-width: 4.2.3
-      strip-ansi: 6.0.1
-      wrap-ansi: 7.0.0
+      domelementtype: 2.3.0
+      domhandler: 5.0.3
+      entities: 4.5.0
 
-  color-convert@2.0.1:
-    dependencies:
-      color-name: 1.1.4
+  domelementtype@2.3.0: {}
 
-  color-name@1.1.4: {}
+  domhandler@5.0.3:
+    dependencies:
+      domelementtype: 2.3.0
 
-  concurrently@9.2.1:
+  domutils@3.2.2:
     dependencies:
-      chalk: 4.1.2
-      rxjs: 7.8.2
-      shell-quote: 1.8.3
-      supports-color: 8.1.1
-      tree-kill: 1.2.2
-      yargs: 17.7.2
+      dom-serializer: 2.0.0
+      domelementtype: 2.3.0
+      domhandler: 5.0.3
 
-  emoji-regex@8.0.0: {}
+  entities@4.5.0: {}
+
+  entities@7.0.1: {}
 
   esbuild@0.27.2:
     optionalDependencies:
@@ -475,70 +424,23 @@ snapshots:
       '@esbuild/win32-ia32': 0.27.2
       '@esbuild/win32-x64': 0.27.2
 
-  escalade@3.2.0: {}
-
   eta@3.5.0: {}
 
-  get-caller-file@2.0.5: {}
-
-  has-flag@4.0.0: {}
-
   hono@4.11.5: {}
 
-  is-fullwidth-code-point@3.0.0: {}
-
-  prettier@3.8.1: {}
-
-  require-directory@2.1.1: {}
-
-  rxjs@7.8.2:
-    dependencies:
-      tslib: 2.8.1
-
-  shell-quote@1.8.3: {}
-
-  string-width@4.2.3:
+  htmlparser2@10.1.0:
     dependencies:
-      emoji-regex: 8.0.0
-      is-fullwidth-code-point: 3.0.0
-      strip-ansi: 6.0.1
+      domelementtype: 2.3.0
+      domhandler: 5.0.3
+      domutils: 3.2.2
+      entities: 7.0.1
 
-  strip-ansi@6.0.1:
+  nth-check@2.1.1:
     dependencies:
-      ansi-regex: 5.0.1
+      boolbase: 1.0.0
 
-  supports-color@7.2.0:
-    dependencies:
-      has-flag: 4.0.0
-
-  supports-color@8.1.1:
-    dependencies:
-      has-flag: 4.0.0
-
-  tree-kill@1.2.2: {}
-
-  tslib@2.8.1: {}
+  prettier@3.8.1: {}
 
   typescript@5.9.3: {}
 
   undici-types@6.21.0: {}
-
-  wrap-ansi@7.0.0:
-    dependencies:
-      ansi-styles: 4.3.0
-      string-width: 4.2.3
-      strip-ansi: 6.0.1
-
-  y18n@5.0.8: {}
-
-  yargs-parser@21.1.1: {}
-
-  yargs@17.7.2:
-    dependencies:
-      cliui: 8.0.1
-      escalade: 3.2.0
-      get-caller-file: 2.0.5
-      require-directory: 2.1.1
-      string-width: 4.2.3
-      y18n: 5.0.8
-      yargs-parser: 21.1.1
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 8c11317..33f00f1 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -172,6 +172,7 @@ describe("GitHub scraper", () => {
       assert.ok(firstRelease.tagName.length > 0);
       assert.ok(firstRelease.title.match(/^v\d/));
       assert.ok(firstRelease.publishedAt.length > 0);
+      assert.ok(firstRelease.bodyHtml.length > 0);
     });
   });
 
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index e52ddac..af39660 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -1,3 +1,8 @@
+import { parseDocument } from "htmlparser2";
+import { type Element } from "domhandler";
+import { getAttributeValue, getInnerHTML, textContent } from "domutils";
+import * as cssSelect from "css-select";
+
 export class GitHubHTTPError extends Error {
   status: number;
 
@@ -418,37 +423,36 @@ export async function getGitHubCommits(
 export async function getGitHubReleases(owner: string, repo: string): Promise<GitHubReleases> {
   const html = await fetchGitHubPage(`${owner}/${repo}/releases`);
 
-  // Releases page doesn't have embedded JSON, so we parse HTML directly
+  const document = parseDocument(html);
   const releases: Release[] = [];
 
-  // Find all h2 tags with version numbers
-  const h2Regex = /<h2[^>]*id="([^"]*)"[^>]*>(v[^<]*)<\/h2>/g;
-
-  for (const h2Match of html.matchAll(h2Regex)) {
-    const id = h2Match[1];
-    const title = h2Match[2].trim();
+  // Not really sure the right way to express this type.
+  const sections = cssSelect.selectAll("section", document) as unknown as Element[];
 
-    // Find the section containing this h2
-    const sectionRegex = new RegExp(
-      `<section[^>]*>\\s*<h2[^>]*id="${id}"[^>]*>[^<]*<\\/h2>([\\s\\S]*?)<\\/section>`,
-      "s",
-    );
-    const sectionMatch = html.match(sectionRegex);
-    if (!sectionMatch) continue;
+  for (const section of sections) {
+    const h2 = cssSelect.selectOne("h2[id]", section);
+    if (!h2) continue;
 
-    const sectionHtml = sectionMatch[1];
+    const title = textContent(h2).trim();
+    if (!title.startsWith("v")) continue;
 
     // Extract tag name from link
-    const tagMatch = sectionHtml.match(/<a[^>]*href="[^"]*\/tree\/([^"]+)"[^>]*>/);
-    const tagName = tagMatch?.[1] || title;
+    const tagLink = cssSelect.selectOne('a[href*="/tree/"]', section);
+    let tagName = title;
+    if (tagLink) {
+      const href = getAttributeValue(tagLink, "href");
+      if (href) {
+        tagName = href.split("/").pop() || title;
+      }
+    }
 
     // Extract published date
-    const dateMatch = sectionHtml.match(/<relative-time[^>]*datetime="([^"]*)"[^>]*>/);
-    const publishedAt = dateMatch?.[1] || "";
+    const relativeTime = cssSelect.selectOne("relative-time", section);
+    const publishedAt = relativeTime ? getAttributeValue(relativeTime, "datetime") || "" : "";
 
     // Extract body HTML (markdown content)
-    const bodyMatch = sectionHtml.match(/<div[^>]*class="[^"]*markdown-body[^"]*"[^>]*>([\s\S]*?)<\/div>/);
-    const bodyHtml = bodyMatch?.[1]?.trim() || "";
+    const body = cssSelect.selectOne(".markdown-body", section);
+    const bodyHtml = body ? getInnerHTML(body).trim() : "";
 
     releases.push({
       tagName,
-- 
cgit v1.3.1


From 371f558ed70d29f63e9fbee48c8db7ffab2f7cba Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Wed, 11 Feb 2026 02:09:26 -0500
Subject: Add single release detail pages and refactor templates

Refactored releases display into reusable partial template (_release.eta) and added support for viewing individual releases at /:owner/:repo/releases/tag/:tag with navigation back to all releases.

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
---
 cgithub/src/index.ts        |  6 +++
 cgithub/src/scraper.test.ts | 13 +++++++
 cgithub/src/scraper.ts      | 93 ++++++++++++++++++++++++++++++++++-----------
 cgithub/static/style.css    |  4 ++
 cgithub/views/_release.eta  | 23 +++++++++++
 cgithub/views/release.eta   |  7 ++++
 cgithub/views/releases.eta  | 20 +---------
 7 files changed, 124 insertions(+), 42 deletions(-)
 create mode 100644 cgithub/views/_release.eta
 create mode 100644 cgithub/views/release.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/index.ts b/cgithub/src/index.ts
index 2f8efef..59c98ee 100644
--- a/cgithub/src/index.ts
+++ b/cgithub/src/index.ts
@@ -9,6 +9,7 @@ import {
   getGitHubBlob,
   getGitHubCommits,
   getGitHubIssues,
+  getGitHubRelease,
   getGitHubReleases,
   getGitHubRepo,
   getGitHubTree,
@@ -74,6 +75,11 @@ app.get("/:owner/:repo/commits/:branch/:path{.*}?", async (c) => {
   return tryRender(c, "commits.eta", getGitHubCommits(owner, repo, branch, path));
 });
 
+app.get("/:owner/:repo/releases/tag/:tag", async (c) => {
+  const { owner, repo, tag } = c.req.param();
+  return tryRender(c, "release.eta", getGitHubRelease(owner, repo, tag));
+});
+
 app.get("/:owner/:repo/releases", async (c) => {
   const { owner, repo } = c.req.param();
   return tryRender(c, "releases.eta", getGitHubReleases(owner, repo));
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 33f00f1..0b91b6a 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -4,6 +4,7 @@ import {
   getGitHubBlob,
   getGitHubCommits,
   getGitHubIssues,
+  getGitHubRelease,
   getGitHubReleases,
   getGitHubRepo,
   getGitHubTree,
@@ -174,6 +175,18 @@ describe("GitHub scraper", () => {
       assert.ok(firstRelease.publishedAt.length > 0);
       assert.ok(firstRelease.bodyHtml.length > 0);
     });
+
+    it("should fetch single release", async () => {
+      const data = await getGitHubRelease("actions", "deploy-pages", "v4.0.5");
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+
+      assert.ok(data.release.tagName.length > 0);
+      assert.ok(data.release.title.length > 0);
+      assert.ok(data.release.publishedAt.length > 0);
+      assert.ok(data.release.bodyHtml.length > 0);
+    });
   });
 
   describe("rich file types", () => {
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index af39660..acfaac1 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -200,6 +200,10 @@ export interface GitHubReleases extends GitHubCommon {
   releases: Release[];
 }
 
+export interface GitHubRelease extends GitHubCommon {
+  release: Release;
+}
+
 async function fetchGitHubPage(path: string): Promise<string> {
   // GitHub throttles/blocks requests without realistic browser headers.
   // These headers make the request appear as a standard browser visit.
@@ -420,6 +424,33 @@ export async function getGitHubCommits(
   };
 }
 
+function parseRelease(section: Element, title: string): Release {
+  // Extract tag name from link
+  const tagLink = cssSelect.selectOne('a[href*="/tree/"]', section);
+  let tagName = title;
+  if (tagLink) {
+    const href = getAttributeValue(tagLink, "href");
+    if (href) {
+      tagName = href.split("/").pop() || title;
+    }
+  }
+
+  // Extract published date
+  const relativeTime = cssSelect.selectOne("relative-time", section);
+  const publishedAt = relativeTime ? getAttributeValue(relativeTime, "datetime") || "" : "";
+
+  // Extract body HTML (markdown content)
+  const body = cssSelect.selectOne(".markdown-body", section);
+  const bodyHtml = body ? getInnerHTML(body).trim() : "";
+
+  return {
+    tagName,
+    title,
+    publishedAt,
+    bodyHtml,
+  };
+}
+
 export async function getGitHubReleases(owner: string, repo: string): Promise<GitHubReleases> {
   const html = await fetchGitHubPage(`${owner}/${repo}/releases`);
 
@@ -436,37 +467,53 @@ export async function getGitHubReleases(owner: string, repo: string): Promise<Gi
     const title = textContent(h2).trim();
     if (!title.startsWith("v")) continue;
 
-    // Extract tag name from link
-    const tagLink = cssSelect.selectOne('a[href*="/tree/"]', section);
-    let tagName = title;
-    if (tagLink) {
-      const href = getAttributeValue(tagLink, "href");
-      if (href) {
-        tagName = href.split("/").pop() || title;
-      }
-    }
+    releases.push(parseRelease(section, title));
+  }
+
+  return {
+    repo: {
+      owner,
+      name: repo,
+    },
+    releases,
+  };
+}
 
-    // Extract published date
-    const relativeTime = cssSelect.selectOne("relative-time", section);
-    const publishedAt = relativeTime ? getAttributeValue(relativeTime, "datetime") || "" : "";
+export async function getGitHubRelease(
+  owner: string,
+  repo: string,
+  tag: string,
+): Promise<GitHubRelease> {
+  const html = await fetchGitHubPage(`${owner}/${repo}/releases/tag/${tag}`);
 
-    // Extract body HTML (markdown content)
-    const body = cssSelect.selectOne(".markdown-body", section);
-    const bodyHtml = body ? getInnerHTML(body).trim() : "";
+  // Find the h1 tag with the release title
+  const h1Match = html.match(/<h1[^>]*class="[^"]*d-inline[^"]*"[^>]*>([^<]*)<\/h1>/);
+  const title = h1Match?.[1]?.trim() || tag;
 
-    releases.push({
-      tagName,
-      title,
-      publishedAt,
-      bodyHtml,
-    });
-  }
+  // Extract tag name from tree link
+  const tagLinkMatch = html.match(/<a[^>]*href="[^"]*\/tree\/([^"]+)"[^>]*>/);
+  const tagName = tagLinkMatch?.[1] || tag;
+
+  // Extract published date
+  const dateMatch = html.match(/<relative-time[^>]*datetime="([^"]*)"[^>]*>/);
+  const publishedAt = dateMatch?.[1] || "";
+
+  // Extract body HTML (markdown content) - find the markdown-body div
+  const bodyMatch = html.match(
+    /<div[^>]*class="[^"]*markdown-body[^"]*"[^>]*>([\s\S]*?)<\/div>\s*<div[^>]*class="mb-3"/,
+  );
+  const bodyHtml = bodyMatch?.[1]?.trim() || "";
 
   return {
     repo: {
       owner,
       name: repo,
     },
-    releases,
+    release: {
+      tagName,
+      title,
+      publishedAt,
+      bodyHtml,
+    },
   };
 }
diff --git a/cgithub/static/style.css b/cgithub/static/style.css
index fdf3285..896f138 100644
--- a/cgithub/static/style.css
+++ b/cgithub/static/style.css
@@ -232,6 +232,10 @@ pre.box {
   font-size: 0.8rem;
 }
 
+.release-nav {
+  margin-bottom: 1rem;
+}
+
 .release {
   margin-bottom: 2rem;
 }
diff --git a/cgithub/views/_release.eta b/cgithub/views/_release.eta
new file mode 100644
index 0000000..85126a2
--- /dev/null
+++ b/cgithub/views/_release.eta
@@ -0,0 +1,23 @@
+<div class="box release">
+  <div class="release-header">
+    <h3>
+      <% if (it.linkToDetail) { %>
+        <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases/tag/<%= it.release.tagName %>">
+          <%= it.release.title %>
+        </a>
+      <% } else { %>
+        <%= it.release.title %>
+      <% } %>
+    </h3>
+    <% if (it.release.publishedAt) { %>
+      <time class="timestamp" title="<%= it.release.publishedAt %>">
+        <%= it.release.publishedAt.split("T")[0] %>
+      </time>
+    <% } %>
+  </div>
+  <% if (it.release.bodyHtml) { %>
+    <div class="release-body">
+      <%~ it.release.bodyHtml %>
+    </div>
+  <% } %>
+</div>
diff --git a/cgithub/views/release.eta b/cgithub/views/release.eta
new file mode 100644
index 0000000..b8c0fe6
--- /dev/null
+++ b/cgithub/views/release.eta
@@ -0,0 +1,7 @@
+<% layout("./layout.eta") %>
+
+<div class="release-nav">
+  <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases">← All releases</a>
+</div>
+
+<%~ include("./_release.eta", { repo: it.repo, release: it.release, linkToDetail: false }) %>
diff --git a/cgithub/views/releases.eta b/cgithub/views/releases.eta
index 39850d8..b3cb254 100644
--- a/cgithub/views/releases.eta
+++ b/cgithub/views/releases.eta
@@ -3,23 +3,5 @@
 <h2>Releases</h2>
 
 <% for (const release of it.releases) { %>
-  <div class="box release">
-    <div class="release-header">
-      <h3>
-        <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases/tag/<%= release.tagName %>">
-          <%= release.title %>
-        </a>
-      </h3>
-      <% if (release.publishedAt) { %>
-        <time class="timestamp" title="<%= release.publishedAt %>">
-          <%= release.publishedAt.split("T")[0] %>
-        </time>
-      <% } %>
-    </div>
-    <% if (release.bodyHtml) { %>
-      <div class="release-body">
-        <%~ release.bodyHtml %>
-      </div>
-    <% } %>
-  </div>
+  <%~ include("./_release.eta", { repo: it.repo, release, linkToDetail: true }) %>
 <% } %>
-- 
cgit v1.3.1


From 7b4521c0227663dadf78f611c1b515f7050015d0 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Sat, 21 Feb 2026 18:41:13 +1100
Subject: Parse repo info from DOM instead of regexes

By Gemini
---
 cgithub/src/scraper.ts | 61 +++++++++++++++++++++++++++++++++++++-------------
 1 file changed, 46 insertions(+), 15 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index acfaac1..d36df5b 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -241,27 +241,58 @@ function parsePayload(html: string, dataTarget: string, prop: string[]): RawPayl
 }
 
 function parseRepoInfo(html: string): RepoInfo {
-  const descMatch = html.match(/<p[^>]*class="[^"]*f4[^"]*"[^>]*>([\s\S]*?)<\/p>/);
-  const description = descMatch?.[1]?.replace(/<[^>]+>/g, "").trim() || null;
-  const websiteMatch = html.match(/<a[^>]*rel="noopener noreferrer nofollow"[^>]*href="([^"]+)"/);
-  const starsMatch = html.match(/id="repo-stars-counter-star"[^>]*>([^<]+)/);
-  const watchersMatch = html.match(/<strong>(\d+)<\/strong>\s*watching/);
-  const forksMatch = html.match(/id="repo-network-counter"[^>]*>([^<]+)/);
+  const document = parseDocument(html);
+
+  const aboutSection = cssSelect.selectOne(".BorderGrid-cell h2:contains('About')", document);
+  let description: string | null = null;
+  let website: string | null = null;
+
+  if (aboutSection) {
+    const parent = aboutSection.parent;
+    if (parent) {
+      const descElem = cssSelect.selectOne("p.f4", parent);
+      if (descElem) {
+        description = textContent(descElem).trim();
+      }
+      const websiteElem = cssSelect.selectOne(
+        "a[rel='noopener noreferrer nofollow']",
+        parent,
+      ) as Element;
+      if (websiteElem) {
+        website = getAttributeValue(websiteElem, "href") || null;
+      }
+    }
+  }
 
+  const starsElem = cssSelect.selectOne("#repo-stars-counter-star", document) as unknown as Element;
+  const stars = starsElem
+    ? getAttributeValue(starsElem, "title") || textContent(starsElem).trim()
+    : null;
+
+  const watchersElem = cssSelect.selectOne("a[href$='/watchers'] strong", document);
+  const watchers = watchersElem ? textContent(watchersElem).trim() : null;
+
+  const forksElem = cssSelect.selectOne("#repo-network-counter", document) as unknown as Element;
+  const forks = forksElem
+    ? getAttributeValue(forksElem, "title") || textContent(forksElem).trim()
+    : null;
+
+  const releaseElem = cssSelect.selectOne(
+    "a[href$='/releases'] .Counter",
+    document,
+  ) as unknown as Element;
   let numReleases: number | undefined;
-  const releaseCountMatch = html.match(
-    /<a[^>]*href="[^"]*\/releases"[^>]*>[\s\S]*?<span[^>]*title="(\d+)"[^>]*(?:class="[^"]*Counter|data-view-component="true")[^>]*>/,
-  );
-  if (releaseCountMatch) {
-    numReleases = parseInt(releaseCountMatch[1], 10);
+  if (releaseElem) {
+    const releaseText = getAttributeValue(releaseElem, "title") || textContent(releaseElem);
+    numReleases = parseInt(releaseText.replace(/,/g, ""), 10);
   }
 
   return {
     description,
-    website: websiteMatch?.[1] || null,
-    stars: starsMatch?.[1]?.trim() || null,
-    watchers: watchersMatch?.[1] || null,
-    forks: forksMatch?.[1]?.trim() || null,
+    website,
+    stars,
+    watchers,
+    forks,
     numReleases,
   };
 }
-- 
cgit v1.3.1


From 8319c4867da25739ee220128e050f0f965f6ed0b Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Sat, 21 Feb 2026 19:01:55 +1100
Subject: Add latest release details

By Gemini
---
 cgithub/src/scraper.test.ts |  2 ++
 cgithub/src/scraper.ts      | 24 ++++++++++++++++++++++++
 cgithub/views/repo.eta      | 11 ++++++++++-
 3 files changed, 36 insertions(+), 1 deletion(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 0b91b6a..5bd609a 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -30,6 +30,8 @@ describe("GitHub scraper", () => {
       assert.strictEqual(data.info.watchers?.length, 2);
       assert.strictEqual(data.info.forks?.length, 3);
       assert.strictEqual(data.info.numReleases, 36);
+      assert.strictEqual(data.info.latestReleaseTag, "v4.0.5");
+      assert.ok(data.info.latestReleaseDate?.startsWith("2024-03-18"));
 
       assert.ok(data.items.length > 0);
       assert.ok(Object.keys(data.overviewHtml ?? {}).includes("README.md"));
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index d36df5b..0ed77b5 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -157,6 +157,8 @@ export interface RepoInfo {
   watchers: string | null;
   forks: string | null;
   numReleases?: number;
+  latestReleaseTag?: string;
+  latestReleaseDate?: string;
 }
 
 export interface GitHubRepo extends GitHubTree {
@@ -282,9 +284,29 @@ function parseRepoInfo(html: string): RepoInfo {
     document,
   ) as unknown as Element;
   let numReleases: number | undefined;
+  let latestReleaseTag: string | undefined;
+  let latestReleaseDate: string | undefined;
+
   if (releaseElem) {
     const releaseText = getAttributeValue(releaseElem, "title") || textContent(releaseElem);
     numReleases = parseInt(releaseText.replace(/,/g, ""), 10);
+
+    // Find the latest release tag and date if available
+    const cell = (releaseElem.parent as Element)?.parent?.parent as Element; // span -> a -> h2 -> cell
+    if (cell) {
+      const latestLink = cssSelect.selectOne("a[href*='/releases/tag/']", cell) as Element;
+      if (latestLink) {
+        const tagElem = cssSelect.selectOne(".css-truncate-target", latestLink);
+        if (tagElem) {
+          latestReleaseTag = textContent(tagElem).trim();
+        }
+
+        const timeElem = cssSelect.selectOne("relative-time", latestLink);
+        if (timeElem) {
+          latestReleaseDate = getAttributeValue(timeElem, "datetime") || undefined;
+        }
+      }
+    }
   }
 
   return {
@@ -294,6 +316,8 @@ function parseRepoInfo(html: string): RepoInfo {
     watchers,
     forks,
     numReleases,
+    latestReleaseTag,
+    latestReleaseDate,
   };
 }
 
diff --git a/cgithub/views/repo.eta b/cgithub/views/repo.eta
index a6fcfb3..6204b69 100644
--- a/cgithub/views/repo.eta
+++ b/cgithub/views/repo.eta
@@ -18,7 +18,16 @@
       <span class="detail"><img src="/static/icons/git-fork.svg" alt="forks" class="icon"> <%= it.info.forks %></span>
     <% } %>
     <% if (it.info.numReleases) { %>
-      <div class="detail"><a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases"><%= it.info.numReleases %> releases</a></div>
+      <div class="detail">
+        <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases"><%= it.info.numReleases %> releases</a>
+        <% if (it.info.latestReleaseTag) { %>
+          (<a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases/tag/<%= it.info.latestReleaseTag %>"><%= it.info.latestReleaseTag %></a>
+          <% if (it.info.latestReleaseDate) { %>
+            <time class="timestamp" title="<%= it.info.latestReleaseDate %>"><%= it.info.latestReleaseDate.split("T")[0] %></time>
+          <% } %>
+          )
+        <% } %>
+      </div>
     <% } %>
   </div>
 <% } %>
-- 
cgit v1.3.1


From 0d95c588319e6f25f4b24612e730f662e209b5dc Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Sat, 21 Feb 2026 21:53:32 +1100
Subject: Release assets

Also fix release content parsing.
---
 cgithub/src/scraper.test.ts |   6 +++
 cgithub/src/scraper.ts      | 100 ++++++++++++++++++++++++++++++++------------
 cgithub/static/style.css    |  14 ++++++-
 cgithub/views/_release.eta  |  13 ++++++
 cgithub/views/layout.eta    |   2 +-
 5 files changed, 106 insertions(+), 29 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 5bd609a..125e6a6 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -188,6 +188,12 @@ describe("GitHub scraper", () => {
       assert.ok(data.release.title.length > 0);
       assert.ok(data.release.publishedAt.length > 0);
       assert.ok(data.release.bodyHtml.length > 0);
+      assert.ok(data.release.assets && data.release.assets.length >= 2);
+      assert.strictEqual(data.release.assets[0].name, "Source code (zip)");
+      assert.strictEqual(
+        data.release.assets[0].url,
+        "https://github.com/actions/deploy-pages/archive/refs/tags/v4.0.5.zip",
+      );
     });
   });
 
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 0ed77b5..769b870 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -196,6 +196,14 @@ interface Release {
   title: string;
   publishedAt: string;
   bodyHtml: string;
+  assets?: ReleaseAsset[];
+}
+
+export interface ReleaseAsset {
+  name: string;
+  url: string;
+  size?: string;
+  publishedAt?: string;
 }
 
 export interface GitHubReleases extends GitHubCommon {
@@ -259,14 +267,14 @@ function parseRepoInfo(html: string): RepoInfo {
       const websiteElem = cssSelect.selectOne(
         "a[rel='noopener noreferrer nofollow']",
         parent,
-      ) as Element;
+      ) as Element | null;
       if (websiteElem) {
         website = getAttributeValue(websiteElem, "href") || null;
       }
     }
   }
 
-  const starsElem = cssSelect.selectOne("#repo-stars-counter-star", document) as unknown as Element;
+  const starsElem = cssSelect.selectOne("#repo-stars-counter-star", document) as Element | null;
   const stars = starsElem
     ? getAttributeValue(starsElem, "title") || textContent(starsElem).trim()
     : null;
@@ -274,7 +282,7 @@ function parseRepoInfo(html: string): RepoInfo {
   const watchersElem = cssSelect.selectOne("a[href$='/watchers'] strong", document);
   const watchers = watchersElem ? textContent(watchersElem).trim() : null;
 
-  const forksElem = cssSelect.selectOne("#repo-network-counter", document) as unknown as Element;
+  const forksElem = cssSelect.selectOne("#repo-network-counter", document) as Element | null;
   const forks = forksElem
     ? getAttributeValue(forksElem, "title") || textContent(forksElem).trim()
     : null;
@@ -282,7 +290,7 @@ function parseRepoInfo(html: string): RepoInfo {
   const releaseElem = cssSelect.selectOne(
     "a[href$='/releases'] .Counter",
     document,
-  ) as unknown as Element;
+  ) as Element | null;
   let numReleases: number | undefined;
   let latestReleaseTag: string | undefined;
   let latestReleaseDate: string | undefined;
@@ -520,7 +528,6 @@ export async function getGitHubReleases(owner: string, repo: string): Promise<Gi
     if (!h2) continue;
 
     const title = textContent(h2).trim();
-    if (!title.startsWith("v")) continue;
 
     releases.push(parseRelease(section, title));
   }
@@ -534,41 +541,82 @@ export async function getGitHubReleases(owner: string, repo: string): Promise<Gi
   };
 }
 
-export async function getGitHubRelease(
+export async function getGitHubReleaseBase(
   owner: string,
   repo: string,
   tag: string,
 ): Promise<GitHubRelease> {
   const html = await fetchGitHubPage(`${owner}/${repo}/releases/tag/${tag}`);
 
-  // Find the h1 tag with the release title
-  const h1Match = html.match(/<h1[^>]*class="[^"]*d-inline[^"]*"[^>]*>([^<]*)<\/h1>/);
-  const title = h1Match?.[1]?.trim() || tag;
+  const document = parseDocument(html);
 
-  // Extract tag name from tree link
-  const tagLinkMatch = html.match(/<a[^>]*href="[^"]*\/tree\/([^"]+)"[^>]*>/);
-  const tagName = tagLinkMatch?.[1] || tag;
+  const box = cssSelect.selectOne(".Box:has(h1)", document) as Element | null;
+  if (!box) throw new Error("Could not find release box");
 
-  // Extract published date
-  const dateMatch = html.match(/<relative-time[^>]*datetime="([^"]*)"[^>]*>/);
-  const publishedAt = dateMatch?.[1] || "";
+  const h1 = cssSelect.selectOne("h1", box);
+  const title = h1 ? textContent(h1).trim() : "";
 
-  // Extract body HTML (markdown content) - find the markdown-body div
-  const bodyMatch = html.match(
-    /<div[^>]*class="[^"]*markdown-body[^"]*"[^>]*>([\s\S]*?)<\/div>\s*<div[^>]*class="mb-3"/,
-  );
-  const bodyHtml = bodyMatch?.[1]?.trim() || "";
+  const release = parseRelease(box, title);
 
   return {
     repo: {
       owner,
       name: repo,
     },
-    release: {
-      tagName,
-      title,
-      publishedAt,
-      bodyHtml,
-    },
+    release,
   };
 }
+
+export async function getGitHubReleaseAssets(
+  owner: string,
+  repo: string,
+  tag: string,
+): Promise<ReleaseAsset[]> {
+  const html = await fetchGitHubPage(`${owner}/${repo}/releases/expanded_assets/${tag}`);
+  const document = parseDocument(html);
+  const assets: ReleaseAsset[] = [];
+
+  for (const row of cssSelect.selectAll("li", document) as unknown as Element[]) {
+    const link = cssSelect.selectOne("a[href]", row);
+    if (!link) continue;
+
+    const url = getAttributeValue(link, "href") || "";
+    // Filename is in the link, often split into multiple spans
+    const name = textContent(link).replace(/\s+/g, " ").trim();
+
+    let size: string | undefined;
+    let publishedAt: string | undefined;
+    for (const s of cssSelect.selectAll("span.color-fg-muted.text-right", row)) {
+      // Size is in a span with color-fg-muted and text-right, but avoid the one with relative-time
+      const timeElem = cssSelect.selectOne("relative-time", row);
+      if (timeElem) {
+        publishedAt = timeElem ? getAttributeValue(timeElem, "datetime") || "" : "";
+      } else {
+        const text = textContent(s).trim();
+        if (text) size = text;
+      }
+    }
+
+    assets.push({
+      name,
+      url: url.startsWith("/") ? `https://github.com${url}` : url,
+      size,
+      publishedAt,
+    });
+  }
+
+  return assets;
+}
+
+export async function getGitHubRelease(
+  owner: string,
+  repo: string,
+  tag: string,
+): Promise<GitHubRelease> {
+  const [base, assets] = await Promise.all([
+    getGitHubReleaseBase(owner, repo, tag),
+    getGitHubReleaseAssets(owner, repo, tag),
+  ]);
+  base.release.assets = assets;
+  return base;
+}
diff --git a/cgithub/static/style.css b/cgithub/static/style.css
index fdf1b34..b3ebb46 100644
--- a/cgithub/static/style.css
+++ b/cgithub/static/style.css
@@ -111,11 +111,16 @@ section {
     margin-bottom: 0;
     padding: .5rem;
     border-bottom: 1px solid var(--border-color);
-    font-weight: bold;
+  }
+
+  > footer:last-child {
+    margin: 0 calc(-1 * var(--box-padding));
+    padding: .5rem;
+    border-top: 1px solid var(--border-color);
   }
 }
 
-footer {
+body > footer {
   margin-top: 1rem;
   text-align: center;
 }
@@ -207,11 +212,16 @@ pre.box {
   }
 }
 
+.overview > header {
+  font-weight: bold;
+}
+
 .release > header {
   display: flex;
   align-items: baseline;
   > :first-child {
     flex-grow: 1;
+    font-weight: bold;
   }
 }
 
diff --git a/cgithub/views/_release.eta b/cgithub/views/_release.eta
index 4769f9b..a8838fb 100644
--- a/cgithub/views/_release.eta
+++ b/cgithub/views/_release.eta
@@ -16,4 +16,17 @@
   <% if (it.release.bodyHtml) { %>
     <%~ it.release.bodyHtml %>
   <% } %>
+  <% if (it.release.assets && it.release.assets.length > 0) { %>
+    <footer>
+      <h3>Assets</h3>
+      <% for (const asset of it.release.assets) { %>
+        <div>
+          <a href="<%= asset.url %>"><%= asset.name %></a>
+          <% if (asset.publishedAt) { %>
+            <time class="timestamp" title="<%= asset.publishedAt %>"><%= asset.publishedAt.split("T")[0] %></time>
+          <% } %>
+        </div>
+      <% } %>
+    </footer>
+  <% } %>
 </div>
diff --git a/cgithub/views/layout.eta b/cgithub/views/layout.eta
index 3910d4b..c42d593 100644
--- a/cgithub/views/layout.eta
+++ b/cgithub/views/layout.eta
@@ -13,10 +13,10 @@ git clone git@github.com:<%= it.repo.owner %>/<%= it.repo.name %>.git</code></pr
     <nav>
       <a href="/<%= it.repo.owner %>/<%= it.repo.name %>">Code</a>
       <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/issues">Issues</a>
+      <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases">Releases</a>
       <% if (it.branch) { %>
         <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/commits/<%= it.branch %><%= it.path && it.path !== '/' ? '/' + it.path : '' %>">Commits</a>
       <% } %>
-      <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases">Releases</a>
     </nav>
   </div>
 
-- 
cgit v1.3.1


From 148272c64a97259ac874c4d8af1a93f197299012 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Thu, 12 Mar 2026 15:46:12 -0400
Subject: Fix GitHub scraping for new JSON structure, update TypeScript
 interfaces

GitHub restructured their embedded JSON: tree/blob data is now split
across codeViewRepoRoute, codeViewTreeRoute, codeViewBlobLayoutRoute,
codeViewBlobLayoutRoute.StyledBlob, codeViewBlobRoute, and the shared
codeViewLayoutRoute (repo, refInfo, path).

Replace RawPayload and OverviewFile with typed interfaces for each
route, add AppPayload as the top-level union, and simplify parsePayload
to always extract react-app.embeddedData.
---
 cgithub/src/scraper.test.ts |   7 +-
 cgithub/src/scraper.ts      | 159 ++++++++++++++++++++++++++------------------
 2 files changed, 99 insertions(+), 67 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 125e6a6..ad3ed46 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -95,8 +95,11 @@ describe("GitHub scraper", () => {
       assert.strictEqual(data.language, "Markdown");
       assert.strictEqual(data.image, false);
 
-      assert.strictEqual(data.textLines, null);
-      assert.strictEqual(data.htmlLines, null);
+      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>',
+      );
       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>/;
       assert.match(data.htmlContent ?? "", firstLine);
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 769b870..ac30259 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -22,7 +22,6 @@ interface TreeItem {
 interface OverviewFile {
   displayName: string;
   richText?: string;
-  loaded: boolean;
 }
 
 interface IssueNode {
@@ -79,34 +78,65 @@ export interface CommitGroup {
   commits: Commit[];
 }
 
-interface RawPayload {
+interface CodeViewLayoutRoute {
   repo: {
     ownerLogin: string;
     name: string;
   };
-  refInfo: { name: string };
+  refInfo: {
+    name: string;
+  };
   path: string;
-  tree?: {
+}
+
+interface CodeViewRepoRoute {
+  tree: {
     items: TreeItem[];
   };
-  blob?: {
-    headerInfo: {
-      blobSize: string;
-      lineInfo: {
-        truncatedLoc: string | null;
-        truncatedSloc: string | null;
-      };
-    };
-    language: string | null;
-    image: boolean;
-    rawLines: string[] | null;
-    colorizedLines: string[] | null;
-    richText: string | null;
-    stylingDirectives: StylingDirective[][] | null;
-  };
   overview?: {
     overviewFiles?: OverviewFile[];
   };
+}
+
+interface CodeViewTreeRoute {
+  tree: {
+    items: TreeItem[];
+  };
+}
+
+interface BlobHeaderInfo {
+  blobSize: string;
+  lineInfo?: {
+    truncatedLoc: string | null;
+    truncatedSloc: string | null;
+  };
+}
+
+interface CodeViewBlobLayoutRoute {
+  blob: {
+    headerInfo: BlobHeaderInfo;
+    language: string | null;
+    image: boolean;
+  };
+}
+
+interface CodeViewBlobLayoutRouteStyledBlob {
+  rawLines: string[] | null;
+  colorizedLines: string[] | null;
+  stylingDirectives: StylingDirective[][] | null;
+}
+
+interface CodeViewBlobRoute {
+  richText: string | null;
+}
+
+interface AppPayload {
+  codeViewLayoutRoute?: CodeViewLayoutRoute;
+  codeViewRepoRoute?: CodeViewRepoRoute;
+  codeViewTreeRoute?: CodeViewTreeRoute;
+  codeViewBlobLayoutRoute?: CodeViewBlobLayoutRoute;
+  "codeViewBlobLayoutRoute.StyledBlob"?: CodeViewBlobLayoutRouteStyledBlob;
+  codeViewBlobRoute?: CodeViewBlobRoute;
   commitGroups?: CommitGroup[];
   // Actually there may be other query types.
   preloadedQueries?: IssueIndexPageQuery[];
@@ -228,9 +258,9 @@ async function fetchGitHubPage(path: string): Promise<string> {
   return response.text();
 }
 
-function parsePayload(html: string, dataTarget: string, prop: string[]): RawPayload | null {
+function parsePayload(html: string): AppPayload | null {
   const regex = new RegExp(
-    `<script type="application/json" data-target="${dataTarget}">([^<]+)</script>`,
+    `<script type="application/json" data-target="react-app.embeddedData">([^<]+)</script>`,
     "g",
   );
 
@@ -241,11 +271,8 @@ function parsePayload(html: string, dataTarget: string, prop: string[]): RawPayl
     } catch {
       continue;
     }
-    for (const key of prop) {
-      data = data[key];
-      if (data === undefined) continue match;
-    }
-    return data;
+    if (data.payload === undefined) continue match;
+    return data.payload;
   }
   return null;
 }
@@ -329,24 +356,31 @@ function parseRepoInfo(html: string): RepoInfo {
   };
 }
 
-function extractOverviewHtml(payload: RawPayload): Record<string, string> | undefined {
+function extractOverviewHtml(
+  overviewFiles: OverviewFile[] = [],
+): Record<string, string> | undefined {
   const result: Record<string, string> = {};
-  for (const file of payload.overview?.overviewFiles ?? []) {
-    if (file.loaded && file.richText !== undefined) {
+  for (const file of overviewFiles) {
+    if (file.richText !== undefined) {
       result[file.displayName] = file.richText;
     }
   }
   return Object.keys(result).length > 0 ? result : undefined;
 }
 
-function extractGitHub<T>(payload: RawPayload, extra: T): GitHubNav & T {
+function extractGitHub<T>(payload: AppPayload | null, extra: T): GitHubNav & T {
+  const layoutRoute = payload?.codeViewLayoutRoute;
+  if (layoutRoute === undefined) {
+    throw new Error("Missing codeViewLayoutRoute");
+  }
+
   return {
     repo: {
-      owner: payload.repo.ownerLogin,
-      name: payload.repo.name,
+      owner: layoutRoute.repo.ownerLogin,
+      name: layoutRoute.repo.name,
     },
-    branch: payload.refInfo.name,
-    path: payload.path,
+    branch: layoutRoute.refInfo.name,
+    path: layoutRoute.path,
     ...extra,
   };
 }
@@ -354,16 +388,17 @@ function extractGitHub<T>(payload: RawPayload, extra: T): GitHubNav & T {
 export async function getGitHubRepo(owner: string, repo: string): Promise<GitHubRepo> {
   const html = await fetchGitHubPage(`${owner}/${repo}`);
 
-  const payload = parsePayload(html, "react-partial.embeddedData", ["props", "initialPayload"]);
+  const payload = parsePayload(html);
+  const repoRoute = payload?.codeViewRepoRoute;
 
-  if (payload?.tree === undefined) {
+  if (repoRoute?.tree === undefined) {
     throw new Error("Could not find tree data in embedded JSON");
   }
 
   const info = parseRepoInfo(html);
-  const overviewHtml = extractOverviewHtml(payload);
+  const overviewHtml = extractOverviewHtml(repoRoute.overview?.overviewFiles);
 
-  return extractGitHub(payload, { items: payload.tree.items, info, overviewHtml });
+  return extractGitHub(payload, { items: repoRoute.tree.items, info, overviewHtml });
 }
 
 export async function getGitHubTree(
@@ -374,20 +409,14 @@ export async function getGitHubTree(
 ): Promise<GitHubTree> {
   const html = await fetchGitHubPage(`${owner}/${repo}/tree/${branch}/${path}`);
 
-  // GitHub uses different formats for root vs subdirectories.
-  const payload =
-    // Root directory format.
-    parsePayload(html, "react-partial.embeddedData", ["props", "initialPayload"]) ||
-    // Subdirectory format.
-    parsePayload(html, "react-app.embeddedData", ["payload"]);
+  const payload = parsePayload(html);
+  const treeRoute = payload?.codeViewTreeRoute;
 
-  if (payload?.tree === undefined) {
+  if (treeRoute?.tree === undefined) {
     throw new Error("Could not find tree data in embedded JSON");
   }
 
-  const overviewHtml = extractOverviewHtml(payload);
-
-  return extractGitHub(payload, { items: payload.tree.items, overviewHtml });
+  return extractGitHub(payload, { items: treeRoute.tree.items });
 }
 
 export async function getGitHubBlob(
@@ -398,42 +427,42 @@ export async function getGitHubBlob(
 ): Promise<GitHubBlob> {
   const html = await fetchGitHubPage(`${owner}/${repo}/blob/${branch}/${path}`);
 
-  const payload = parsePayload(html, "react-app.embeddedData", ["payload"]);
+  const payload = parsePayload(html);
+  const blob = payload?.codeViewBlobLayoutRoute?.blob;
+  const styledBlob = payload?.["codeViewBlobLayoutRoute.StyledBlob"];
 
-  if (payload?.blob === undefined) {
+  if (blob === undefined) {
     throw new Error("Could not find blob data in embedded JSON");
   }
 
-  const blob = payload.blob;
-  const { headerInfo, stylingDirectives } = blob;
-
-  let htmlLines = blob.colorizedLines;
-  if (!htmlLines && stylingDirectives && blob.rawLines) {
-    htmlLines = blob.rawLines.map((line, i) => applyStyling(line, stylingDirectives[i]));
+  let htmlLines = styledBlob?.colorizedLines ?? null;
+  const stylingDirectives = styledBlob?.stylingDirectives;
+  if (!htmlLines && stylingDirectives && styledBlob?.rawLines) {
+    htmlLines = styledBlob.rawLines.map((line, i) => applyStyling(line, stylingDirectives[i]));
   }
 
-  let size = headerInfo.blobSize;
-  if (headerInfo.lineInfo.truncatedLoc) {
-    size += ` / ${headerInfo.lineInfo.truncatedLoc} lines`;
+  let size = blob.headerInfo.blobSize;
+  if (blob.headerInfo.lineInfo?.truncatedLoc) {
+    size += ` / ${blob.headerInfo.lineInfo.truncatedLoc} lines`;
   }
-  if (headerInfo.lineInfo.truncatedSloc) {
-    size += ` / ${headerInfo.lineInfo.truncatedSloc} loc`;
+  if (blob.headerInfo.lineInfo?.truncatedSloc) {
+    size += ` / ${blob.headerInfo.lineInfo.truncatedSloc} loc`;
   }
 
   return extractGitHub(payload, {
     language: blob.language || null,
     size,
     image: blob.image,
-    textLines: blob.rawLines,
+    textLines: styledBlob?.rawLines ?? null,
     htmlLines,
-    htmlContent: blob.richText,
+    htmlContent: payload?.codeViewBlobRoute?.richText ?? null,
   });
 }
 
 export async function getGitHubIssues(owner: string, repo: string): Promise<GitHubIssues> {
   const html = await fetchGitHubPage(`${owner}/${repo}/issues`);
 
-  const payload = parsePayload(html, "react-app.embeddedData", ["payload"]);
+  const payload = parsePayload(html);
 
   let issuesQuery: IssueIndexPageQuery | undefined;
   for (const query of payload?.preloadedQueries ?? []) {
@@ -470,7 +499,7 @@ export async function getGitHubCommits(
     : `${owner}/${repo}/commits/${branch}`;
   const html = await fetchGitHubPage(urlPath);
 
-  const payload = parsePayload(html, "react-app.embeddedData", ["payload"]);
+  const payload = parsePayload(html);
 
   if (payload?.commitGroups === undefined) {
     throw new Error("Could not find commit history in embedded JSON");
-- 
cgit v1.3.1


From 4662a229e06468a0805a18c6175bc0fd3a4ea985 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Sat, 14 Mar 2026 00:35:48 -0400
Subject: Fix tree overviewHtml: use tree.readme instead of
 overview.overviewFiles

---
 cgithub/src/scraper.test.ts | 25 +++++++++++--------------
 cgithub/src/scraper.ts      |  5 ++++-
 2 files changed, 15 insertions(+), 15 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index ad3ed46..1abab29 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -38,30 +38,27 @@ describe("GitHub scraper", () => {
     });
 
     it("should fetch tree", async () => {
-      const data = await getGitHubTree("actions", "deploy-pages", "main", "src");
+      const data = await getGitHubTree("github", "rally", "main", "lib");
 
-      assert.strictEqual(data.repo.owner, "actions");
-      assert.strictEqual(data.repo.name, "deploy-pages");
+      assert.strictEqual(data.repo.owner, "github");
+      assert.strictEqual(data.repo.name, "rally");
       assert.strictEqual(data.branch, "main");
-      assert.strictEqual(data.path, "src");
+      assert.strictEqual(data.path, "lib");
 
       assert.deepStrictEqual(data.items, [
         {
-          name: "__tests__",
-          path: "src/__tests__",
-          contentType: "directory",
-        },
-        {
-          name: "internal",
-          path: "src/internal",
-          contentType: "directory",
+          name: "README.md",
+          path: "lib/README.md",
+          contentType: "file",
         },
         {
-          name: "index.js",
-          path: "src/index.js",
+          name: "RallyValidate.js",
+          path: "lib/RallyValidate.js",
           contentType: "file",
         },
       ]);
+
+      assert.ok(Object.keys(data.overviewHtml ?? {}).includes("README.md"));
     });
 
     it("should fetch text blob", async () => {
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index ac30259..cd743a0 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -101,6 +101,7 @@ interface CodeViewRepoRoute {
 interface CodeViewTreeRoute {
   tree: {
     items: TreeItem[];
+    readme?: OverviewFile;
   };
 }
 
@@ -416,7 +417,9 @@ export async function getGitHubTree(
     throw new Error("Could not find tree data in embedded JSON");
   }
 
-  return extractGitHub(payload, { items: treeRoute.tree.items });
+  const overviewHtml = extractOverviewHtml(treeRoute.tree.readme ? [treeRoute.tree.readme] : []);
+
+  return extractGitHub(payload, { items: treeRoute.tree.items, overviewHtml });
 }
 
 export async function getGitHubBlob(
-- 
cgit v1.3.1


From d9a6f6cc7c6a374647486530c6246df980c4be6b Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Mon, 20 Apr 2026 17:14:11 -0400
Subject: Fallback to issue titleHtml

---
 cgithub/src/scraper.ts | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index cd743a0..5d17e3f 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -26,7 +26,8 @@ interface OverviewFile {
 
 interface IssueNode {
   number: number;
-  title: string;
+  title?: string;
+  titleHtml?: string;
   author?: {
     login: string;
   };
@@ -478,8 +479,9 @@ export async function getGitHubIssues(owner: string, repo: string): Promise<GitH
   }
 
   const issues: Issue[] = issuesQuery.result.data.repository.search.edges.map((edge) => {
-    const { number, title, state, createdAt } = edge.node;
-    return { number, title, state, createdAt };
+    const { number, title, titleHtml, state, createdAt } = edge.node;
+    const finalTitle = (title || titleHtml || "").replace(/<[^>]+>/g, "");
+    return { number, title: finalTitle, state, createdAt };
   });
 
   return {
-- 
cgit v1.3.1


From dbe8141dce07559c1e16f17800dfd505b0ecdb46 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Mon, 20 Apr 2026 17:14:11 -0400
Subject: Apply overlapping styling drectives instead of duplicating content

---
 cgithub/src/scraper.test.ts | 24 +++++++++++++++++++++++-
 cgithub/src/scraper.ts      | 45 ++++++++++++++++++++++++++++++++++++++-------
 2 files changed, 61 insertions(+), 8 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index a843561..4432231 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/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/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 5d17e3f..2dcd1d0 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/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;
 }
 
-- 
cgit v1.3.1


From cfe74db6c177151da3d1bc8549732e3d39448040 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Sat, 25 Apr 2026 13:35:17 -0400
Subject: Support redirects

---
 cgithub/src/index.ts        |  5 ++++-
 cgithub/src/scraper.test.ts | 26 ++++++++++++++++++++++++++
 cgithub/src/scraper.ts      | 21 ++++++++++++++++++++-
 3 files changed, 50 insertions(+), 2 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/index.ts b/cgithub/src/index.ts
index 59c98ee..eb3a0ee 100644
--- a/cgithub/src/index.ts
+++ b/cgithub/src/index.ts
@@ -14,6 +14,7 @@ import {
   getGitHubRepo,
   getGitHubTree,
   GitHubHTTPError,
+  RedirectError,
 } from "./scraper.ts";
 
 const __dirname = path.dirname(url.fileURLToPath(import.meta.url));
@@ -32,7 +33,9 @@ async function tryRender<T extends object>(c: Context, template: string, promise
   try {
     data = await promise;
   } catch (e) {
-    if (e instanceof GitHubHTTPError) {
+    if (e instanceof RedirectError) {
+      return c.redirect(e.location);
+    } else if (e instanceof GitHubHTTPError) {
       c.status(e.status as StatusCode);
       const message = `GitHub responded with HTTP ${e.status} ${e.message}`;
       return c.html(await eta.renderAsync("error.eta", { title: e.message, message }));
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 4432231..c20a5ff 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -9,6 +9,7 @@ import {
   getGitHubRepo,
   getGitHubTree,
   GitHubHTTPError,
+  RedirectError,
 } from "./scraper.ts";
 
 describe("GitHub scraper", () => {
@@ -245,5 +246,30 @@ describe("GitHub scraper", () => {
         return true;
       });
     });
+
+    it("should redirect when fetching a directory as a blob", async () => {
+      await assert.rejects(
+        getGitHubBlob("badlogic", "pi-mono", "main", "packages/agent"),
+        (err) => {
+          assert(err instanceof RedirectError, "error should be a RedirectError");
+          assert.strictEqual(err.location, "/badlogic/pi-mono/tree/main/packages/agent");
+          return true;
+        },
+      );
+    });
+
+    it("should redirect when fetching a file as a tree", async () => {
+      await assert.rejects(
+        getGitHubTree("badlogic", "pi-mono", "main", "packages/agent/package.json"),
+        (err) => {
+          assert(err instanceof RedirectError, "error should be a RedirectError");
+          assert.strictEqual(
+            err.location,
+            "/badlogic/pi-mono/blob/main/packages/agent/package.json",
+          );
+          return true;
+        },
+      );
+    });
   });
 });
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 2dcd1d0..04a685b 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -13,6 +13,16 @@ export class GitHubHTTPError extends Error {
   }
 }
 
+export class RedirectError extends Error {
+  location: string;
+
+  constructor(location: string) {
+    super(`Redirect to ${location}`);
+    this.name = "RedirectError";
+    this.location = location;
+  }
+}
+
 interface TreeItem {
   contentType: "directory" | "file";
   name: string;
@@ -286,7 +296,16 @@ async function fetchGitHubPage(path: string): Promise<string> {
     Accept: "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
     "Accept-Language": "en-US,en;q=0.5",
   };
-  const response = await fetch(`https://github.com/${path}`, { headers });
+  const response = await fetch(`https://github.com/${path}`, { headers, redirect: "manual" });
+  if (response.status === 301 || response.status === 302) {
+    const location = response.headers.get("location");
+    if (location) {
+      const url = new URL(location);
+      if (url.hostname === "github.com") {
+        throw new RedirectError(url.pathname);
+      }
+    }
+  }
   if (!response.ok) throw new GitHubHTTPError(response.status, response.statusText);
   return response.text();
 }
-- 
cgit v1.3.1


From c57123139794c3ef174b62c6bd2a54a60bbf3327 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Wed, 27 May 2026 14:02:36 -0400
Subject: Add repo-scoped search support for issues and PRs

We do not support hybrid issue/PR search. Any searches on the issues
endpoint will redirect to an explicit issue-only search.
---
 cgithub/src/index.ts        | 26 +++++++++++++++++++++++++-
 cgithub/src/scraper.test.ts | 15 +++++++++++++++
 cgithub/src/scraper.ts      | 20 ++++++++++----------
 cgithub/static/style.css    | 13 ++++++++++---
 cgithub/views/blob.eta      |  2 +-
 cgithub/views/issues.eta    |  2 +-
 cgithub/views/layout.eta    |  8 ++++++++
 cgithub/views/tree.eta      |  2 +-
 8 files changed, 71 insertions(+), 17 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/index.ts b/cgithub/src/index.ts
index eb3a0ee..db9915f 100644
--- a/cgithub/src/index.ts
+++ b/cgithub/src/index.ts
@@ -70,7 +70,31 @@ app.get("/:owner/:repo/raw/:branch/:path{.*}", async (c) => {
 
 app.get("/:owner/:repo/issues", async (c) => {
   const { owner, repo } = c.req.param();
-  return tryRender(c, "issues.eta", getGitHubIssues(owner, repo));
+  const q = c.req.query("q");
+
+  // GitHub supports combined issue/PR search, apparently for legacy. We just
+  // redirect to issue-only search.
+  if (q && !/\bis:issue\b/.test(q)) {
+    return c.redirect(`/${owner}/${repo}/issues?q=${encodeURIComponent(`is:issue ${q}`)}`);
+  }
+
+  return tryRender(c, "issues.eta", getGitHubIssues(owner, repo, { q }));
+});
+
+app.get("/:owner/:repo/search", async (c) => {
+  const { owner, repo } = c.req.param();
+  const q = c.req.query("q") ?? "";
+  const type = c.req.query("type") ?? "issues";
+
+  if (type === "issues") {
+    return c.redirect(`/${owner}/${repo}/issues?q=${encodeURIComponent(q)}`);
+  }
+
+  // Redirect unhandled search types (like code, which requires sign in anyway).
+  c.header("Referrer-Policy", "no-referrer");
+  return c.redirect(
+    `https://github.com/${owner}/${repo}/search?${new URLSearchParams(c.req.query()).toString()}`,
+  );
 });
 
 app.get("/:owner/:repo/commits/:branch/:path{.*}?", async (c) => {
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 42ddd6e..236b0f5 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -127,6 +127,7 @@ describe("GitHub scraper", () => {
 
       assert.strictEqual(data.repo.owner, "actions");
       assert.strictEqual(data.repo.name, "deploy-pages");
+      assert.strictEqual(data.q, undefined);
 
       const issue402 = data.issues.find((issue) => issue.number === 402);
       assert.ok(issue402);
@@ -135,6 +136,20 @@ describe("GitHub scraper", () => {
       assert.strictEqual(issue402.createdAt, "2025-07-03T19:07:08Z");
     });
 
+    it("should fetch filtered issues with q", async () => {
+      const data = await getGitHubIssues("actions", "deploy-pages", {
+        q: "is:issue state:closed",
+      });
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+      assert.strictEqual(data.q, "is:issue state:closed");
+      const issue417 = data.issues.find((issue) => issue.number === 417);
+      assert.ok(issue417);
+      assert.strictEqual(issue417.title, "Chats");
+      assert.ok(data.issues.every((issue) => issue.title !== ""));
+    });
+
     it("should fetch commits", async () => {
       const data = await getGitHubCommits("actions", "deploy-pages", "main", "");
 
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 04a685b..2412e49 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -258,6 +258,7 @@ interface Issue {
 
 export interface GitHubIssues extends GitHubCommon {
   issues: Issue[];
+  q?: string;
 }
 
 export interface GitHubCommits extends GitHubNav {
@@ -513,9 +514,14 @@ export async function getGitHubBlob(
   });
 }
 
-export async function getGitHubIssues(owner: string, repo: string): Promise<GitHubIssues> {
-  const html = await fetchGitHubPage(`${owner}/${repo}/issues`);
-
+export async function getGitHubIssues(
+  owner: string,
+  repo: string,
+  // We only support is:issue searches (no PR results).
+  { q }: { q?: string } = {},
+): Promise<GitHubIssues> {
+  const suffix = q ? `?q=${encodeURIComponent(q)}` : "";
+  const html = await fetchGitHubPage(`${owner}/${repo}/issues${suffix}`);
   const payload = parsePayload(html);
 
   let issuesQuery: IssueIndexPageQuery | undefined;
@@ -534,13 +540,7 @@ export async function getGitHubIssues(owner: string, repo: string): Promise<GitH
     return { number, title: finalTitle, state, createdAt };
   });
 
-  return {
-    repo: {
-      owner,
-      name: repo,
-    },
-    issues,
-  };
+  return { repo: { owner, name: repo }, issues, q };
 }
 
 export async function getGitHubCommits(
diff --git a/cgithub/static/style.css b/cgithub/static/style.css
index b3ebb46..86a67d8 100644
--- a/cgithub/static/style.css
+++ b/cgithub/static/style.css
@@ -43,14 +43,20 @@ body {
   color: var(--text-color);
 }
 
+a {
+  text-decoration: none;
+  color: var(--link-color);
+}
+
 h1, h2, h3, h4, h5, h6,
 p, pre, ul, ol {
   margin: .5em 0;
 }
 
-a {
-  text-decoration: none;
-  color: var(--link-color);
+input, select, button {
+  background-color: var(--body-background);
+  color: var(--text-color);
+  border: 1px solid var(--border-color);
 }
 
 table {
@@ -75,6 +81,7 @@ td {
 
 nav {
   display: flex;
+  margin: .5em 0;
   > a {
     display: block;
     padding-right: .5rem;
diff --git a/cgithub/views/blob.eta b/cgithub/views/blob.eta
index ea07059..50f92a9 100644
--- a/cgithub/views/blob.eta
+++ b/cgithub/views/blob.eta
@@ -1,4 +1,4 @@
-<% layout("./layout.eta") %>
+<% layout("./layout.eta"); it.searchType = "code" %>
 
 <%~ include("./_nav.eta", it) %>
 
diff --git a/cgithub/views/issues.eta b/cgithub/views/issues.eta
index 6e4c9ba..646da46 100644
--- a/cgithub/views/issues.eta
+++ b/cgithub/views/issues.eta
@@ -1,4 +1,4 @@
-<% layout("./layout") %>
+<% layout("./layout"); it.searchType = "issues" %>
 
 <table>
   <% for (const issue of it.issues) { %>
diff --git a/cgithub/views/layout.eta b/cgithub/views/layout.eta
index ec3c7b8..f31c6e0 100644
--- a/cgithub/views/layout.eta
+++ b/cgithub/views/layout.eta
@@ -19,6 +19,14 @@ git clone git@github.com:<%= it.repo.owner %>/<%= it.repo.name %>.git</code></pr
         <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/commits/<%= it.branch %><%= it.path && it.path !== '/' ? '/' + it.path : '' %>">Commits</a>
       <% } %>
     </nav>
+    <form action="/<%= it.repo.owner %>/<%= it.repo.name %>/search" method="get">
+      <input type="text" name="q" placeholder="Search…" value="<%= it.q ?? "" %>">
+      <select name="type">
+        <option value="issues" <%= it.searchType === "issues" ? "selected" : "" %>>Issues</option>
+        <option value="code" <%= it.searchType === "code" || it.searchType == null ? "selected" : "" %>>Code</option>
+      </select>
+      <button type="submit">Search</button>
+    </form>
   </div>
 
   <%~ it.body %>
diff --git a/cgithub/views/tree.eta b/cgithub/views/tree.eta
index 6e675e1..41ded06 100644
--- a/cgithub/views/tree.eta
+++ b/cgithub/views/tree.eta
@@ -1,3 +1,3 @@
-<% layout("./layout.eta") %>
+<% layout("./layout.eta"); it.searchType = "code" %>
 
 <%~ include("./_tree.eta", it) %>
-- 
cgit v1.3.1


From 745ac9d83234cb060bf802527efd90d301ecc491 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Wed, 27 May 2026 14:02:36 -0400
Subject: Pull request index

---
 cgithub/src/index.ts        |  9 +++++++++
 cgithub/src/scraper.test.ts | 11 +++++++++++
 cgithub/src/scraper.ts      | 40 ++++++++++++++++++++++++++++++++++++++++
 cgithub/views/layout.eta    |  2 ++
 cgithub/views/pulls.eta     | 15 +++++++++++++++
 5 files changed, 77 insertions(+)
 create mode 100644 cgithub/views/pulls.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/index.ts b/cgithub/src/index.ts
index db9915f..514f2e2 100644
--- a/cgithub/src/index.ts
+++ b/cgithub/src/index.ts
@@ -9,6 +9,7 @@ import {
   getGitHubBlob,
   getGitHubCommits,
   getGitHubIssues,
+  getGitHubPulls,
   getGitHubRelease,
   getGitHubReleases,
   getGitHubRepo,
@@ -81,6 +82,12 @@ app.get("/:owner/:repo/issues", async (c) => {
   return tryRender(c, "issues.eta", getGitHubIssues(owner, repo, { q }));
 });
 
+app.get("/:owner/:repo/pulls", async (c) => {
+  const { owner, repo } = c.req.param();
+  const q = c.req.query("q");
+  return tryRender(c, "pulls.eta", getGitHubPulls(owner, repo, { q }));
+});
+
 app.get("/:owner/:repo/search", async (c) => {
   const { owner, repo } = c.req.param();
   const q = c.req.query("q") ?? "";
@@ -88,6 +95,8 @@ app.get("/:owner/:repo/search", async (c) => {
 
   if (type === "issues") {
     return c.redirect(`/${owner}/${repo}/issues?q=${encodeURIComponent(q)}`);
+  } else if (type === "pullrequests") {
+    return c.redirect(`/${owner}/${repo}/pulls?q=${encodeURIComponent(q)}`);
   }
 
   // Redirect unhandled search types (like code, which requires sign in anyway).
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 236b0f5..a4b834e 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -4,6 +4,7 @@ import {
   getGitHubBlob,
   getGitHubCommits,
   getGitHubIssues,
+  getGitHubPulls,
   getGitHubRelease,
   getGitHubReleases,
   getGitHubRepo,
@@ -150,6 +151,16 @@ describe("GitHub scraper", () => {
       assert.ok(data.issues.every((issue) => issue.title !== ""));
     });
 
+    it("should fetch pull requests", async () => {
+      const data = await getGitHubPulls("actions", "deploy-pages", { q: "state:closed" });
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+      const pr411 = data.pulls.find((pull) => pull.number === 411);
+      assert.ok(pr411);
+      assert.strictEqual(pr411.title, "update node version");
+    });
+
     it("should fetch commits", async () => {
       const data = await getGitHubCommits("actions", "deploy-pages", "main", "");
 
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 2412e49..f32a8da 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -261,6 +261,11 @@ export interface GitHubIssues extends GitHubCommon {
   q?: string;
 }
 
+export interface GitHubPulls extends GitHubCommon {
+  pulls: Issue[];
+  q?: string;
+}
+
 export interface GitHubCommits extends GitHubNav {
   commitGroups: CommitGroup[];
 }
@@ -543,6 +548,41 @@ export async function getGitHubIssues(
   return { repo: { owner, name: repo }, issues, q };
 }
 
+export async function getGitHubPulls(
+  owner: string,
+  repo: string,
+  { q }: { q?: string } = {},
+): Promise<GitHubPulls> {
+  const suffix = q ? `?q=${encodeURIComponent(q)}` : "";
+  const html = await fetchGitHubPage(`${owner}/${repo}/pulls${suffix}`);
+  const document = parseDocument(html);
+
+  const pulls: Issue[] = [];
+  for (const row of cssSelect.selectAll(".js-issue-row", document) as unknown as Element[]) {
+    const id = getAttributeValue(row, "id"); // "issue_NNN"
+    const number = id ? parseInt(id.replace("issue_", ""), 10) : 0;
+    if (!number) continue;
+
+    const titleLink = cssSelect.selectOne("a.markdown-title", row);
+    const title = titleLink ? textContent(titleLink).trim() : "";
+
+    const timeElem = cssSelect.selectOne("relative-time", row);
+    const createdAt = timeElem ? getAttributeValue(timeElem, "datetime") || "" : "";
+
+    const stateSpan = cssSelect.selectOne(".tooltipped[aria-label]", row) as Element | null;
+    const ariaLabel = stateSpan ? (getAttributeValue(stateSpan, "aria-label") ?? "") : "";
+    const state = ariaLabel.includes("Merged")
+      ? "MERGED"
+      : ariaLabel.includes("Closed")
+        ? "CLOSED"
+        : "OPEN";
+
+    pulls.push({ number, title, state, createdAt });
+  }
+
+  return { repo: { owner, name: repo }, pulls, q };
+}
+
 export async function getGitHubCommits(
   owner: string,
   repo: string,
diff --git a/cgithub/views/layout.eta b/cgithub/views/layout.eta
index f31c6e0..85f431c 100644
--- a/cgithub/views/layout.eta
+++ b/cgithub/views/layout.eta
@@ -14,6 +14,7 @@ git clone git@github.com:<%= it.repo.owner %>/<%= it.repo.name %>.git</code></pr
     <nav>
       <a href="/<%= it.repo.owner %>/<%= it.repo.name %>">Code</a>
       <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/issues">Issues</a>
+      <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/pulls">Pull requests</a>
       <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases">Releases</a>
       <% if (it.branch) { %>
         <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/commits/<%= it.branch %><%= it.path && it.path !== '/' ? '/' + it.path : '' %>">Commits</a>
@@ -23,6 +24,7 @@ git clone git@github.com:<%= it.repo.owner %>/<%= it.repo.name %>.git</code></pr
       <input type="text" name="q" placeholder="Search…" value="<%= it.q ?? "" %>">
       <select name="type">
         <option value="issues" <%= it.searchType === "issues" ? "selected" : "" %>>Issues</option>
+        <option value="pullrequests" <%= it.searchType === "pullrequests" ? "selected" : "" %>>Pull requests</option>
         <option value="code" <%= it.searchType === "code" || it.searchType == null ? "selected" : "" %>>Code</option>
       </select>
       <button type="submit">Search</button>
diff --git a/cgithub/views/pulls.eta b/cgithub/views/pulls.eta
new file mode 100644
index 0000000..9ec5ca1
--- /dev/null
+++ b/cgithub/views/pulls.eta
@@ -0,0 +1,15 @@
+<% layout("./layout"); it.searchType = "pullrequests" %>
+
+<table>
+  <% for (const pull of it.pulls) { %>
+    <tr>
+      <td>#<%= pull.number %></td>
+      <td>
+        <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/pull/<%= pull.number %>">
+          <%= pull.title %>
+        </a>
+      </td>
+      <td class="timestamp"><time title="<%= pull.createdAt %>"><%= pull.createdAt.split("T")[0] %></time></td>
+    </tr>
+  <% } %>
+</table>
-- 
cgit v1.3.1


From c07865241ce113cbb26a939fd43fdf3461092c94 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Wed, 15 Jul 2026 00:16:30 -0400
Subject: User profile pages

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---
 cgithub/CLAUDE.md           |  1 +
 cgithub/src/index.ts        |  9 +++++
 cgithub/src/scraper.test.ts | 49 +++++++++++++++++++++++
 cgithub/src/scraper.ts      | 96 ++++++++++++++++++++++++++++++++++++++++++++-
 cgithub/static/style.css    |  5 +++
 cgithub/views/home.eta      |  2 +-
 cgithub/views/user.eta      | 58 +++++++++++++++++++++++++++
 7 files changed, 218 insertions(+), 2 deletions(-)
 create mode 100644 cgithub/views/user.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/CLAUDE.md b/cgithub/CLAUDE.md
index 8feb529..f6d8524 100644
--- a/cgithub/CLAUDE.md
+++ b/cgithub/CLAUDE.md
@@ -26,6 +26,7 @@ pnpm test
 **Templates:** `views/*.eta` - Eta templates with layout inheritance (`layout.eta` as base)
 
 **Routes:**
+- `GET /:user` - User profile with pinned repositories
 - `GET /:owner/:repo` - Repository info with directory listing
 - `GET /:owner/:repo/tree/:branch/:path*` - Directory listing
 - `GET /:owner/:repo/blob/:branch/:path` - File content view
diff --git a/cgithub/src/index.ts b/cgithub/src/index.ts
index 514f2e2..d8bdfec 100644
--- a/cgithub/src/index.ts
+++ b/cgithub/src/index.ts
@@ -14,6 +14,7 @@ import {
   getGitHubReleases,
   getGitHubRepo,
   getGitHubTree,
+  getGitHubUser,
   GitHubHTTPError,
   RedirectError,
 } from "./scraper.ts";
@@ -35,6 +36,9 @@ async function tryRender<T extends object>(c: Context, template: string, promise
     data = await promise;
   } catch (e) {
     if (e instanceof RedirectError) {
+      if (e.location.startsWith("https://")) {
+        c.header("Referrer-Policy", "no-referrer");
+      }
       return c.redirect(e.location);
     } else if (e instanceof GitHubHTTPError) {
       c.status(e.status as StatusCode);
@@ -48,6 +52,11 @@ async function tryRender<T extends object>(c: Context, template: string, promise
   return c.html(await eta.renderAsync(template, data));
 }
 
+app.get("/:user", async (c) => {
+  const { user } = c.req.param();
+  return tryRender(c, "user.eta", getGitHubUser(user));
+});
+
 app.get("/:owner/:repo", async (c) => {
   const { owner, repo } = c.req.param();
   return tryRender(c, "repo.eta", getGitHubRepo(owner, repo));
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index a4b834e..4a01853 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -9,6 +9,7 @@ import {
   getGitHubReleases,
   getGitHubRepo,
   getGitHubTree,
+  getGitHubUser,
   GitHubHTTPError,
   RedirectError,
 } from "./scraper.ts";
@@ -264,6 +265,54 @@ describe("GitHub scraper", () => {
     });
   });
 
+  describe("user pages", () => {
+    it("should fetch user profile", async () => {
+      const data = await getGitHubUser("torvalds");
+
+      assert.strictEqual(data.login, "torvalds");
+      assert.strictEqual(data.name, "Linus Torvalds");
+      assert.strictEqual(data.company, "Linux Foundation");
+      assert.strictEqual(data.location, "Portland, OR");
+      assert.ok(data.avatarUrl?.startsWith("https://avatars.githubusercontent.com/"));
+      assert.ok(data.followers);
+      assert.ok(data.following !== null);
+
+      assert.ok(data.pinned.length > 0);
+      const linux = data.pinned.find((repo) => repo.name === "linux");
+      assert.ok(linux);
+      assert.strictEqual(linux.owner, "torvalds");
+      assert.strictEqual(linux.description, "Linux kernel source tree");
+      assert.strictEqual(linux.language, "C");
+      assert.ok(linux.stars);
+      assert.ok(linux.forks);
+    });
+
+    it("should fetch user bio and links", async () => {
+      const data = await getGitHubUser("sindresorhus");
+
+      assert.strictEqual(data.login, "sindresorhus");
+      assert.ok(data.bio && data.bio.length > 0);
+      assert.ok(data.links.length > 0);
+      assert.ok(data.links.every((link) => link.href.startsWith("https://")));
+    });
+
+    it("should redirect to GitHub for organizations", async () => {
+      await assert.rejects(getGitHubUser("github"), (err) => {
+        assert(err instanceof RedirectError, "error should be a RedirectError");
+        assert.strictEqual(err.location, "https://github.com/github");
+        return true;
+      });
+    });
+
+    it("should throw for non-existent user", async () => {
+      await assert.rejects(getGitHubUser("nosuchuserzzzz"), (err) => {
+        assert(err instanceof GitHubHTTPError, "error should be an HTTPError");
+        assert.strictEqual(err.status, 404);
+        return true;
+      });
+    });
+  });
+
   describe("error handling", () => {
     it("should throw for non-existent repository", async () => {
       await assert.rejects(getGitHubRepo("nosuchowner", "nosuchrepo"), (err) => {
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index f32a8da..a44f0a7 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -1,5 +1,5 @@
 import { parseDocument } from "htmlparser2";
-import { type Element } from "domhandler";
+import { type AnyNode, type Element } from "domhandler";
 import { getAttributeValue, getInnerHTML, textContent } from "domutils";
 import * as cssSelect from "css-select";
 
@@ -293,6 +293,28 @@ export interface GitHubRelease extends GitHubCommon {
   release: Release;
 }
 
+interface UserPinnedRepo {
+  owner: string;
+  name: string;
+  description?: string;
+  language?: string;
+  stars?: string;
+  forks?: string;
+}
+
+export interface GitHubUser {
+  login: string;
+  name: string | null;
+  bio: string | null;
+  avatarUrl: string | null;
+  followers: string | null;
+  following: string | null;
+  company: string | null;
+  location: string | null;
+  links: { text: string; href: string }[];
+  pinned: UserPinnedRepo[];
+}
+
 async function fetchGitHubPage(path: string): Promise<string> {
   // GitHub throttles/blocks requests without realistic browser headers.
   // These headers make the request appear as a standard browser visit.
@@ -744,3 +766,75 @@ export async function getGitHubRelease(
   base.release.assets = assets;
   return base;
 }
+
+function selectText(selector: string, context: AnyNode): string | null {
+  const elem = cssSelect.selectOne(selector, context);
+  if (!elem) return null;
+  return textContent(elem).replace(/\s+/g, " ").trim() || null;
+}
+
+export async function getGitHubUser(user: string): Promise<GitHubUser> {
+  const html = await fetchGitHubPage(user);
+  const document = parseDocument(html);
+
+  // Organizations and other non-profile pages lack the user vcard.
+  const login = selectText(".p-nickname.vcard-username", document);
+  if (!login) {
+    throw new RedirectError(`https://github.com/${user}`);
+  }
+
+  const avatarElem = cssSelect.selectOne("a[itemprop='image'] img", document) as Element | null;
+  const avatarUrl = avatarElem ? getAttributeValue(avatarElem, "src") || null : null;
+
+  const followers =
+    selectText("a[href$='tab=followers']", document)?.replace(/ followers?$/, "") ?? null;
+  const following =
+    selectText("a[href$='tab=following']", document)?.replace(/ following$/, "") ?? null;
+
+  const links: { text: string; href: string }[] = [];
+  const linkItems = cssSelect.selectAll(
+    "li[itemprop='url'], li[itemprop='social']",
+    document,
+  ) as unknown as Element[];
+  for (const item of linkItems) {
+    const anchor = cssSelect.selectOne("a[href]", item) as Element | null;
+    const href = anchor ? getAttributeValue(anchor, "href") : undefined;
+    if (!anchor || !href) continue;
+    links.push({ text: textContent(anchor).replace(/\s+/g, " ").trim(), href });
+  }
+
+  const pinned: UserPinnedRepo[] = [];
+  const pinnedItems = cssSelect.selectAll(
+    ".pinned-item-list-item",
+    document,
+  ) as unknown as Element[];
+  for (const item of pinnedItems) {
+    const link = cssSelect.selectOne("a:has(span.repo)", item) as Element | null;
+    const href = link ? getAttributeValue(link, "href") : undefined;
+    if (!href) continue;
+    const [owner, name] = href.replace(/^\//, "").split("/");
+    if (!owner || !name) continue;
+
+    pinned.push({
+      owner,
+      name,
+      description: selectText("p.pinned-item-desc", item) ?? undefined,
+      language: selectText("[itemprop='programmingLanguage']", item) ?? undefined,
+      stars: selectText("a[href$='/stargazers']", item) ?? undefined,
+      forks: selectText("a[href$='/forks']", item) ?? undefined,
+    });
+  }
+
+  return {
+    login,
+    name: selectText(".p-name.vcard-fullname", document),
+    bio: selectText(".p-note.user-profile-bio", document),
+    avatarUrl,
+    followers,
+    following,
+    company: selectText("li[itemprop='worksFor']", document),
+    location: selectText("li[itemprop='homeLocation']", document),
+    links,
+    pinned,
+  };
+}
diff --git a/cgithub/static/style.css b/cgithub/static/style.css
index 86a67d8..e698993 100644
--- a/cgithub/static/style.css
+++ b/cgithub/static/style.css
@@ -148,6 +148,11 @@ body > footer {
   font-weight: bold;
 }
 
+.avatar {
+  float: right;
+  clip-path: circle();
+}
+
 .branch {
   display: inline-block;
   margin-right: 1rem;
diff --git a/cgithub/views/home.eta b/cgithub/views/home.eta
index adb71f2..0026ea3 100644
--- a/cgithub/views/home.eta
+++ b/cgithub/views/home.eta
@@ -12,7 +12,7 @@
   </header>
 
   <main>
-    <p>Browse any GitHub repository by navigating to <code>/:owner/:repo</code></p>
+    <p>Browse any GitHub repository by navigating to <code>/:owner/:repo</code>, or a user profile at <code>/:user</code></p>
   </main>
 
   <h2>Credits</h2>
diff --git a/cgithub/views/user.eta b/cgithub/views/user.eta
new file mode 100644
index 0000000..cf2da94
--- /dev/null
+++ b/cgithub/views/user.eta
@@ -0,0 +1,58 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+  <meta charset="UTF-8">
+  <meta name="viewport" content="width=device-width, initial-scale=1.0">
+  <title><%= it.name ? `${it.name} (${it.login})` : it.login %></title>
+  <link rel="stylesheet" href="/static/style.css">
+</head>
+<body>
+  <div class="box repo-info">
+    <% if (it.avatarUrl) { %>
+      <img class="avatar" src="<%= it.avatarUrl %>" alt="" width="64" height="64">
+    <% } %>
+    <div class="repo"><%= it.name || it.login %></div>
+    <div class="detail"><%= it.login %></div>
+    <% if (it.bio) { %>
+      <div class="description"><%= it.bio %></div>
+    <% } %>
+    <% if (it.followers) { %>
+      <div class="detail"><img src="/static/icons/eye.svg" alt="" class="icon"> <%= it.followers %> followers<%= it.following ? ` · ${it.following} following` : "" %></div>
+    <% } %>
+    <% if (it.company) { %>
+      <div class="detail"><%= it.company %></div>
+    <% } %>
+    <% if (it.location) { %>
+      <div class="detail"><%= it.location %></div>
+    <% } %>
+    <% for (const link of it.links) { %>
+      <div class="detail"><img src="/static/icons/link.svg" alt="" class="icon"> <a href="<%= link.href %>" rel="noopener noreferrer"><%= link.text %></a></div>
+    <% } %>
+  </div>
+
+  <% if (it.pinned.length > 0) { %>
+    <section>
+      <h2>Pinned</h2>
+      <% for (const repo of it.pinned) { %>
+        <div class="box repo-info">
+          <div><a href="/<%= repo.owner %>/<%= repo.name %>"><%= repo.owner === it.login ? repo.name : `${repo.owner}/${repo.name}` %></a></div>
+          <% if (repo.description) { %>
+            <div class="description"><%= repo.description %></div>
+          <% } %>
+          <% if (repo.language) { %>
+            <span class="detail"><%= repo.language %></span>
+          <% } %>
+          <% if (repo.stars) { %>
+            <span class="detail"><img src="/static/icons/star.svg" alt="stars" class="icon"> <%= repo.stars %></span>
+          <% } %>
+          <% if (repo.forks) { %>
+            <span class="detail"><img src="/static/icons/git-fork.svg" alt="forks" class="icon"> <%= repo.forks %></span>
+          <% } %>
+        </div>
+      <% } %>
+    </section>
+  <% } %>
+
+  <footer>Generated by cgithub</footer>
+</body>
+</html>
-- 
cgit v1.3.1


From 47ddd107c2f19a977798b8af230c652ba731cf45 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Wed, 15 Jul 2026 00:21:11 -0400
Subject: Remove watchers, no longer rendered by GitHub

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---
 cgithub/src/scraper.test.ts | 1 -
 cgithub/src/scraper.ts      | 5 -----
 cgithub/views/repo.eta      | 3 ---
 3 files changed, 9 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 4a01853..2438361 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -30,7 +30,6 @@ describe("GitHub scraper", () => {
       );
       assert.strictEqual(data.info.website, "https://pages.github.com");
       assert.strictEqual(data.info.stars?.length, 3);
-      assert.strictEqual(data.info.watchers?.length, 2);
       assert.strictEqual(data.info.forks?.length, 3);
       assert.strictEqual(data.info.numReleases, 38);
       assert.strictEqual(data.info.latestReleaseTag, "v5.0.0");
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index a44f0a7..3434434 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -227,7 +227,6 @@ export interface RepoInfo {
   description: string | null;
   website: string | null;
   stars: string | null;
-  watchers: string | null;
   forks: string | null;
   numReleases?: number;
   latestReleaseTag?: string;
@@ -386,9 +385,6 @@ function parseRepoInfo(html: string): RepoInfo {
     ? getAttributeValue(starsElem, "title") || textContent(starsElem).trim()
     : null;
 
-  const watchersElem = cssSelect.selectOne("a[href$='/watchers'] strong", document);
-  const watchers = watchersElem ? textContent(watchersElem).trim() : null;
-
   const forksElem = cssSelect.selectOne("#repo-network-counter", document) as Element | null;
   const forks = forksElem
     ? getAttributeValue(forksElem, "title") || textContent(forksElem).trim()
@@ -428,7 +424,6 @@ function parseRepoInfo(html: string): RepoInfo {
     description,
     website,
     stars,
-    watchers,
     forks,
     numReleases,
     latestReleaseTag,
diff --git a/cgithub/views/repo.eta b/cgithub/views/repo.eta
index 6204b69..683a9ad 100644
--- a/cgithub/views/repo.eta
+++ b/cgithub/views/repo.eta
@@ -11,9 +11,6 @@
     <% if (it.info.stars) { %>
       <span class="detail"><img src="/static/icons/star.svg" alt="stars" class="icon"> <%= it.info.stars %></span>
     <% } %>
-    <% if (it.info.watchers) { %>
-      <span class="detail"><img src="/static/icons/eye.svg" alt="watchers" class="icon"> <%= it.info.watchers %></span>
-    <% } %>
     <% if (it.info.forks) { %>
       <span class="detail"><img src="/static/icons/git-fork.svg" alt="forks" class="icon"> <%= it.info.forks %></span>
     <% } %>
-- 
cgit v1.3.1


From 7f38fbb0e3c852fda02986f3a755557938810062 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Wed, 15 Jul 2026 01:18:50 -0400
Subject: Add organization support to owner pages

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---
 cgithub/CLAUDE.md           |   2 +-
 cgithub/src/index.ts        |   8 +--
 cgithub/src/scraper.test.ts |  37 ++++++++++----
 cgithub/src/scraper.ts      | 121 ++++++++++++++++++++++++++++++++------------
 cgithub/views/home.eta      |   2 +-
 cgithub/views/owner.eta     |  58 +++++++++++++++++++++
 cgithub/views/user.eta      |  58 ---------------------
 7 files changed, 180 insertions(+), 106 deletions(-)
 create mode 100644 cgithub/views/owner.eta
 delete mode 100644 cgithub/views/user.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/CLAUDE.md b/cgithub/CLAUDE.md
index f6d8524..827cc3b 100644
--- a/cgithub/CLAUDE.md
+++ b/cgithub/CLAUDE.md
@@ -26,7 +26,7 @@ pnpm test
 **Templates:** `views/*.eta` - Eta templates with layout inheritance (`layout.eta` as base)
 
 **Routes:**
-- `GET /:user` - User profile with pinned repositories
+- `GET /:owner` - User or organization profile with pinned repositories
 - `GET /:owner/:repo` - Repository info with directory listing
 - `GET /:owner/:repo/tree/:branch/:path*` - Directory listing
 - `GET /:owner/:repo/blob/:branch/:path` - File content view
diff --git a/cgithub/src/index.ts b/cgithub/src/index.ts
index d8bdfec..f676bd6 100644
--- a/cgithub/src/index.ts
+++ b/cgithub/src/index.ts
@@ -9,12 +9,12 @@ import {
   getGitHubBlob,
   getGitHubCommits,
   getGitHubIssues,
+  getGitHubOwner,
   getGitHubPulls,
   getGitHubRelease,
   getGitHubReleases,
   getGitHubRepo,
   getGitHubTree,
-  getGitHubUser,
   GitHubHTTPError,
   RedirectError,
 } from "./scraper.ts";
@@ -52,9 +52,9 @@ async function tryRender<T extends object>(c: Context, template: string, promise
   return c.html(await eta.renderAsync(template, data));
 }
 
-app.get("/:user", async (c) => {
-  const { user } = c.req.param();
-  return tryRender(c, "user.eta", getGitHubUser(user));
+app.get("/:owner", async (c) => {
+  const { owner } = c.req.param();
+  return tryRender(c, "owner.eta", getGitHubOwner(owner));
 });
 
 app.get("/:owner/:repo", async (c) => {
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 2438361..3048915 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -8,8 +8,8 @@ import {
   getGitHubRelease,
   getGitHubReleases,
   getGitHubRepo,
+  getGitHubOwner,
   getGitHubTree,
-  getGitHubUser,
   GitHubHTTPError,
   RedirectError,
 } from "./scraper.ts";
@@ -264,10 +264,11 @@ describe("GitHub scraper", () => {
     });
   });
 
-  describe("user pages", () => {
+  describe("owner pages", () => {
     it("should fetch user profile", async () => {
-      const data = await getGitHubUser("torvalds");
+      const data = await getGitHubOwner("torvalds");
 
+      assert.strictEqual(data.type, "user");
       assert.strictEqual(data.login, "torvalds");
       assert.strictEqual(data.name, "Linus Torvalds");
       assert.strictEqual(data.company, "Linux Foundation");
@@ -287,7 +288,7 @@ describe("GitHub scraper", () => {
     });
 
     it("should fetch user bio and links", async () => {
-      const data = await getGitHubUser("sindresorhus");
+      const data = await getGitHubOwner("sindresorhus");
 
       assert.strictEqual(data.login, "sindresorhus");
       assert.ok(data.bio && data.bio.length > 0);
@@ -295,16 +296,34 @@ describe("GitHub scraper", () => {
       assert.ok(data.links.every((link) => link.href.startsWith("https://")));
     });
 
-    it("should redirect to GitHub for organizations", async () => {
-      await assert.rejects(getGitHubUser("github"), (err) => {
+    it("should fetch organization profile", async () => {
+      const data = await getGitHubOwner("github");
+
+      assert.strictEqual(data.type, "organization");
+      assert.strictEqual(data.login, "github");
+      assert.strictEqual(data.name, "GitHub");
+      assert.strictEqual(data.bio, "How people build software.");
+      assert.strictEqual(data.location, "United States of America");
+      assert.ok(data.avatarUrl?.startsWith("https://avatars.githubusercontent.com/"));
+      assert.ok(data.followers);
+      assert.strictEqual(data.following, null);
+      assert.ok(data.links.some((link) => link.href === "https://github.com/about"));
+
+      assert.ok(data.pinned.length > 0);
+      assert.ok(data.pinned.every((repo) => repo.owner === "github"));
+      assert.ok(data.pinned.some((repo) => repo.stars && repo.forks));
+    });
+
+    it("should redirect to GitHub for non-profile pages", async () => {
+      await assert.rejects(getGitHubOwner("features"), (err) => {
         assert(err instanceof RedirectError, "error should be a RedirectError");
-        assert.strictEqual(err.location, "https://github.com/github");
+        assert.strictEqual(err.location, "https://github.com/features");
         return true;
       });
     });
 
-    it("should throw for non-existent user", async () => {
-      await assert.rejects(getGitHubUser("nosuchuserzzzz"), (err) => {
+    it("should throw for non-existent owner", async () => {
+      await assert.rejects(getGitHubOwner("nosuchuserzzzz"), (err) => {
         assert(err instanceof GitHubHTTPError, "error should be an HTTPError");
         assert.strictEqual(err.status, 404);
         return true;
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 3434434..3e3ad75 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -292,7 +292,7 @@ export interface GitHubRelease extends GitHubCommon {
   release: Release;
 }
 
-interface UserPinnedRepo {
+interface OwnerPinnedRepo {
   owner: string;
   name: string;
   description?: string;
@@ -301,17 +301,19 @@ interface UserPinnedRepo {
   forks?: string;
 }
 
-export interface GitHubUser {
+export interface GitHubOwner {
+  type: "user" | "organization";
   login: string;
   name: string | null;
   bio: string | null;
   avatarUrl: string | null;
   followers: string | null;
+  // Users only; always null for organizations.
   following: string | null;
   company: string | null;
   location: string | null;
   links: { text: string; href: string }[];
-  pinned: UserPinnedRepo[];
+  pinned: OwnerPinnedRepo[];
 }
 
 async function fetchGitHubPage(path: string): Promise<string> {
@@ -768,16 +770,37 @@ function selectText(selector: string, context: AnyNode): string | null {
   return textContent(elem).replace(/\s+/g, " ").trim() || null;
 }
 
-export async function getGitHubUser(user: string): Promise<GitHubUser> {
-  const html = await fetchGitHubPage(user);
-  const document = parseDocument(html);
+// Pinned repositories appear as .pinned-item-list-item on user pages and as
+// [itemprop='owns'] blocks on organization pages.
+function parsePinnedRepos(document: AnyNode): OwnerPinnedRepo[] {
+  const pinned: OwnerPinnedRepo[] = [];
+  const items = cssSelect.selectAll(
+    ".pinned-item-list-item, [itemprop='owns']",
+    document,
+  ) as unknown as Element[];
+  for (const item of items) {
+    const link = cssSelect.selectOne(
+      "a:has(span.repo), a[itemprop~='codeRepository']",
+      item,
+    ) as Element | null;
+    const href = link ? getAttributeValue(link, "href") : undefined;
+    if (!href) continue;
+    const [owner, name] = href.replace(/^\//, "").split("/");
+    if (!owner || !name) continue;
 
-  // Organizations and other non-profile pages lack the user vcard.
-  const login = selectText(".p-nickname.vcard-username", document);
-  if (!login) {
-    throw new RedirectError(`https://github.com/${user}`);
+    pinned.push({
+      owner,
+      name,
+      description: selectText("p.pinned-item-desc, [itemprop='description']", item) ?? undefined,
+      language: selectText("[itemprop='programmingLanguage']", item) ?? undefined,
+      stars: selectText("a[href$='/stargazers']", item) ?? undefined,
+      forks: selectText("a[href$='/forks']", item) ?? undefined,
+    });
   }
+  return pinned;
+}
 
+function parseUser(document: AnyNode, login: string): GitHubOwner {
   const avatarElem = cssSelect.selectOne("a[itemprop='image'] img", document) as Element | null;
   const avatarUrl = avatarElem ? getAttributeValue(avatarElem, "src") || null : null;
 
@@ -798,29 +821,8 @@ export async function getGitHubUser(user: string): Promise<GitHubUser> {
     links.push({ text: textContent(anchor).replace(/\s+/g, " ").trim(), href });
   }
 
-  const pinned: UserPinnedRepo[] = [];
-  const pinnedItems = cssSelect.selectAll(
-    ".pinned-item-list-item",
-    document,
-  ) as unknown as Element[];
-  for (const item of pinnedItems) {
-    const link = cssSelect.selectOne("a:has(span.repo)", item) as Element | null;
-    const href = link ? getAttributeValue(link, "href") : undefined;
-    if (!href) continue;
-    const [owner, name] = href.replace(/^\//, "").split("/");
-    if (!owner || !name) continue;
-
-    pinned.push({
-      owner,
-      name,
-      description: selectText("p.pinned-item-desc", item) ?? undefined,
-      language: selectText("[itemprop='programmingLanguage']", item) ?? undefined,
-      stars: selectText("a[href$='/stargazers']", item) ?? undefined,
-      forks: selectText("a[href$='/forks']", item) ?? undefined,
-    });
-  }
-
   return {
+    type: "user",
     login,
     name: selectText(".p-name.vcard-fullname", document),
     bio: selectText(".p-note.user-profile-bio", document),
@@ -830,6 +832,59 @@ export async function getGitHubUser(user: string): Promise<GitHubUser> {
     company: selectText("li[itemprop='worksFor']", document),
     location: selectText("li[itemprop='homeLocation']", document),
     links,
-    pinned,
+    pinned: parsePinnedRepos(document),
   };
 }
+
+function parseOrganization(document: AnyNode, owner: string): GitHubOwner {
+  const avatarElem = cssSelect.selectOne(".orghead img[itemprop='image']", document) as
+    | Element
+    | null;
+  const avatarUrl = avatarElem ? getAttributeValue(avatarElem, "src") || null : null;
+  // The login only appears in the avatar's "@login" alt text.
+  const alt = avatarElem ? getAttributeValue(avatarElem, "alt") : undefined;
+  const login = alt?.replace(/^@/, "") || owner;
+
+  const followers =
+    selectText("a[href$='/followers']", document)?.replace(/ followers?$/, "") ?? null;
+
+  const links: { text: string; href: string }[] = [];
+  const anchors = cssSelect.selectAll(
+    ".orghead a[itemprop='url'], .orghead [itemprop='email'] a",
+    document,
+  ) as unknown as Element[];
+  for (const anchor of anchors) {
+    const href = getAttributeValue(anchor, "href");
+    if (!href) continue;
+    links.push({ text: textContent(anchor).replace(/\s+/g, " ").trim(), href });
+  }
+
+  return {
+    type: "organization",
+    login,
+    name: selectText(".orghead h1", document),
+    bio: selectText(".orghead h1 + div", document),
+    avatarUrl,
+    followers,
+    following: null,
+    company: null,
+    location: selectText(".orghead [itemprop='location']", document),
+    links,
+    pinned: parsePinnedRepos(document),
+  };
+}
+
+export async function getGitHubOwner(owner: string): Promise<GitHubOwner> {
+  const html = await fetchGitHubPage(owner);
+  const document = parseDocument(html);
+
+  const login = selectText(".p-nickname.vcard-username", document);
+  if (login) {
+    return parseUser(document, login);
+  }
+  if (cssSelect.selectOne(".orghead", document)) {
+    return parseOrganization(document, owner);
+  }
+  // Not a profile page (marketing pages like github.com/features, etc.).
+  throw new RedirectError(`https://github.com/${owner}`);
+}
diff --git a/cgithub/views/home.eta b/cgithub/views/home.eta
index 0026ea3..604e4d0 100644
--- a/cgithub/views/home.eta
+++ b/cgithub/views/home.eta
@@ -12,7 +12,7 @@
   </header>
 
   <main>
-    <p>Browse any GitHub repository by navigating to <code>/:owner/:repo</code>, or a user profile at <code>/:user</code></p>
+    <p>Browse any GitHub repository by navigating to <code>/:owner/:repo</code>, or a user or organization profile at <code>/:owner</code></p>
   </main>
 
   <h2>Credits</h2>
diff --git a/cgithub/views/owner.eta b/cgithub/views/owner.eta
new file mode 100644
index 0000000..4426400
--- /dev/null
+++ b/cgithub/views/owner.eta
@@ -0,0 +1,58 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+  <meta charset="UTF-8">
+  <meta name="viewport" content="width=device-width, initial-scale=1.0">
+  <title><%= it.name ? `${it.name} (${it.login})` : it.login %></title>
+  <link rel="stylesheet" href="/static/style.css">
+</head>
+<body>
+  <div class="box repo-info">
+    <% if (it.avatarUrl) { %>
+      <img class="avatar" src="<%= it.avatarUrl %>" alt="" width="64" height="64">
+    <% } %>
+    <div class="repo"><%= it.name || it.login %></div>
+    <div class="detail"><%= it.login %><%= it.type === "organization" ? " (organization)" : "" %></div>
+    <% if (it.bio) { %>
+      <div class="description"><%= it.bio %></div>
+    <% } %>
+    <% if (it.followers) { %>
+      <div class="detail"><img src="/static/icons/eye.svg" alt="" class="icon"> <%= it.followers %> followers<%= it.following ? ` · ${it.following} following` : "" %></div>
+    <% } %>
+    <% if (it.company) { %>
+      <div class="detail"><%= it.company %></div>
+    <% } %>
+    <% if (it.location) { %>
+      <div class="detail"><%= it.location %></div>
+    <% } %>
+    <% for (const link of it.links) { %>
+      <div class="detail"><img src="/static/icons/link.svg" alt="" class="icon"> <a href="<%= link.href %>" rel="noopener noreferrer"><%= link.text %></a></div>
+    <% } %>
+  </div>
+
+  <% if (it.pinned.length > 0) { %>
+    <section>
+      <h2>Pinned</h2>
+      <% for (const repo of it.pinned) { %>
+        <div class="box repo-info">
+          <div><a href="/<%= repo.owner %>/<%= repo.name %>"><%= repo.owner === it.login ? repo.name : `${repo.owner}/${repo.name}` %></a></div>
+          <% if (repo.description) { %>
+            <div class="description"><%= repo.description %></div>
+          <% } %>
+          <% if (repo.language) { %>
+            <span class="detail"><%= repo.language %></span>
+          <% } %>
+          <% if (repo.stars) { %>
+            <span class="detail"><img src="/static/icons/star.svg" alt="stars" class="icon"> <%= repo.stars %></span>
+          <% } %>
+          <% if (repo.forks) { %>
+            <span class="detail"><img src="/static/icons/git-fork.svg" alt="forks" class="icon"> <%= repo.forks %></span>
+          <% } %>
+        </div>
+      <% } %>
+    </section>
+  <% } %>
+
+  <footer>Generated by cgithub</footer>
+</body>
+</html>
diff --git a/cgithub/views/user.eta b/cgithub/views/user.eta
deleted file mode 100644
index cf2da94..0000000
--- a/cgithub/views/user.eta
+++ /dev/null
@@ -1,58 +0,0 @@
-<!DOCTYPE html>
-<html lang="en">
-<head>
-  <meta charset="UTF-8">
-  <meta name="viewport" content="width=device-width, initial-scale=1.0">
-  <title><%= it.name ? `${it.name} (${it.login})` : it.login %></title>
-  <link rel="stylesheet" href="/static/style.css">
-</head>
-<body>
-  <div class="box repo-info">
-    <% if (it.avatarUrl) { %>
-      <img class="avatar" src="<%= it.avatarUrl %>" alt="" width="64" height="64">
-    <% } %>
-    <div class="repo"><%= it.name || it.login %></div>
-    <div class="detail"><%= it.login %></div>
-    <% if (it.bio) { %>
-      <div class="description"><%= it.bio %></div>
-    <% } %>
-    <% if (it.followers) { %>
-      <div class="detail"><img src="/static/icons/eye.svg" alt="" class="icon"> <%= it.followers %> followers<%= it.following ? ` · ${it.following} following` : "" %></div>
-    <% } %>
-    <% if (it.company) { %>
-      <div class="detail"><%= it.company %></div>
-    <% } %>
-    <% if (it.location) { %>
-      <div class="detail"><%= it.location %></div>
-    <% } %>
-    <% for (const link of it.links) { %>
-      <div class="detail"><img src="/static/icons/link.svg" alt="" class="icon"> <a href="<%= link.href %>" rel="noopener noreferrer"><%= link.text %></a></div>
-    <% } %>
-  </div>
-
-  <% if (it.pinned.length > 0) { %>
-    <section>
-      <h2>Pinned</h2>
-      <% for (const repo of it.pinned) { %>
-        <div class="box repo-info">
-          <div><a href="/<%= repo.owner %>/<%= repo.name %>"><%= repo.owner === it.login ? repo.name : `${repo.owner}/${repo.name}` %></a></div>
-          <% if (repo.description) { %>
-            <div class="description"><%= repo.description %></div>
-          <% } %>
-          <% if (repo.language) { %>
-            <span class="detail"><%= repo.language %></span>
-          <% } %>
-          <% if (repo.stars) { %>
-            <span class="detail"><img src="/static/icons/star.svg" alt="stars" class="icon"> <%= repo.stars %></span>
-          <% } %>
-          <% if (repo.forks) { %>
-            <span class="detail"><img src="/static/icons/git-fork.svg" alt="forks" class="icon"> <%= repo.forks %></span>
-          <% } %>
-        </div>
-      <% } %>
-    </section>
-  <% } %>
-
-  <footer>Generated by cgithub</footer>
-</body>
-</html>
-- 
cgit v1.3.1


From 59f587d9ffac3d20486ec6773aedebb54012cc32 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Wed, 15 Jul 2026 01:36:33 -0400
Subject: Owner page nav menu and layout restructure

Add a menu bar of profile tab links, split the org repository list from
pinned repos, and restructure the owner template to match repo pages.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---
 cgithub/src/scraper.test.ts  | 22 +++++++++++++++
 cgithub/src/scraper.ts       | 57 ++++++++++++++++++++++++++++++--------
 cgithub/views/_ownerRepo.eta | 15 ++++++++++
 cgithub/views/owner.eta      | 65 ++++++++++++++++++++++++--------------------
 4 files changed, 118 insertions(+), 41 deletions(-)
 create mode 100644 cgithub/views/_ownerRepo.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 3048915..d6991d7 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -285,6 +285,12 @@ describe("GitHub scraper", () => {
       assert.strictEqual(linux.language, "C");
       assert.ok(linux.stars);
       assert.ok(linux.forks);
+
+      assert.deepStrictEqual(data.repos, []);
+      const navTexts = data.nav.map((item) => item.text);
+      assert.deepStrictEqual(navTexts, ["Repositories", "Projects", "Packages"]);
+      const reposTab = data.nav.find((item) => item.text === "Repositories");
+      assert.strictEqual(reposTab?.href, "/torvalds?tab=repositories");
     });
 
     it("should fetch user bio and links", async () => {
@@ -294,6 +300,7 @@ describe("GitHub scraper", () => {
       assert.ok(data.bio && data.bio.length > 0);
       assert.ok(data.links.length > 0);
       assert.ok(data.links.every((link) => link.href.startsWith("https://")));
+      assert.ok(data.nav.some((item) => item.text === "Sponsoring"));
     });
 
     it("should fetch organization profile", async () => {
@@ -310,8 +317,23 @@ describe("GitHub scraper", () => {
       assert.ok(data.links.some((link) => link.href === "https://github.com/about"));
 
       assert.ok(data.pinned.length > 0);
+      assert.ok(data.pinned.length <= 6, "pinned should not include the repository list");
       assert.ok(data.pinned.every((repo) => repo.owner === "github"));
       assert.ok(data.pinned.some((repo) => repo.stars && repo.forks));
+
+      assert.ok(data.repos.length > 0);
+      assert.ok(data.repos.every((repo) => repo.owner === "github"));
+
+      const navTexts = data.nav.map((item) => item.text);
+      assert.deepStrictEqual(navTexts, [
+        "Repositories",
+        "Projects",
+        "Packages",
+        "People",
+        "Sponsoring",
+      ]);
+      const reposTab = data.nav.find((item) => item.text === "Repositories");
+      assert.strictEqual(reposTab?.href, "/orgs/github/repositories");
     });
 
     it("should redirect to GitHub for non-profile pages", async () => {
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 3e3ad75..760a306 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -292,7 +292,7 @@ export interface GitHubRelease extends GitHubCommon {
   release: Release;
 }
 
-interface OwnerPinnedRepo {
+interface OwnerRepo {
   owner: string;
   name: string;
   description?: string;
@@ -301,6 +301,11 @@ interface OwnerPinnedRepo {
   forks?: string;
 }
 
+interface OwnerNavItem {
+  text: string;
+  href: string;
+}
+
 export interface GitHubOwner {
   type: "user" | "organization";
   login: string;
@@ -313,7 +318,10 @@ export interface GitHubOwner {
   company: string | null;
   location: string | null;
   links: { text: string; href: string }[];
-  pinned: OwnerPinnedRepo[];
+  nav: OwnerNavItem[];
+  pinned: OwnerRepo[];
+  // The repository list on organization overviews; empty for users.
+  repos: OwnerRepo[];
 }
 
 async function fetchGitHubPage(path: string): Promise<string> {
@@ -770,14 +778,8 @@ function selectText(selector: string, context: AnyNode): string | null {
   return textContent(elem).replace(/\s+/g, " ").trim() || null;
 }
 
-// Pinned repositories appear as .pinned-item-list-item on user pages and as
-// [itemprop='owns'] blocks on organization pages.
-function parsePinnedRepos(document: AnyNode): OwnerPinnedRepo[] {
-  const pinned: OwnerPinnedRepo[] = [];
-  const items = cssSelect.selectAll(
-    ".pinned-item-list-item, [itemprop='owns']",
-    document,
-  ) as unknown as Element[];
+function parseRepoItems(items: Element[]): OwnerRepo[] {
+  const repos: OwnerRepo[] = [];
   for (const item of items) {
     const link = cssSelect.selectOne(
       "a:has(span.repo), a[itemprop~='codeRepository']",
@@ -788,7 +790,7 @@ function parsePinnedRepos(document: AnyNode): OwnerPinnedRepo[] {
     const [owner, name] = href.replace(/^\//, "").split("/");
     if (!owner || !name) continue;
 
-    pinned.push({
+    repos.push({
       owner,
       name,
       description: selectText("p.pinned-item-desc, [itemprop='description']", item) ?? undefined,
@@ -797,7 +799,32 @@ function parsePinnedRepos(document: AnyNode): OwnerPinnedRepo[] {
       forks: selectText("a[href$='/forks']", item) ?? undefined,
     });
   }
-  return pinned;
+  return repos;
+}
+
+function parsePinnedRepos(document: AnyNode): OwnerRepo[] {
+  return parseRepoItems(
+    cssSelect.selectAll(".pinned-item-list-item", document) as unknown as Element[],
+  );
+}
+
+// The tabs we surface from the profile page's own navigation. Stars and
+// Overview are deliberately excluded.
+const OWNER_NAV_TABS = ["Repositories", "Projects", "Packages", "People", "Sponsoring"];
+
+function parseOwnerNav(document: AnyNode): OwnerNavItem[] {
+  const nav: OwnerNavItem[] = [];
+  const seen = new Set<string>();
+  const anchors = cssSelect.selectAll("a.UnderlineNav-item", document) as unknown as Element[];
+  for (const anchor of anchors) {
+    const href = getAttributeValue(anchor, "href");
+    // Label text is e.g. "Repositories 12" including the counter.
+    const text = textContent(anchor).replace(/\s+/g, " ").trim().split(" ")[0];
+    if (!href || seen.has(href) || !OWNER_NAV_TABS.includes(text)) continue;
+    seen.add(href);
+    nav.push({ text, href });
+  }
+  return nav;
 }
 
 function parseUser(document: AnyNode, login: string): GitHubOwner {
@@ -832,7 +859,9 @@ function parseUser(document: AnyNode, login: string): GitHubOwner {
     company: selectText("li[itemprop='worksFor']", document),
     location: selectText("li[itemprop='homeLocation']", document),
     links,
+    nav: parseOwnerNav(document),
     pinned: parsePinnedRepos(document),
+    repos: [],
   };
 }
 
@@ -870,7 +899,11 @@ function parseOrganization(document: AnyNode, owner: string): GitHubOwner {
     company: null,
     location: selectText(".orghead [itemprop='location']", document),
     links,
+    nav: parseOwnerNav(document),
     pinned: parsePinnedRepos(document),
+    repos: parseRepoItems(
+      cssSelect.selectAll("#org-repositories [itemprop='owns']", document) as unknown as Element[],
+    ),
   };
 }
 
diff --git a/cgithub/views/_ownerRepo.eta b/cgithub/views/_ownerRepo.eta
new file mode 100644
index 0000000..a857aff
--- /dev/null
+++ b/cgithub/views/_ownerRepo.eta
@@ -0,0 +1,15 @@
+<div class="box repo-info">
+  <div><a href="/<%= it.repo.owner %>/<%= it.repo.name %>"><%= it.repo.owner === it.login ? it.repo.name : `${it.repo.owner}/${it.repo.name}` %></a></div>
+  <% if (it.repo.description) { %>
+    <div class="description"><%= it.repo.description %></div>
+  <% } %>
+  <% if (it.repo.language) { %>
+    <span class="detail"><%= it.repo.language %></span>
+  <% } %>
+  <% if (it.repo.stars) { %>
+    <span class="detail"><img src="/static/icons/star.svg" alt="stars" class="icon"> <%= it.repo.stars %></span>
+  <% } %>
+  <% if (it.repo.forks) { %>
+    <span class="detail"><img src="/static/icons/git-fork.svg" alt="forks" class="icon"> <%= it.repo.forks %></span>
+  <% } %>
+</div>
diff --git a/cgithub/views/owner.eta b/cgithub/views/owner.eta
index 4426400..e5667fb 100644
--- a/cgithub/views/owner.eta
+++ b/cgithub/views/owner.eta
@@ -7,48 +7,55 @@
   <link rel="stylesheet" href="/static/style.css">
 </head>
 <body>
-  <div class="box repo-info">
+  <div class="box">
     <% if (it.avatarUrl) { %>
       <img class="avatar" src="<%= it.avatarUrl %>" alt="" width="64" height="64">
     <% } %>
     <div class="repo"><%= it.name || it.login %></div>
-    <div class="detail"><%= it.login %><%= it.type === "organization" ? " (organization)" : "" %></div>
-    <% if (it.bio) { %>
-      <div class="description"><%= it.bio %></div>
-    <% } %>
-    <% if (it.followers) { %>
-      <div class="detail"><img src="/static/icons/eye.svg" alt="" class="icon"> <%= it.followers %> followers<%= it.following ? ` · ${it.following} following` : "" %></div>
-    <% } %>
-    <% if (it.company) { %>
-      <div class="detail"><%= it.company %></div>
-    <% } %>
+    <div><%= it.login %><%= it.type === "organization" ? " (organization)" : "" %></div>
     <% if (it.location) { %>
-      <div class="detail"><%= it.location %></div>
+      <div><%= it.location %></div>
     <% } %>
-    <% for (const link of it.links) { %>
-      <div class="detail"><img src="/static/icons/link.svg" alt="" class="icon"> <a href="<%= link.href %>" rel="noopener noreferrer"><%= link.text %></a></div>
+    <% if (it.nav.length > 0) { %>
+      <nav>
+        <% for (const item of it.nav) { %>
+          <a href="<%= item.href %>"><%= item.text %></a>
+        <% } %>
+      </nav>
     <% } %>
   </div>
 
+  <% if (it.bio || it.links.length > 0 || it.followers || it.company) { %>
+    <div class="box repo-info">
+      <% if (it.bio) { %>
+        <div class="description"><%= it.bio %></div>
+      <% } %>
+      <% for (const link of it.links) { %>
+        <div class="detail"><img src="/static/icons/link.svg" alt="" class="icon"> <a href="<%= link.href %>" rel="noopener noreferrer"><%= link.text %></a></div>
+      <% } %>
+      <% if (it.followers) { %>
+        <div class="detail"><img src="/static/icons/eye.svg" alt="" class="icon"> <%= it.followers %> followers<%= it.following ? ` · ${it.following} following` : "" %></div>
+      <% } %>
+      <% if (it.company) { %>
+        <div class="detail"><%= it.company %></div>
+      <% } %>
+    </div>
+  <% } %>
+
   <% if (it.pinned.length > 0) { %>
     <section>
       <h2>Pinned</h2>
       <% for (const repo of it.pinned) { %>
-        <div class="box repo-info">
-          <div><a href="/<%= repo.owner %>/<%= repo.name %>"><%= repo.owner === it.login ? repo.name : `${repo.owner}/${repo.name}` %></a></div>
-          <% if (repo.description) { %>
-            <div class="description"><%= repo.description %></div>
-          <% } %>
-          <% if (repo.language) { %>
-            <span class="detail"><%= repo.language %></span>
-          <% } %>
-          <% if (repo.stars) { %>
-            <span class="detail"><img src="/static/icons/star.svg" alt="stars" class="icon"> <%= repo.stars %></span>
-          <% } %>
-          <% if (repo.forks) { %>
-            <span class="detail"><img src="/static/icons/git-fork.svg" alt="forks" class="icon"> <%= repo.forks %></span>
-          <% } %>
-        </div>
+        <%~ include("./_ownerRepo", { repo, login: it.login }) %>
+      <% } %>
+    </section>
+  <% } %>
+
+  <% if (it.repos.length > 0) { %>
+    <section>
+      <h2>Repositories</h2>
+      <% for (const repo of it.repos) { %>
+        <%~ include("./_ownerRepo", { repo, login: it.login }) %>
       <% } %>
     </section>
   <% } %>
-- 
cgit v1.3.1


From 9fbb5f0d68a131b55bc172d46b75705f2842af04 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Wed, 15 Jul 2026 01:46:06 -0400
Subject: Replace null overview files with HTML comments

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---
 cgithub/src/scraper.test.ts |  4 ++--
 cgithub/src/scraper.ts      |  9 +++++----
 cgithub/views/_tree.eta     | 12 ++++++++----
 3 files changed, 15 insertions(+), 10 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index d6991d7..f9b60e8 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -36,7 +36,7 @@ describe("GitHub scraper", () => {
       assert.ok(data.info.latestReleaseDate?.startsWith("2026-03-25"));
 
       assert.ok(data.items.length > 0);
-      assert.ok(Object.keys(data.overviewHtml ?? {}).includes("README.md"));
+      assert.ok(typeof data.overviewHtml?.["README.md"] === "string");
     });
 
     it("should fetch tree", async () => {
@@ -60,7 +60,7 @@ describe("GitHub scraper", () => {
         },
       ]);
 
-      assert.ok(Object.keys(data.overviewHtml ?? {}).includes("README.md"));
+      assert.ok(typeof data.overviewHtml?.["README.md"] === "string");
     });
 
     it("should fetch text blob", async () => {
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 760a306..53f2bb1 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -31,7 +31,8 @@ interface TreeItem {
 
 interface OverviewFile {
   displayName: string;
-  richText?: string;
+  // null when GitHub includes the file without rendered content.
+  richText?: string | null;
 }
 
 interface IssueNode {
@@ -220,7 +221,7 @@ interface GitHubNav extends GitHubCommon {
 
 export interface GitHubTree extends GitHubNav {
   items: TreeItem[];
-  overviewHtml?: Record<string, string>;
+  overviewHtml?: Record<string, string | null>;
 }
 
 export interface RepoInfo {
@@ -443,8 +444,8 @@ function parseRepoInfo(html: string): RepoInfo {
 
 function extractOverviewHtml(
   overviewFiles: OverviewFile[] = [],
-): Record<string, string> | undefined {
-  const result: Record<string, string> = {};
+): Record<string, string | null> | undefined {
+  const result: Record<string, string | null> = {};
   for (const file of overviewFiles) {
     if (file.richText !== undefined) {
       result[file.displayName] = file.richText;
diff --git a/cgithub/views/_tree.eta b/cgithub/views/_tree.eta
index 311762f..ed6eee9 100644
--- a/cgithub/views/_tree.eta
+++ b/cgithub/views/_tree.eta
@@ -18,9 +18,13 @@
 
 <% if (it.overviewHtml) { %>
   <% for (const [filename, html] of Object.entries(it.overviewHtml)) { %>
-    <div class="overview box">
-      <header><%= filename %></header>
-      <%~ html %>
-    </div>
+    <% if (html !== null) { %>
+      <div class="overview box">
+        <header><%= filename %></header>
+        <%~ html %>
+      </div>
+    <% } else { %>
+      <!-- No rendered content for <%= filename %> -->
+    <% } %>
   <% } %>
 <% } %>
-- 
cgit v1.3.1


From 6240cf1c6b73b7ed34ff5816b53333382c229f5c Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Tue, 21 Jul 2026 17:46:58 -0400
Subject: Add async branch/tag dropdown and latest-commit bar

Adds two lazily-fetched UI pieces to match vanilla GitHub while keeping
initial page loads plain server-rendered HTML: a branch/tag selector
(native <dialog>, CSS anchor-positioned to its button, light-dismissed
via closedby="any") and a latest-commit bar above the file listing.
Both are populated by new /api/:owner/:repo/refs/:type and
/api/:owner/:repo/latest-commit/:branch/:path routes; the latter calls
GitHub's own /latest-commit JSON endpoint (the same one the React UI
uses) rather than scraping the full commit history page.
---
 cgithub/public/static/refs.js   |  60 ++++++++++++++
 cgithub/public/static/style.css |  62 ++++++++++++++
 cgithub/src/app.ts              |  46 +++++++++++
 cgithub/src/scraper.test.ts     |  41 ++++++++++
 cgithub/src/scraper.ts          | 177 ++++++++++++++++++++++++++++++++++++++--
 cgithub/views/_latestCommit.eta |  16 ++++
 cgithub/views/_nav.eta          |  11 ++-
 cgithub/views/_refs.eta         |  17 ++++
 cgithub/views/_tree.eta         |   3 +
 cgithub/views/layout.eta        |   1 +
 10 files changed, 426 insertions(+), 8 deletions(-)
 create mode 100644 cgithub/public/static/refs.js
 create mode 100644 cgithub/views/_latestCommit.eta
 create mode 100644 cgithub/views/_refs.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/public/static/refs.js b/cgithub/public/static/refs.js
new file mode 100644
index 0000000..28b8373
--- /dev/null
+++ b/cgithub/public/static/refs.js
@@ -0,0 +1,60 @@
+// Progressive enhancement for two bits of data GitHub itself loads lazily:
+// the branch/tag dropdown and the latest-commit bar above a file listing.
+// Both degrade gracefully without JS: the button's command/commandfor still
+// opens the (empty) dialog natively; the commit placeholder just never fills
+// in.
+document.addEventListener("DOMContentLoaded", () => {
+  const dialog = document.querySelector(".ref-selector-menu");
+  if (dialog) {
+    const list = dialog.querySelector(".ref-selector-list");
+    const { owner, repo } = dialog.dataset;
+    const cache = {};
+
+    async function load(type) {
+      if (Object.prototype.hasOwnProperty.call(cache, type)) {
+        list.innerHTML = cache[type];
+        return;
+      }
+      list.textContent = "Loading…";
+      try {
+        const res = await fetch(`/api/${owner}/${repo}/refs/${type}`);
+        const html = res.ok ? await res.text() : "";
+        cache[type] = html || "Failed to load.";
+        list.innerHTML = cache[type];
+      } catch {
+        list.textContent = "Failed to load.";
+      }
+    }
+
+    let opened = false;
+    dialog.addEventListener("toggle", () => {
+      if (dialog.open && !opened) {
+        opened = true;
+        load("branches");
+      }
+    });
+
+    for (const tab of dialog.querySelectorAll("[data-ref-type]")) {
+      tab.addEventListener("click", () => {
+        for (const other of dialog.querySelectorAll("[data-ref-type]")) {
+          other.classList.toggle("active", other === tab);
+        }
+        load(tab.dataset.refType);
+      });
+    }
+  }
+
+  const commitPlaceholder = document.querySelector(".latest-commit-placeholder");
+  if (commitPlaceholder) {
+    fetch(commitPlaceholder.dataset.src)
+      .then((res) => (res.ok ? res.text() : ""))
+      .then((html) => {
+        if (html.trim()) {
+          commitPlaceholder.outerHTML = html;
+        } else {
+          commitPlaceholder.remove();
+        }
+      })
+      .catch(() => commitPlaceholder.remove());
+  }
+});
diff --git a/cgithub/public/static/style.css b/cgithub/public/static/style.css
index e698993..983cb86 100644
--- a/cgithub/public/static/style.css
+++ b/cgithub/public/static/style.css
@@ -160,6 +160,68 @@ body > footer {
   border: 1px solid var(--border-color);
 }
 
+.ref-selector-button {
+  anchor-name: --ref-selector-anchor;
+}
+
+.ref-selector-menu {
+  position: absolute;
+  position-anchor: --ref-selector-anchor;
+  position-area: bottom span-right;
+  width: 16rem;
+  max-height: 20rem;
+  overflow-y: auto;
+  overflow-wrap: break-word;
+  background: var(--body-background);
+  border: 1px solid var(--border-color);
+  padding: 0.5rem;
+
+  &::backdrop {
+    background: transparent;
+  }
+}
+
+.ref-selector-tabs {
+  display: flex;
+  margin-bottom: 0.5rem;
+
+  button {
+    flex: 1;
+    padding: 0.25rem;
+    cursor: pointer;
+    &.active {
+      font-weight: bold;
+      border-bottom: 2px solid var(--link-color);
+    }
+  }
+}
+
+.ref-selector-items {
+  list-style: none;
+  margin: 0;
+  padding: 0;
+
+  li {
+    padding: 0.2rem 0;
+  }
+}
+
+.ref-selector-badge {
+  margin-left: 0.3rem;
+  font-size: 0.75rem;
+  opacity: 0.7;
+}
+
+.ref-selector-empty,
+.ref-selector-more {
+  padding: 0.2rem 0;
+  font-size: 0.85rem;
+}
+
+.latest-commit-placeholder {
+  opacity: 0.7;
+}
+
 .repo-info {
   .description {
     margin-bottom: .2rem;
diff --git a/cgithub/src/app.ts b/cgithub/src/app.ts
index 78b59ae..ca79b1f 100644
--- a/cgithub/src/app.ts
+++ b/cgithub/src/app.ts
@@ -5,8 +5,10 @@ import {
   getGitHubBlob,
   getGitHubCommits,
   getGitHubIssues,
+  getGitHubLatestCommit,
   getGitHubOwner,
   getGitHubPulls,
+  getGitHubRefs,
   getGitHubRelease,
   getGitHubReleases,
   getGitHubRepo,
@@ -44,6 +46,29 @@ export function createApp(eta: Eta) {
     return c.html(eta.render(template, data));
   }
 
+  // For fragments fetched asynchronously by client-side JS (see public/static/refs.js):
+  // errors render as an empty body rather than the full error.eta page.
+  async function tryRenderFragment<T extends object>(
+    c: Context,
+    template: string,
+    promise: Promise<T>,
+  ) {
+    let data: T;
+    try {
+      data = await promise;
+    } catch (e) {
+      const status =
+        e instanceof GitHubHTTPError
+          ? (e.status as StatusCode)
+          : e instanceof RedirectError
+            ? 404
+            : 500;
+      c.status(status);
+      return c.body(null);
+    }
+    return c.html(eta.render(template, data));
+  }
+
   app.get("/:owner", async (c) => {
     const { owner } = c.req.param();
     return tryRender(c, "owner.eta", getGitHubOwner(owner));
@@ -112,6 +137,27 @@ export function createApp(eta: Eta) {
     return tryRender(c, "commits.eta", getGitHubCommits(owner, repo, branch, path));
   });
 
+  // Internal endpoints fetched asynchronously by client-side JS, not meant to
+  // be visited directly. Keep the site's initial page loads plain server-
+  // rendered HTML while still surfacing data (branches/tags, latest commit)
+  // that would otherwise require an extra scrape on every page view.
+  app.get("/api/:owner/:repo/refs/:type", async (c) => {
+    const { owner, repo, type } = c.req.param();
+    if (type !== "branches" && type !== "tags") {
+      return c.notFound();
+    }
+    return tryRenderFragment(c, "_refs.eta", getGitHubRefs(owner, repo, type));
+  });
+
+  app.get("/api/:owner/:repo/latest-commit/:branch/:path{.*}?", async (c) => {
+    const { owner, repo, branch, path = "" } = c.req.param();
+    return tryRenderFragment(
+      c,
+      "_latestCommit.eta",
+      getGitHubLatestCommit(owner, repo, branch, path),
+    );
+  });
+
   app.get("/:owner/:repo/releases/tag/:tag", async (c) => {
     const { owner, repo, tag } = c.req.param();
     return tryRender(c, "release.eta", getGitHubRelease(owner, repo, tag));
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index f9b60e8..56f14b0 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -4,7 +4,9 @@ import {
   getGitHubBlob,
   getGitHubCommits,
   getGitHubIssues,
+  getGitHubLatestCommit,
   getGitHubPulls,
+  getGitHubRefs,
   getGitHubRelease,
   getGitHubReleases,
   getGitHubRepo,
@@ -205,6 +207,45 @@ describe("GitHub scraper", () => {
       assert.ok(firstRelease.bodyHtml.length > 0);
     });
 
+    it("should fetch latest commit for a path", async () => {
+      const data = await getGitHubLatestCommit("actions", "deploy-pages", "main", "src");
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+      assert.ok(data.commit);
+      assert.strictEqual(data.commit.oid.length, 40);
+      assert.ok(data.commit.shortMessageHtml.length > 0);
+      assert.ok(data.commit.date.length > 0);
+      assert.ok(data.commit.authors.length > 0);
+    });
+
+    it("should fetch branches", async () => {
+      const data = await getGitHubRefs("actions", "deploy-pages", "branches");
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+      assert.strictEqual(data.type, "branches");
+
+      assert.ok(data.refs.length > 0);
+      const main = data.refs.find((ref) => ref.name === "main");
+      assert.ok(main);
+      assert.strictEqual(main.isDefault, true);
+    });
+
+    it("should fetch tags", async () => {
+      const data = await getGitHubRefs("actions", "deploy-pages", "tags");
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+      assert.strictEqual(data.type, "tags");
+
+      assert.ok(data.refs.length > 0);
+      const tag = data.refs.find((ref) => ref.name === "v5.0.0");
+      assert.ok(tag);
+      assert.ok(tag.date?.length);
+      assert.strictEqual(data.hasMore, true);
+    });
+
     it("should fetch single release", async () => {
       const data = await getGitHubRelease("actions", "deploy-pages", "v4.0.5");
 
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 53f2bb1..39ad2cd 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -70,7 +70,7 @@ interface IssueIndexPageQuery {
 
 type StylingDirective = [number, number, string];
 
-interface Commit {
+export interface Commit {
   oid: string;
   url: string;
   authoredDate: string;
@@ -270,6 +270,34 @@ export interface GitHubCommits extends GitHubNav {
   commitGroups: CommitGroup[];
 }
 
+export interface LatestCommit {
+  oid: string;
+  date: string;
+  shortMessageHtml: string;
+  authors: {
+    login: string;
+    displayName: string;
+    avatarUrl: string;
+  }[];
+}
+
+export interface GitHubLatestCommit extends GitHubCommon {
+  commit: LatestCommit;
+}
+
+export interface Ref {
+  name: string;
+  isDefault?: boolean;
+  date?: string;
+  authorLogin?: string;
+}
+
+export interface GitHubRefs extends GitHubCommon {
+  type: "branches" | "tags";
+  refs: Ref[];
+  hasMore: boolean;
+}
+
 interface Release {
   tagName: string;
   title: string;
@@ -325,13 +353,13 @@ export interface GitHubOwner {
   repos: OwnerRepo[];
 }
 
-async function fetchGitHubPage(path: string): Promise<string> {
+async function fetchGitHubResponse(path: string, accept: string): Promise<Response> {
   // GitHub throttles/blocks requests without realistic browser headers.
   // These headers make the request appear as a standard browser visit.
   const headers = {
     "User-Agent":
       "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36",
-    Accept: "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
+    Accept: accept,
     "Accept-Language": "en-US,en;q=0.5",
   };
   const response = await fetch(`https://github.com/${path}`, { headers, redirect: "manual" });
@@ -345,10 +373,25 @@ async function fetchGitHubPage(path: string): Promise<string> {
     }
   }
   if (!response.ok) throw new GitHubHTTPError(response.status, response.statusText);
+  return response;
+}
+
+async function fetchGitHubPage(path: string): Promise<string> {
+  const response = await fetchGitHubResponse(
+    path,
+    "text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8",
+  );
   return response.text();
 }
 
-function parsePayload(html: string): AppPayload | null {
+// Some routes (e.g. latest-commit) are the same endpoints the React UI itself
+// calls to lazily fill in data; they 406 without an explicit JSON Accept.
+async function fetchGitHubJson<T>(path: string): Promise<T> {
+  const response = await fetchGitHubResponse(path, "application/json");
+  return (await response.json()) as T;
+}
+
+function parseEmbeddedPayload<T>(html: string): T | null {
   const regex = new RegExp(
     `<script type="application/json" data-target="react-app.embeddedData">([^<]+)</script>`,
     "g",
@@ -367,6 +410,10 @@ function parsePayload(html: string): AppPayload | null {
   return null;
 }
 
+function parsePayload(html: string): AppPayload | null {
+  return parseEmbeddedPayload<AppPayload>(html);
+}
+
 function parseRepoInfo(html: string): RepoInfo {
   const document = parseDocument(html);
 
@@ -639,6 +686,121 @@ export async function getGitHubCommits(
   };
 }
 
+interface LatestCommitPayload {
+  oid: string;
+  date: string;
+  shortMessageHtmlLink: string;
+  authors: {
+    login: string;
+    displayName: string;
+    avatarUrl: string;
+  }[];
+}
+
+// This is the same endpoint the React UI itself calls to lazily fill in the
+// commit bar above a file listing, rather than the much heavier /commits
+// history page: it returns just the single latest commit as JSON.
+export async function getGitHubLatestCommit(
+  owner: string,
+  repo: string,
+  branch: string,
+  path: string,
+): Promise<GitHubLatestCommit> {
+  const urlPath = path
+    ? `${owner}/${repo}/latest-commit/${branch}/${path}`
+    : `${owner}/${repo}/latest-commit/${branch}`;
+  const payload = await fetchGitHubJson<LatestCommitPayload>(urlPath);
+
+  return {
+    repo: { owner, name: repo },
+    commit: {
+      oid: payload.oid,
+      date: payload.date,
+      shortMessageHtml: payload.shortMessageHtmlLink,
+      authors: payload.authors,
+    },
+  };
+}
+
+interface BranchRefEntry {
+  name: string;
+  isDefault: boolean;
+  authoredDate?: string;
+  author?: { login: string };
+}
+
+interface BranchesPayload {
+  branches: {
+    default?: BranchRefEntry;
+    yours?: BranchRefEntry[];
+    active?: BranchRefEntry[];
+    stale?: BranchRefEntry[];
+  };
+  hasMore?: Record<string, boolean>;
+}
+
+async function getGitHubBranchRefs(owner: string, repo: string): Promise<GitHubRefs> {
+  const html = await fetchGitHubPage(`${owner}/${repo}/branches`);
+  const payload = parseEmbeddedPayload<BranchesPayload>(html);
+
+  if (payload?.branches === undefined) {
+    throw new Error("Could not find branches data in embedded JSON");
+  }
+
+  const { default: defaultBranch, yours = [], active = [], stale = [] } = payload.branches;
+  const entries = [...(defaultBranch ? [defaultBranch] : []), ...yours, ...active, ...stale];
+
+  const seen = new Set<string>();
+  const refs: Ref[] = [];
+  for (const entry of entries) {
+    if (seen.has(entry.name)) continue;
+    seen.add(entry.name);
+    refs.push({
+      name: entry.name,
+      isDefault: entry.isDefault || undefined,
+      date: entry.authoredDate,
+      authorLogin: entry.author?.login,
+    });
+  }
+
+  const hasMore = Object.values(payload.hasMore ?? {}).some(Boolean);
+
+  return { repo: { owner, name: repo }, type: "branches", refs, hasMore };
+}
+
+async function getGitHubTagRefs(owner: string, repo: string): Promise<GitHubRefs> {
+  const html = await fetchGitHubPage(`${owner}/${repo}/tags`);
+  const document = parseDocument(html);
+
+  const refs: Ref[] = [];
+  for (const row of cssSelect.selectAll(".Box-row", document) as unknown as Element[]) {
+    const nameElem = cssSelect.selectOne("h2 a", row);
+    if (!nameElem) continue;
+    const name = textContent(nameElem).trim();
+    if (!name) continue;
+
+    // The row has two <relative-time> elements: a release/tag creation
+    // timestamp, then the tagged commit's authored date. We want the latter.
+    const timeElems = cssSelect.selectAll("relative-time", row);
+    const timeElem = timeElems[1] as Element | undefined;
+    const date = timeElem ? getAttributeValue(timeElem, "datetime") || undefined : undefined;
+
+    refs.push({ name, date });
+  }
+
+  const hasMore = cssSelect.selectOne(".pagination a", document) !== null;
+
+  return { repo: { owner, name: repo }, type: "tags", refs, hasMore };
+}
+
+export async function getGitHubRefs(
+  owner: string,
+  repo: string,
+  type: "branches" | "tags",
+): Promise<GitHubRefs> {
+  return type === "branches" ? getGitHubBranchRefs(owner, repo) : getGitHubTagRefs(owner, repo);
+}
+
 function parseRelease(section: Element, title: string): Release {
   // Extract tag name from link
   const tagLink = cssSelect.selectOne('a[href*="/tree/"]', section);
@@ -867,9 +1029,10 @@ function parseUser(document: AnyNode, login: string): GitHubOwner {
 }
 
 function parseOrganization(document: AnyNode, owner: string): GitHubOwner {
-  const avatarElem = cssSelect.selectOne(".orghead img[itemprop='image']", document) as
-    | Element
-    | null;
+  const avatarElem = cssSelect.selectOne(
+    ".orghead img[itemprop='image']",
+    document,
+  ) as Element | null;
   const avatarUrl = avatarElem ? getAttributeValue(avatarElem, "src") || null : null;
   // The login only appears in the avatar's "@login" alt text.
   const alt = avatarElem ? getAttributeValue(avatarElem, "alt") : undefined;
diff --git a/cgithub/views/_latestCommit.eta b/cgithub/views/_latestCommit.eta
new file mode 100644
index 0000000..18d22a1
--- /dev/null
+++ b/cgithub/views/_latestCommit.eta
@@ -0,0 +1,16 @@
+<div class="commit box">
+  <div>
+    <div>
+      <%~ it.commit.shortMessageHtml %>
+    </div>
+    <div class="commit-author">
+      <% for (const author of it.commit.authors) { %>
+        <a href="/<%= author.login %>"><%= author.login %></a>
+      <% } %>
+      <time class="timestamp" title="<%= it.commit.date %>"><%= it.commit.date.split("T")[0] %></time>
+    </div>
+  </div>
+  <div class="commit-sha">
+    <code><%= it.commit.oid.substring(0, 7) %></code>
+  </div>
+</div>
diff --git a/cgithub/views/_nav.eta b/cgithub/views/_nav.eta
index 338a6a6..e3fbdcf 100644
--- a/cgithub/views/_nav.eta
+++ b/cgithub/views/_nav.eta
@@ -1,5 +1,14 @@
 <section>
-  <div class="branch"><img src="/static/icons/git-branch.svg" alt="branch" class="icon"> <%= it.branch %></div>
+  <button type="button" class="ref-selector-button branch" command="show-modal" commandfor="ref-selector-menu">
+    <img src="/static/icons/git-branch.svg" alt="branch" class="icon"> <%= it.branch %> ▾
+  </button>
+  <dialog class="ref-selector-menu" id="ref-selector-menu" closedby="any" data-owner="<%= it.repo.owner %>" data-repo="<%= it.repo.name %>">
+    <div class="ref-selector-tabs">
+      <button type="button" class="active" data-ref-type="branches">Branches</button>
+      <button type="button" data-ref-type="tags">Tags</button>
+    </div>
+    <div class="ref-selector-list">Loading…</div>
+  </dialog>
   <% if (it.path && it.path !== "/") { %>
     <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/tree/<%= it.branch %>">(top)</a><%
     const pathParts = it.path.split('/').filter(p => p);
diff --git a/cgithub/views/_refs.eta b/cgithub/views/_refs.eta
new file mode 100644
index 0000000..2450c4b
--- /dev/null
+++ b/cgithub/views/_refs.eta
@@ -0,0 +1,17 @@
+<% if (it.refs.length === 0) { %>
+  <div class="ref-selector-empty">No <%= it.type %> found.</div>
+<% } else { %>
+  <ul class="ref-selector-items">
+    <% for (const ref of it.refs) { %>
+      <li>
+        <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/tree/<%= ref.name %>"><%= ref.name %></a>
+        <% if (ref.isDefault) { %><span class="ref-selector-badge">default</span><% } %>
+      </li>
+    <% } %>
+  </ul>
+<% } %>
+<% if (it.hasMore) { %>
+  <div class="ref-selector-more">
+    <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/<%= it.type %>">View all <%= it.type %></a>
+  </div>
+<% } %>
diff --git a/cgithub/views/_tree.eta b/cgithub/views/_tree.eta
index ed6eee9..a0ecfab 100644
--- a/cgithub/views/_tree.eta
+++ b/cgithub/views/_tree.eta
@@ -1,5 +1,8 @@
 <%~ include("./_nav.eta", {...it, isTree: true}) %>
 
+<% const commitPath = it.path && it.path !== "/" ? it.path : ""; %>
+<div class="latest-commit-placeholder box" data-src="/api/<%= it.repo.owner %>/<%= it.repo.name %>/latest-commit/<%= it.branch %>/<%= commitPath %>">Loading latest commit…</div>
+
 <table>
   <% for (const item of it.items) { %>
     <tr>
diff --git a/cgithub/views/layout.eta b/cgithub/views/layout.eta
index dd523d1..a80bd1d 100644
--- a/cgithub/views/layout.eta
+++ b/cgithub/views/layout.eta
@@ -5,6 +5,7 @@
   <meta name="viewport" content="width=device-width, initial-scale=1.0">
   <title><%= it.repo.owner %>/<%= it.repo.name %> - <%= it.path || '/' %></title>
   <link rel="stylesheet" href="/static/style.css">
+  <script src="/static/refs.js" defer></script>
 </head>
 <body>
   <div class="box">
-- 
cgit v1.3.1


From 0be1fdd2635fbc49d7451d3711b5772e1e9990e2 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Tue, 11 Aug 2026 08:09:47 -0400
Subject: Add language percentages to repo pages

GitHub no longer server-renders the repository sidebar; its React UI
lazily fetches /:owner/:repo/_sidebar as JSON, which carries the
language breakdown (name, percentage, color). getGitHubLanguages() uses
that endpoint, and a new /api/:owner/:repo/languages fragment route
renders the bar and legend the same way refs and latest-commit already
work, so repo page loads still cost a single scrape.

The latest-commit placeholder loader in refs.js is generalized over
.placeholder[data-src] so both placeholders share it.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
---
 cgithub/public/static/refs.js   | 21 ++++++++++-----------
 cgithub/public/static/style.css | 37 ++++++++++++++++++++++++++++++++++++-
 cgithub/src/app.ts              |  6 ++++++
 cgithub/src/scraper.test.ts     | 28 ++++++++++++++++++++++++++++
 cgithub/src/scraper.ts          | 39 +++++++++++++++++++++++++++++++++++++++
 cgithub/views/_languages.eta    | 17 +++++++++++++++++
 cgithub/views/_tree.eta         |  2 +-
 cgithub/views/repo.eta          |  2 ++
 8 files changed, 139 insertions(+), 13 deletions(-)
 create mode 100644 cgithub/views/_languages.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/public/static/refs.js b/cgithub/public/static/refs.js
index 28b8373..ffff505 100644
--- a/cgithub/public/static/refs.js
+++ b/cgithub/public/static/refs.js
@@ -1,8 +1,8 @@
-// Progressive enhancement for two bits of data GitHub itself loads lazily:
-// the branch/tag dropdown and the latest-commit bar above a file listing.
-// Both degrade gracefully without JS: the button's command/commandfor still
-// opens the (empty) dialog natively; the commit placeholder just never fills
-// in.
+// Progressive enhancement for the bits of data GitHub itself loads lazily:
+// the branch/tag dropdown, the latest-commit bar above a file listing, and the
+// language breakdown. All degrade gracefully without JS: the button's
+// command/commandfor still opens the (empty) dialog natively; the placeholders
+// just never fill in.
 document.addEventListener("DOMContentLoaded", () => {
   const dialog = document.querySelector(".ref-selector-menu");
   if (dialog) {
@@ -44,17 +44,16 @@ document.addEventListener("DOMContentLoaded", () => {
     }
   }
 
-  const commitPlaceholder = document.querySelector(".latest-commit-placeholder");
-  if (commitPlaceholder) {
-    fetch(commitPlaceholder.dataset.src)
+  for (const placeholder of document.querySelectorAll(".placeholder[data-src]")) {
+    fetch(placeholder.dataset.src)
       .then((res) => (res.ok ? res.text() : ""))
       .then((html) => {
         if (html.trim()) {
-          commitPlaceholder.outerHTML = html;
+          placeholder.outerHTML = html;
         } else {
-          commitPlaceholder.remove();
+          placeholder.remove();
         }
       })
-      .catch(() => commitPlaceholder.remove());
+      .catch(() => placeholder.remove());
   }
 });
diff --git a/cgithub/public/static/style.css b/cgithub/public/static/style.css
index 983cb86..ed64bc6 100644
--- a/cgithub/public/static/style.css
+++ b/cgithub/public/static/style.css
@@ -218,10 +218,45 @@ body > footer {
   font-size: 0.85rem;
 }
 
-.latest-commit-placeholder {
+.placeholder {
   opacity: 0.7;
 }
 
+.languages {
+  .language-bar {
+    display: flex;
+    height: .5rem;
+    border-radius: .25rem;
+    overflow: hidden;
+
+    span {
+      /* Keep slivers of tiny languages visible. */
+      min-width: 1px;
+    }
+  }
+
+  .language-list {
+    display: flex;
+    flex-wrap: wrap;
+    gap: .2rem 1rem;
+    list-style: none;
+    margin: .5rem 0 0;
+    padding: 0;
+    font-size: 0.85rem;
+  }
+
+  .language-dot {
+    display: inline-block;
+    width: .6rem;
+    height: .6rem;
+    border-radius: 50%;
+  }
+
+  .language-percentage {
+    opacity: 0.7;
+  }
+}
+
 .repo-info {
   .description {
     margin-bottom: .2rem;
diff --git a/cgithub/src/app.ts b/cgithub/src/app.ts
index ca79b1f..e34c7b1 100644
--- a/cgithub/src/app.ts
+++ b/cgithub/src/app.ts
@@ -5,6 +5,7 @@ import {
   getGitHubBlob,
   getGitHubCommits,
   getGitHubIssues,
+  getGitHubLanguages,
   getGitHubLatestCommit,
   getGitHubOwner,
   getGitHubPulls,
@@ -158,6 +159,11 @@ export function createApp(eta: Eta) {
     );
   });
 
+  app.get("/api/:owner/:repo/languages", async (c) => {
+    const { owner, repo } = c.req.param();
+    return tryRenderFragment(c, "_languages.eta", getGitHubLanguages(owner, repo));
+  });
+
   app.get("/:owner/:repo/releases/tag/:tag", async (c) => {
     const { owner, repo, tag } = c.req.param();
     return tryRender(c, "release.eta", getGitHubRelease(owner, repo, tag));
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 56f14b0..5b5e78b 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -4,6 +4,7 @@ import {
   getGitHubBlob,
   getGitHubCommits,
   getGitHubIssues,
+  getGitHubLanguages,
   getGitHubLatestCommit,
   getGitHubPulls,
   getGitHubRefs,
@@ -219,6 +220,33 @@ describe("GitHub scraper", () => {
       assert.ok(data.commit.authors.length > 0);
     });
 
+    it("should fetch languages", async () => {
+      const data = await getGitHubLanguages("actions", "deploy-pages");
+
+      assert.strictEqual(data.repo.owner, "actions");
+      assert.strictEqual(data.repo.name, "deploy-pages");
+
+      assert.deepStrictEqual(data.languages, [
+        { name: "JavaScript", percentage: 100, color: "#f1e05a" },
+      ]);
+    });
+
+    it("should fetch languages for a multi-language repository", async () => {
+      const data = await getGitHubLanguages("rust-lang", "rust");
+
+      const rust = data.languages.find((language) => language.name === "Rust");
+      assert.ok(rust);
+      assert.ok(rust.percentage > 50);
+      assert.strictEqual(rust.color, "#dea584");
+
+      // The aggregate bucket comes back without a name.
+      const other = data.languages.find((language) => language.name === "Other");
+      assert.ok(other);
+
+      const total = data.languages.reduce((sum, language) => sum + language.percentage, 0);
+      assert.ok(Math.abs(total - 100) < 1);
+    });
+
     it("should fetch branches", async () => {
       const data = await getGitHubRefs("actions", "deploy-pages", "branches");
 
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 39ad2cd..223839f 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -285,6 +285,17 @@ export interface GitHubLatestCommit extends GitHubCommon {
   commit: LatestCommit;
 }
 
+export interface Language {
+  // "Other" for the aggregate bucket GitHub reports without a name.
+  name: string;
+  percentage: number;
+  color: string | null;
+}
+
+export interface GitHubLanguages extends GitHubCommon {
+  languages: Language[];
+}
+
 export interface Ref {
   name: string;
   isDefault?: boolean;
@@ -722,6 +733,34 @@ export async function getGitHubLatestCommit(
   };
 }
 
+interface SidebarPayload {
+  languages?: {
+    languages: {
+      // Absent for the "Other" bucket.
+      name?: string;
+      percentage: number;
+      color?: string;
+    }[];
+  };
+}
+
+// The repository sidebar (languages, contributors, ...) is not rendered into
+// the repo page at all; GitHub's own UI lazily fetches it as JSON from here.
+export async function getGitHubLanguages(owner: string, repo: string): Promise<GitHubLanguages> {
+  const payload = await fetchGitHubJson<SidebarPayload>(`${owner}/${repo}/_sidebar`);
+
+  const languages: Language[] = [];
+  for (const entry of payload.languages?.languages ?? []) {
+    languages.push({
+      name: entry.name ?? "Other",
+      percentage: entry.percentage,
+      color: entry.color ?? null,
+    });
+  }
+
+  return { repo: { owner, name: repo }, languages };
+}
+
 interface BranchRefEntry {
   name: string;
   isDefault: boolean;
diff --git a/cgithub/views/_languages.eta b/cgithub/views/_languages.eta
new file mode 100644
index 0000000..8aca447
--- /dev/null
+++ b/cgithub/views/_languages.eta
@@ -0,0 +1,17 @@
+<% if (it.languages.length > 0) { %>
+  <div class="languages box">
+    <div class="language-bar">
+      <% for (const lang of it.languages) { %>
+        <span style="width: <%= lang.percentage %>%; background: <%= lang.color ?? "#ededed" %>" title="<%= lang.name %> <%= lang.percentage %>%"></span>
+      <% } %>
+    </div>
+    <ul class="language-list">
+      <% for (const lang of it.languages) { %>
+        <li>
+          <span class="language-dot" style="background: <%= lang.color ?? "#ededed" %>"></span>
+          <%= lang.name %> <span class="language-percentage"><%= lang.percentage %>%</span>
+        </li>
+      <% } %>
+    </ul>
+  </div>
+<% } %>
diff --git a/cgithub/views/_tree.eta b/cgithub/views/_tree.eta
index a0ecfab..2eb445f 100644
--- a/cgithub/views/_tree.eta
+++ b/cgithub/views/_tree.eta
@@ -1,7 +1,7 @@
 <%~ include("./_nav.eta", {...it, isTree: true}) %>
 
 <% const commitPath = it.path && it.path !== "/" ? it.path : ""; %>
-<div class="latest-commit-placeholder box" data-src="/api/<%= it.repo.owner %>/<%= it.repo.name %>/latest-commit/<%= it.branch %>/<%= commitPath %>">Loading latest commit…</div>
+<div class="placeholder box" data-src="/api/<%= it.repo.owner %>/<%= it.repo.name %>/latest-commit/<%= it.branch %>/<%= commitPath %>">Loading latest commit…</div>
 
 <table>
   <% for (const item of it.items) { %>
diff --git a/cgithub/views/repo.eta b/cgithub/views/repo.eta
index 683a9ad..9122594 100644
--- a/cgithub/views/repo.eta
+++ b/cgithub/views/repo.eta
@@ -29,4 +29,6 @@
   </div>
 <% } %>
 
+<div class="placeholder box" data-src="/api/<%= it.repo.owner %>/<%= it.repo.name %>/languages">Loading languages…</div>
+
 <%~ include("./_tree", it) %>
-- 
cgit v1.3.1


From 21310d0366c0f21e73c1f4f78d38370edfb59445 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Tue, 11 Aug 2026 17:41:01 -0400
Subject: Read the About box and latest release from JSON

GitHub rebuilt the repository sidebar, so the selectors parseRepoInfo used
matched nothing and the About box rendered empty. Everything but the latest
release is in the page's embedded payload as sidebarAbout; the latest release
is only in the _sidebar JSON already fetched for languages, so it arrives with
that fragment and is moved beside the release count.
---
 cgithub/public/static/refs.js |  10 ++++
 cgithub/src/app.test.ts       |  34 ++++++++++++
 cgithub/src/app.ts            |   6 +-
 cgithub/src/scraper.test.ts   |  15 ++---
 cgithub/src/scraper.ts        | 126 ++++++++++++++----------------------------
 cgithub/views/_languages.eta  |  17 ------
 cgithub/views/_sidebar.eta    |  22 ++++++++
 cgithub/views/repo.eta        |  10 +---
 8 files changed, 122 insertions(+), 118 deletions(-)
 delete mode 100644 cgithub/views/_languages.eta
 create mode 100644 cgithub/views/_sidebar.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/public/static/refs.js b/cgithub/public/static/refs.js
index f3e07e8..1035593 100644
--- a/cgithub/public/static/refs.js
+++ b/cgithub/public/static/refs.js
@@ -52,10 +52,20 @@ document.addEventListener("DOMContentLoaded", () => {
       .then((html) => {
         if (html.trim()) {
           placeholder.outerHTML = html;
+          relocate();
         } else {
           placeholder.remove();
         }
       })
       .catch(() => placeholder.remove());
   }
+
+  // A fragment may carry a node that belongs elsewhere on the page, where an
+  // empty element of that id holds its spot: the sidebar's latest release
+  // reads as part of the release count, far from the languages it arrives with.
+  function relocate() {
+    for (const node of document.querySelectorAll("[data-target]")) {
+      document.getElementById(node.dataset.target)?.replaceWith(node);
+    }
+  }
 });
diff --git a/cgithub/src/app.test.ts b/cgithub/src/app.test.ts
index 6544f03..5474f8d 100644
--- a/cgithub/src/app.test.ts
+++ b/cgithub/src/app.test.ts
@@ -58,6 +58,40 @@ describe("branch picker", () => {
   });
 });
 
+describe("sidebar fragment", () => {
+  const repo = { owner: "actions", name: "deploy-pages" };
+  const languages = [{ name: "JavaScript", percentage: 100, color: "#f1e05a" }];
+  const latestRelease = { tag: "v5.0.0", publishedAt: "2026-03-25T16:59:14Z" };
+
+  // The release replaces the empty element it names, beside the release count.
+  it("should render the latest release and the languages", () => {
+    const html = eta.render("_sidebar.eta", { repo, languages, latestRelease });
+
+    assert.match(html, /<span data-target="latest-release">\(<a href/);
+    assert.match(html, /href="\/actions\/deploy-pages\/releases\/tag\/v5.0.0"/);
+    assert.match(html, /<time class="timestamp" title="2026-03-25T16:59:14Z">2026-03-25<\/time>/);
+    assert.match(html, /<div class="languages box">/);
+  });
+
+  it("should name a spot the repo page holds open", () => {
+    const html = eta.render("repo.eta", {
+      repo,
+      branch: "main",
+      path: "/",
+      items: [],
+      info: { description: null, website: null, stars: null, forks: null, numReleases: 38 },
+    });
+
+    assert.match(html, /<span id="latest-release"><\/span>/);
+  });
+
+  it("should render nothing when the repository has neither", () => {
+    const html = eta.render("_sidebar.eta", { repo, languages: [], latestRelease: null });
+
+    assert.strictEqual(html.trim(), "");
+  });
+});
+
 describe("redirects to GitHub", () => {
   // An HTTP redirect would be caught by the extension's declarativeNetRequest
   // rules and bounced straight back here.
diff --git a/cgithub/src/app.ts b/cgithub/src/app.ts
index 81e47fd..e27db59 100644
--- a/cgithub/src/app.ts
+++ b/cgithub/src/app.ts
@@ -6,7 +6,7 @@ import {
   getGitHubBlob,
   getGitHubCommits,
   getGitHubIssues,
-  getGitHubLanguages,
+  getGitHubSidebar,
   getGitHubLatestCommit,
   getGitHubOwner,
   getGitHubPulls,
@@ -178,9 +178,9 @@ export function createApp(eta: Eta) {
     );
   });
 
-  app.get("/api/:owner/:repo/languages", async (c) => {
+  app.get("/api/:owner/:repo/sidebar", async (c) => {
     const { owner, repo } = c.req.param();
-    return tryRenderFragment(c, "_languages.eta", getGitHubLanguages(owner, repo));
+    return tryRenderFragment(c, "_sidebar.eta", getGitHubSidebar(owner, repo));
   });
 
   app.get("/:owner/:repo/releases/tag/:tag", async (c) => {
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 5b5e78b..24fdfc0 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -4,7 +4,7 @@ import {
   getGitHubBlob,
   getGitHubCommits,
   getGitHubIssues,
-  getGitHubLanguages,
+  getGitHubSidebar,
   getGitHubLatestCommit,
   getGitHubPulls,
   getGitHubRefs,
@@ -35,8 +35,6 @@ describe("GitHub scraper", () => {
       assert.strictEqual(data.info.stars?.length, 3);
       assert.strictEqual(data.info.forks?.length, 3);
       assert.strictEqual(data.info.numReleases, 38);
-      assert.strictEqual(data.info.latestReleaseTag, "v5.0.0");
-      assert.ok(data.info.latestReleaseDate?.startsWith("2026-03-25"));
 
       assert.ok(data.items.length > 0);
       assert.ok(typeof data.overviewHtml?.["README.md"] === "string");
@@ -220,8 +218,8 @@ describe("GitHub scraper", () => {
       assert.ok(data.commit.authors.length > 0);
     });
 
-    it("should fetch languages", async () => {
-      const data = await getGitHubLanguages("actions", "deploy-pages");
+    it("should fetch the sidebar", async () => {
+      const data = await getGitHubSidebar("actions", "deploy-pages");
 
       assert.strictEqual(data.repo.owner, "actions");
       assert.strictEqual(data.repo.name, "deploy-pages");
@@ -229,10 +227,13 @@ describe("GitHub scraper", () => {
       assert.deepStrictEqual(data.languages, [
         { name: "JavaScript", percentage: 100, color: "#f1e05a" },
       ]);
+
+      assert.strictEqual(data.latestRelease?.tag, "v5.0.0");
+      assert.ok(data.latestRelease?.publishedAt?.startsWith("2026-03-25"));
     });
 
-    it("should fetch languages for a multi-language repository", async () => {
-      const data = await getGitHubLanguages("rust-lang", "rust");
+    it("should fetch the sidebar for a multi-language repository", async () => {
+      const data = await getGitHubSidebar("rust-lang", "rust");
 
       const rust = data.languages.find((language) => language.name === "Rust");
       assert.ok(rust);
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 223839f..4220dc4 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -143,7 +143,17 @@ interface CodeViewBlobRoute {
   richText: string | null;
 }
 
+interface SidebarAbout {
+  description?: string | null;
+  website?: string | null;
+  stargazerCount?: number;
+  forksCount?: number;
+  // A section is false when the repository has none of that thing.
+  sections?: { releases?: { releaseCount?: number } | false };
+}
+
 interface AppPayload {
+  sidebarAbout?: SidebarAbout;
   codeViewLayoutRoute?: CodeViewLayoutRoute;
   codeViewRepoRoute?: CodeViewRepoRoute;
   codeViewTreeRoute?: CodeViewTreeRoute;
@@ -230,8 +240,6 @@ export interface RepoInfo {
   stars: string | null;
   forks: string | null;
   numReleases?: number;
-  latestReleaseTag?: string;
-  latestReleaseDate?: string;
 }
 
 export interface GitHubRepo extends GitHubTree {
@@ -292,8 +300,14 @@ export interface Language {
   color: string | null;
 }
 
-export interface GitHubLanguages extends GitHubCommon {
+export interface LatestRelease {
+  tag: string;
+  publishedAt: string | null;
+}
+
+export interface GitHubSidebar extends GitHubCommon {
   languages: Language[];
+  latestRelease: LatestRelease | null;
 }
 
 export interface Ref {
@@ -425,81 +439,6 @@ function parsePayload(html: string): AppPayload | null {
   return parseEmbeddedPayload<AppPayload>(html);
 }
 
-function parseRepoInfo(html: string): RepoInfo {
-  const document = parseDocument(html);
-
-  const aboutSection = cssSelect.selectOne(".BorderGrid-cell h2:contains('About')", document);
-  let description: string | null = null;
-  let website: string | null = null;
-
-  if (aboutSection) {
-    const parent = aboutSection.parent;
-    if (parent) {
-      const descElem = cssSelect.selectOne("p.f4", parent);
-      if (descElem) {
-        description = textContent(descElem).trim();
-      }
-      const websiteElem = cssSelect.selectOne(
-        "a[rel='noopener noreferrer nofollow']",
-        parent,
-      ) as Element | null;
-      if (websiteElem) {
-        website = getAttributeValue(websiteElem, "href") || null;
-      }
-    }
-  }
-
-  const starsElem = cssSelect.selectOne("#repo-stars-counter-star", document) as Element | null;
-  const stars = starsElem
-    ? getAttributeValue(starsElem, "title") || textContent(starsElem).trim()
-    : null;
-
-  const forksElem = cssSelect.selectOne("#repo-network-counter", document) as Element | null;
-  const forks = forksElem
-    ? getAttributeValue(forksElem, "title") || textContent(forksElem).trim()
-    : null;
-
-  const releaseElem = cssSelect.selectOne(
-    "a[href$='/releases'] .Counter",
-    document,
-  ) as Element | null;
-  let numReleases: number | undefined;
-  let latestReleaseTag: string | undefined;
-  let latestReleaseDate: string | undefined;
-
-  if (releaseElem) {
-    const releaseText = getAttributeValue(releaseElem, "title") || textContent(releaseElem);
-    numReleases = parseInt(releaseText.replace(/,/g, ""), 10);
-
-    // Find the latest release tag and date if available
-    const cell = (releaseElem.parent as Element)?.parent?.parent as Element; // span -> a -> h2 -> cell
-    if (cell) {
-      const latestLink = cssSelect.selectOne("a[href*='/releases/tag/']", cell) as Element;
-      if (latestLink) {
-        const tagElem = cssSelect.selectOne(".css-truncate-target", latestLink);
-        if (tagElem) {
-          latestReleaseTag = textContent(tagElem).trim();
-        }
-
-        const timeElem = cssSelect.selectOne("relative-time", latestLink);
-        if (timeElem) {
-          latestReleaseDate = getAttributeValue(timeElem, "datetime") || undefined;
-        }
-      }
-    }
-  }
-
-  return {
-    description,
-    website,
-    stars,
-    forks,
-    numReleases,
-    latestReleaseTag,
-    latestReleaseDate,
-  };
-}
-
 function extractOverviewHtml(
   overviewFiles: OverviewFile[] = [],
 ): Record<string, string | null> | undefined {
@@ -539,7 +478,15 @@ export async function getGitHubRepo(owner: string, repo: string): Promise<GitHub
     throw new Error("Could not find tree data in embedded JSON");
   }
 
-  const info = parseRepoInfo(html);
+  const about = payload?.sidebarAbout;
+  const releases = about?.sections?.releases;
+  const info = {
+    description: about?.description ?? null,
+    website: about?.website ?? null,
+    stars: about?.stargazerCount?.toLocaleString("en-US") ?? null,
+    forks: about?.forksCount?.toLocaleString("en-US") ?? null,
+    numReleases: releases === false ? undefined : releases?.releaseCount,
+  };
   const overviewHtml = extractOverviewHtml(repoRoute.overview?.overviewFiles);
 
   return extractGitHub(payload, { items: repoRoute.tree.items, info, overviewHtml });
@@ -742,11 +689,16 @@ interface SidebarPayload {
       color?: string;
     }[];
   };
+  releases?: {
+    // The name is the release title, which need not be the tag; the path is.
+    latestRelease?: { name?: string; path: string; publishedAt?: string } | null;
+  };
 }
 
-// The repository sidebar (languages, contributors, ...) is not rendered into
-// the repo page at all; GitHub's own UI lazily fetches it as JSON from here.
-export async function getGitHubLanguages(owner: string, repo: string): Promise<GitHubLanguages> {
+// The repository sidebar (languages, releases, contributors, ...) is not
+// rendered into the repo page at all; GitHub's own UI lazily fetches it as
+// JSON from here.
+export async function getGitHubSidebar(owner: string, repo: string): Promise<GitHubSidebar> {
   const payload = await fetchGitHubJson<SidebarPayload>(`${owner}/${repo}/_sidebar`);
 
   const languages: Language[] = [];
@@ -758,7 +710,15 @@ export async function getGitHubLanguages(owner: string, repo: string): Promise<G
     });
   }
 
-  return { repo: { owner, name: repo }, languages };
+  const release = payload.releases?.latestRelease;
+  const latestRelease = release
+    ? {
+        tag: decodeURIComponent(release.path.split("/releases/tag/")[1] ?? release.name ?? ""),
+        publishedAt: release.publishedAt ?? null,
+      }
+    : null;
+
+  return { repo: { owner, name: repo }, languages, latestRelease };
 }
 
 interface BranchRefEntry {
diff --git a/cgithub/views/_languages.eta b/cgithub/views/_languages.eta
deleted file mode 100644
index 8aca447..0000000
--- a/cgithub/views/_languages.eta
+++ /dev/null
@@ -1,17 +0,0 @@
-<% if (it.languages.length > 0) { %>
-  <div class="languages box">
-    <div class="language-bar">
-      <% for (const lang of it.languages) { %>
-        <span style="width: <%= lang.percentage %>%; background: <%= lang.color ?? "#ededed" %>" title="<%= lang.name %> <%= lang.percentage %>%"></span>
-      <% } %>
-    </div>
-    <ul class="language-list">
-      <% for (const lang of it.languages) { %>
-        <li>
-          <span class="language-dot" style="background: <%= lang.color ?? "#ededed" %>"></span>
-          <%= lang.name %> <span class="language-percentage"><%= lang.percentage %>%</span>
-        </li>
-      <% } %>
-    </ul>
-  </div>
-<% } %>
diff --git a/cgithub/views/_sidebar.eta b/cgithub/views/_sidebar.eta
new file mode 100644
index 0000000..bceef7c
--- /dev/null
+++ b/cgithub/views/_sidebar.eta
@@ -0,0 +1,22 @@
+<% if (it.latestRelease) { %>
+  <span data-target="latest-release">(<a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases/tag/<%= it.latestRelease.tag %>"><%= it.latestRelease.tag %></a><%
+    if (it.latestRelease.publishedAt) { %> <time class="timestamp" title="<%= it.latestRelease.publishedAt %>"><%= it.latestRelease.publishedAt.split("T")[0] %></time><%
+    } %>)</span>
+<% } %>
+<% if (it.languages.length > 0) { %>
+  <div class="languages box">
+    <div class="language-bar">
+      <% for (const lang of it.languages) { %>
+        <span style="width: <%= lang.percentage %>%; background: <%= lang.color ?? "#ededed" %>" title="<%= lang.name %> <%= lang.percentage %>%"></span>
+      <% } %>
+    </div>
+    <ul class="language-list">
+      <% for (const lang of it.languages) { %>
+        <li>
+          <span class="language-dot" style="background: <%= lang.color ?? "#ededed" %>"></span>
+          <%= lang.name %> <span class="language-percentage"><%= lang.percentage %>%</span>
+        </li>
+      <% } %>
+    </ul>
+  </div>
+<% } %>
diff --git a/cgithub/views/repo.eta b/cgithub/views/repo.eta
index 1bff97f..5dc6beb 100644
--- a/cgithub/views/repo.eta
+++ b/cgithub/views/repo.eta
@@ -17,19 +17,13 @@
     <% if (it.info.numReleases) { %>
       <div class="detail">
         <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases"><%= it.info.numReleases %> releases</a>
-        <% if (it.info.latestReleaseTag) { %>
-          (<a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases/tag/<%= it.info.latestReleaseTag %>"><%= it.info.latestReleaseTag %></a>
-          <% if (it.info.latestReleaseDate) { %>
-            <time class="timestamp" title="<%= it.info.latestReleaseDate %>"><%= it.info.latestReleaseDate.split("T")[0] %></time>
-          <% } %>
-          )
-        <% } %>
+        <span id="latest-release"></span>
       </div>
     <% } %>
   </div>
 <% } %>
 
-<div class="placeholder languages box" data-src="/api/<%= it.repo.owner %>/<%= it.repo.name %>/languages">
+<div class="placeholder languages box" data-src="/api/<%= it.repo.owner %>/<%= it.repo.name %>/sidebar">
     <div class="language-bar"><span style="width: 100%; background: #ededed"></span></div>
     <ul class="language-list"><li>…</li></ul>
 </div>
-- 
cgit v1.3.1


From 0d60b8c8235c3c07f675aba7a56dd37b62519f9e Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Thu, 27 Aug 2026 14:08:06 -0400
Subject: Display PDF blobs

GitHub renders a PDF with a viewer of its own and so ships no content for
one, leaving the blob view with nothing to show. Its raw bytes come as
application/octet-stream, which browsers download rather than display, so
serve them from a new embed route that retypes them and point an <object>
at that.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BErPR92xppzGFXkPqXPYj2
---
 cgithub/public/static/style.css |  7 +++++
 cgithub/src/app.test.ts         | 68 +++++++++++++++++++++++++++++++++++++++--
 cgithub/src/app.ts              | 26 ++++++++++++++++
 cgithub/src/scraper.test.ts     | 32 +++++++++++++++++++
 cgithub/src/scraper.ts          | 28 ++++++++++++++---
 cgithub/views/blob.eta          |  7 ++++-
 6 files changed, 161 insertions(+), 7 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/public/static/style.css b/cgithub/public/static/style.css
index ed64bc6..157cc15 100644
--- a/cgithub/public/static/style.css
+++ b/cgithub/public/static/style.css
@@ -276,6 +276,13 @@ body > footer {
   }
 }
 
+object.pdf {
+  display: block;
+  width: 100%;
+  height: 80vh;
+  padding: 0;
+}
+
 pre.box {
   /* Instead of scrolling overflow for long lines, grow the box. */
   width: fit-content;
diff --git a/cgithub/src/app.test.ts b/cgithub/src/app.test.ts
index 94bf205..e28c375 100644
--- a/cgithub/src/app.test.ts
+++ b/cgithub/src/app.test.ts
@@ -1,6 +1,6 @@
 import assert from "node:assert";
 import path from "node:path";
-import { describe, it } from "node:test";
+import { describe, it, type TestContext } from "node:test";
 import { Eta } from "eta";
 import { createApp } from "./app.ts";
 
@@ -100,12 +100,30 @@ describe("blob view", () => {
     language: null,
     size: "815 KB",
     image: false,
+    renderFileType: null,
     textLines: null,
     htmlLines: null,
     htmlContent: null,
   };
 
-  // GitHub previews neither the text nor the bytes of most binary files.
+  it("should embed a PDF", () => {
+    const html = eta.render("blob.eta", {
+      ...blob,
+      path: "test/pdfs/basicapi.pdf",
+      renderFileType: "pdf",
+    });
+
+    assert.match(html, /<object[^>]*type="application\/pdf"/);
+    assert.match(html, /data="\/justrajdeep\/fonts\/embed\/pdf\/master\/test\/pdfs\/basicapi.pdf"/);
+  });
+
+  it("should not embed a file type it has no viewer for", () => {
+    const html = eta.render("blob.eta", { ...blob, renderFileType: "stl" });
+
+    assert.doesNotMatch(html, /<object/);
+    assert.match(html, /This file cannot be displayed\./);
+  });
+
   it("should offer the raw file when there is nothing to display", () => {
     const html = eta.render("blob.eta", blob);
 
@@ -114,6 +132,52 @@ describe("blob view", () => {
   });
 });
 
+describe("embedded files", () => {
+  function stubGitHub(t: TestContext, body: string, contentType: string, status = 200) {
+    t.mock.method(
+      globalThis,
+      "fetch",
+      async () => new Response(body, { status, headers: { "Content-Type": contentType } }),
+    );
+  }
+
+  it("should retype a PDF GitHub sends as a download", async (t) => {
+    stubGitHub(t, "%PDF-1.6", "application/octet-stream");
+
+    const res = await app.request("http://cgithub.example/mozilla/pdf.js/embed/pdf/master/doc.pdf");
+
+    assert.strictEqual(res.status, 200);
+    assert.strictEqual(res.headers.get("Content-Type"), "application/pdf");
+    assert.strictEqual(await res.text(), "%PDF-1.6");
+  });
+
+  it("should pass on a type GitHub already got right", async (t) => {
+    stubGitHub(t, "%PDF-1.6", "application/pdf");
+
+    const res = await app.request("http://cgithub.example/mozilla/pdf.js/embed/pdf/master/doc.pdf");
+
+    assert.strictEqual(res.headers.get("Content-Type"), "application/pdf");
+  });
+
+  it("should not serve a file type it has no viewer for", async (t) => {
+    stubGitHub(t, "<script>", "text/html");
+
+    const res = await app.request("http://cgithub.example/mozilla/pdf.js/embed/html/master/x.html");
+
+    assert.strictEqual(res.status, 404);
+  });
+
+  it("should pass on the status of a file GitHub does not have", async (t) => {
+    stubGitHub(t, "404: Not Found", "text/plain", 404);
+
+    const res = await app.request(
+      "http://cgithub.example/mozilla/pdf.js/embed/pdf/master/nope.pdf",
+    );
+
+    assert.strictEqual(res.status, 404);
+  });
+});
+
 describe("redirects to GitHub", () => {
   // An HTTP redirect would be caught by the extension's declarativeNetRequest
   // rules and bounced straight back here.
diff --git a/cgithub/src/app.ts b/cgithub/src/app.ts
index 9781a98..33b408e 100644
--- a/cgithub/src/app.ts
+++ b/cgithub/src/app.ts
@@ -10,6 +10,7 @@ import {
   getGitHubLatestCommit,
   getGitHubOwner,
   getGitHubPulls,
+  getGitHubRaw,
   getGitHubRefs,
   getGitHubRelease,
   getGitHubReleases,
@@ -120,6 +121,31 @@ export function createApp(eta: Eta) {
     );
   });
 
+  // This is not a native GitHub route, but instead it lets us render blob
+  // content whose raw headers prevent hotlinking.
+  app.get("/:owner/:repo/embed/:type/:branch/:path{.*}", async (c) => {
+    const { owner, repo, type, branch, path } = c.req.param();
+
+    if (type !== "pdf") {
+      return c.notFound();
+    }
+
+    let response: Response;
+    try {
+      response = await getGitHubRaw(owner, repo, branch, path);
+    } catch (e) {
+      c.status(e instanceof GitHubHTTPError ? (e.status as StatusCode) : 500);
+      return c.body(null);
+    }
+
+    let contentType = response.headers.get("Content-Type");
+    if (contentType === null || contentType === "application/octet-stream") {
+      contentType = "application/pdf";
+    }
+
+    return new Response(response.body, { headers: { "Content-Type": contentType } });
+  });
+
   app.get("/:owner/:repo/issues", async (c) => {
     const { owner, repo } = c.req.param();
     const q = c.req.query("q");
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 836235e..88e9271 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -7,6 +7,7 @@ import {
   getGitHubSidebar,
   getGitHubLatestCommit,
   getGitHubPulls,
+  getGitHubRaw,
   getGitHubRefs,
   getGitHubRelease,
   getGitHubReleases,
@@ -328,11 +329,41 @@ describe("GitHub scraper", () => {
       assert.strictEqual(data.size, "1.26 KB");
       assert.strictEqual(data.language, null);
       assert.strictEqual(data.image, true);
+      assert.strictEqual(data.renderFileType, null);
       assert.strictEqual(data.textLines, null);
       assert.strictEqual(data.htmlLines, null);
       assert.strictEqual(data.htmlContent, null);
     });
 
+    // GitHub renders a PDF with a viewer of its own, so the page carries no
+    // content for it beyond the file type.
+    it("should support PDF", async () => {
+      const data = await getGitHubBlob("mozilla", "pdf.js", "master", "test/pdfs/basicapi.pdf");
+
+      assert.strictEqual(data.repo.owner, "mozilla");
+      assert.strictEqual(data.repo.name, "pdf.js");
+      assert.strictEqual(data.branch, "master");
+      assert.strictEqual(data.path, "test/pdfs/basicapi.pdf");
+
+      assert.strictEqual(data.size, "103 KB");
+      assert.strictEqual(data.language, null);
+      assert.strictEqual(data.image, false);
+      assert.strictEqual(data.renderFileType, "pdf");
+      assert.strictEqual(data.textLines, null);
+      assert.strictEqual(data.htmlLines, null);
+      assert.strictEqual(data.htmlContent, null);
+    });
+
+    // Why the embed route exists rather than the blob view pointing a viewer
+    // at GitHub: as application/octet-stream a browser downloads the file
+    // instead of displaying it.
+    it("should fetch a raw PDF that GitHub types as a download", async () => {
+      const response = await getGitHubRaw("mozilla", "pdf.js", "master", "test/pdfs/basicapi.pdf");
+
+      assert.strictEqual(response.headers.get("Content-Type"), "application/octet-stream");
+      assert.ok((await response.text()).startsWith("%PDF-"));
+    });
+
     // A binary file GitHub has no preview for: no content of any kind, and the
     // path exercises escaping of a space.
     it("should support an unpreviewable binary file", async () => {
@@ -346,6 +377,7 @@ describe("GitHub scraper", () => {
       assert.strictEqual(data.size, "815 KB");
       assert.strictEqual(data.language, null);
       assert.strictEqual(data.image, false);
+      assert.strictEqual(data.renderFileType, null);
       assert.strictEqual(data.textLines, null);
       assert.strictEqual(data.htmlLines, null);
       assert.strictEqual(data.htmlContent, null);
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 4220dc4..1a67d05 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -141,6 +141,7 @@ interface CodeViewBlobLayoutRouteStyledBlob {
 
 interface CodeViewBlobRoute {
   richText: string | null;
+  renderedFileInfo?: { renderFileType: string | null } | null;
 }
 
 interface SidebarAbout {
@@ -250,6 +251,7 @@ export interface GitHubBlob extends GitHubNav {
   language: string | null;
   size: string;
   image: boolean;
+  renderFileType: string | null;
   textLines: string[] | null;
   htmlLines: string[] | null;
   htmlContent: string | null;
@@ -378,15 +380,19 @@ export interface GitHubOwner {
   repos: OwnerRepo[];
 }
 
-async function fetchGitHubResponse(path: string, accept: string): Promise<Response> {
-  // GitHub throttles/blocks requests without realistic browser headers.
-  // These headers make the request appear as a standard browser visit.
-  const headers = {
+// GitHub throttles/blocks requests without realistic browser headers.
+// These headers make the request appear as a standard browser visit.
+function browserHeaders(accept: string): Record<string, string> {
+  return {
     "User-Agent":
       "Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/131.0.0.0 Safari/537.36",
     Accept: accept,
     "Accept-Language": "en-US,en;q=0.5",
   };
+}
+
+async function fetchGitHubResponse(path: string, accept: string): Promise<Response> {
+  const headers = browserHeaders(accept);
   const response = await fetch(`https://github.com/${path}`, { headers, redirect: "manual" });
   if (response.status === 301 || response.status === 302) {
     const location = response.headers.get("location");
@@ -546,12 +552,26 @@ export async function getGitHubBlob(
     language: blob.language || null,
     size,
     image: blob.image,
+    renderFileType: payload?.codeViewBlobRoute?.renderedFileInfo?.renderFileType ?? null,
     textLines: styledBlob?.rawLines ?? null,
     htmlLines,
     htmlContent: payload?.codeViewBlobRoute?.richText ?? null,
   });
 }
 
+// For proxying embeds.
+export async function getGitHubRaw(
+  owner: string,
+  repo: string,
+  branch: string,
+  path: string,
+): Promise<Response> {
+  const url = `https://raw.githubusercontent.com/${owner}/${repo}/${branch}/${path}`;
+  const response = await fetch(url, { headers: browserHeaders("*/*") });
+  if (!response.ok) throw new GitHubHTTPError(response.status, response.statusText);
+  return response;
+}
+
 export async function getGitHubIssues(
   owner: string,
   repo: string,
diff --git a/cgithub/views/blob.eta b/cgithub/views/blob.eta
index c7315f6..444508e 100644
--- a/cgithub/views/blob.eta
+++ b/cgithub/views/blob.eta
@@ -14,6 +14,11 @@
 
 <% if (it.image) { %>
   <div class="box"><img src="/<%= it.repo.owner %>/<%= it.repo.name %>/raw/<%= it.branch %>/<%= it.path %>"></div>
+<% } else if (it.renderFileType === "pdf") { %>
+  <object class="box pdf" type="application/pdf"
+          data="/<%= it.repo.owner %>/<%= it.repo.name %>/embed/<%= it.renderFileType %>/<%= it.branch %>/<%= it.path %>">
+    This file cannot be displayed.
+  </object>
 <% } else if (it.htmlContent) { %>
   <div class="box"><%~ it.htmlContent %></div>
 <% } else if (it.htmlLines) { %>
@@ -21,5 +26,5 @@
 <% } else if (it.textLines) { %>
   <pre class="box"><code><%= it.textLines.join("\n") %></code></pre>
 <% } else { %>
-  <div class="box">This file cannot be displayed</div>
+  <div class="box">This file cannot be displayed.</div>
 <% } %>
-- 
cgit v1.3.1


From 1f014a1f2a51cc9ced0349fe6353a6cc79303161 Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Thu, 27 Aug 2026 15:02:13 -0400
Subject: Follow the commit history to its new home in the payload

GitHub moved commitGroups from the top of the embedded JSON into a
commitsRefRoute, alongside the code view's own routes, so every commits
page threw. The groups themselves are unchanged.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BErPR92xppzGFXkPqXPYj2
---
 cgithub/src/scraper.ts | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 1a67d05..c02cb5e 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -139,6 +139,10 @@ interface CodeViewBlobLayoutRouteStyledBlob {
   stylingDirectives: StylingDirective[][] | null;
 }
 
+interface CommitsRefRoute {
+  commitGroups: CommitGroup[];
+}
+
 interface CodeViewBlobRoute {
   richText: string | null;
   renderedFileInfo?: { renderFileType: string | null } | null;
@@ -161,7 +165,7 @@ interface AppPayload {
   codeViewBlobLayoutRoute?: CodeViewBlobLayoutRoute;
   "codeViewBlobLayoutRoute.StyledBlob"?: CodeViewBlobLayoutRouteStyledBlob;
   codeViewBlobRoute?: CodeViewBlobRoute;
-  commitGroups?: CommitGroup[];
+  commitsRefRoute?: CommitsRefRoute;
   // Actually there may be other query types.
   preloadedQueries?: IssueIndexPageQuery[];
 }
@@ -648,8 +652,9 @@ export async function getGitHubCommits(
   const html = await fetchGitHubPage(urlPath);
 
   const payload = parsePayload(html);
+  const commitsRoute = payload?.commitsRefRoute;
 
-  if (payload?.commitGroups === undefined) {
+  if (commitsRoute === undefined) {
     throw new Error("Could not find commit history in embedded JSON");
   }
 
@@ -660,7 +665,7 @@ export async function getGitHubCommits(
     },
     branch,
     path,
-    commitGroups: payload.commitGroups,
+    commitGroups: commitsRoute.commitGroups,
   };
 }
 
-- 
cgit v1.3.1


From 4c66a07a995c0d551cd284d6e72826b4f7748bab Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Thu, 3 Sep 2026 00:08:06 -0400
Subject: Render repository wikis

Wikis are one of the last parts of GitHub still served as plain Rails
HTML, with no embedded JSON to read, so this scrapes the markup: a page
and any older revision of it, the _Sidebar and _Footer a wiki can
define, and the page index. Every page is already in the sidebar of
every page, since the "Show N more pages" button only unhides the rest,
so linking the whole wiki costs no second request.

Routes mirror GitHub's and resolve in its order, so an underscore name
reserved for its own actions (_pages, _history, _new, _edit) can never
be mistaken for a page; the ones we don't serve fall through to the
usual redirect. A slug reaches us decoded and has to be encoded again
rather than interpolated the way a ref is, or a page named C# would
truncate at the fragment.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_013rQG1io1i4GJQW5RFcXNGd
---
 cgithub/CLAUDE.md               |   3 ++
 cgithub/public/static/style.css |  21 +++++++-
 cgithub/src/app.test.ts         |  98 ++++++++++++++++++++++++++++++++++
 cgithub/src/app.ts              |  32 +++++++++++
 cgithub/src/scraper.test.ts     |  80 ++++++++++++++++++++++++++++
 cgithub/src/scraper.ts          | 115 +++++++++++++++++++++++++++++++++++++++-
 cgithub/views/_wikiNav.eta      |   4 ++
 cgithub/views/layout.eta        |   1 +
 cgithub/views/wiki.eta          |  37 +++++++++++++
 cgithub/views/wikiPages.eta     |  16 ++++++
 10 files changed, 404 insertions(+), 3 deletions(-)
 create mode 100644 cgithub/views/_wikiNav.eta
 create mode 100644 cgithub/views/wiki.eta
 create mode 100644 cgithub/views/wikiPages.eta

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/CLAUDE.md b/cgithub/CLAUDE.md
index 0ef4b47..c0c211d 100644
--- a/cgithub/CLAUDE.md
+++ b/cgithub/CLAUDE.md
@@ -33,6 +33,9 @@ pnpm test
 - `GET /:owner/:repo` - Repository info with directory listing
 - `GET /:owner/:repo/tree/:branch/:path*` - Directory listing
 - `GET /:owner/:repo/blob/:branch/:path` - File content view
+- `GET /:owner/:repo/wiki` - Wiki home page
+- `GET /:owner/:repo/wiki/:page` - Wiki page (also `/:page/:oid` for a revision)
+- `GET /:owner/:repo/wiki/_pages` - Wiki page index
 
 ## Instructions
 
diff --git a/cgithub/public/static/style.css b/cgithub/public/static/style.css
index 157cc15..a83e128 100644
--- a/cgithub/public/static/style.css
+++ b/cgithub/public/static/style.css
@@ -288,7 +288,7 @@ pre.box {
   width: fit-content;
 }
 
-.overview, .release {
+.overview, .release, .wiki {
   overflow-wrap: break-word;
 
   /* This is a bit convoluted: we want the effect of padding but using margins
@@ -332,7 +332,7 @@ pre.box {
   font-weight: bold;
 }
 
-.release > header {
+.release > header, .wiki > header {
   display: flex;
   align-items: baseline;
   > :first-child {
@@ -341,6 +341,23 @@ pre.box {
   }
 }
 
+.wiki > footer {
+  /* The wiki's _Footer page is rendered content, not a caption for the page. */
+  font-size: .9rem;
+}
+
+.wiki-pages {
+  summary {
+    cursor: pointer;
+  }
+
+  ul {
+    columns: 16rem;
+    list-style: none;
+    padding: 0;
+  }
+}
+
 .timestamp {
   white-space: nowrap;
   font-size: .8rem;
diff --git a/cgithub/src/app.test.ts b/cgithub/src/app.test.ts
index e28c375..3ecebd0 100644
--- a/cgithub/src/app.test.ts
+++ b/cgithub/src/app.test.ts
@@ -178,6 +178,104 @@ describe("embedded files", () => {
   });
 });
 
+describe("wiki", () => {
+  const repo = { owner: "ariya", name: "phantomjs" };
+  const page = {
+    repo,
+    title: "Home",
+    updatedBy: "pixiuPL",
+    updatedAt: "2018-02-08T18:57:56Z",
+    bodyHtml: "<p>New to <strong>PhantomJS</strong>?</p>",
+    sidebarHtml: null,
+    footerHtml: null,
+    pages: [
+      { name: "Home", href: "/ariya/phantomjs/wiki" },
+      { name: "API Reference WebPage", href: "/ariya/phantomjs/wiki/API-Reference-WebPage" },
+    ],
+  };
+
+  it("should render the page beside who last edited it", () => {
+    const html = eta.render("wiki.eta", page);
+
+    assert.match(html, /<p>New to <strong>PhantomJS<\/strong>\?<\/p>/);
+    assert.match(html, /pixiuPL edited this page/);
+    assert.match(html, /<time title="2018-02-08T18:57:56Z">2018-02-08<\/time>/);
+  });
+
+  // The _Sidebar and _Footer pages a wiki can define, which GitHub renders
+  // alongside every page rather than as pages of their own.
+  it("should render the sidebar and footer only when the wiki has them", () => {
+    // The page's own content is the last thing in its box, and the box is the
+    // only one on the page.
+    const bare = eta.render("wiki.eta", page);
+    assert.match(bare, /New to <strong>PhantomJS<\/strong>\?<\/p>\s*<\/div>/);
+    assert.doesNotMatch(bare, /Sidebar/);
+
+    const html = eta.render("wiki.eta", {
+      ...page,
+      sidebarHtml: "<ul><li>nav</li></ul>",
+      footerHtml: "<p>fine print</p>",
+    });
+    assert.match(html, /<footer>\s*<p>fine print<\/p>/);
+    assert.match(html, /<header><span>Sidebar<\/span><\/header>\s*<ul><li>nav<\/li><\/ul>/);
+  });
+
+  it("should link every page of the wiki", () => {
+    const html = eta.render("wiki.eta", page);
+
+    assert.match(html, /<summary>2 pages<\/summary>/);
+    assert.match(
+      html,
+      /href="\/ariya\/phantomjs\/wiki\/API-Reference-WebPage">API Reference WebPage/,
+    );
+  });
+
+  it("should list the pages with when each was last updated", () => {
+    const html = eta.render("wikiPages.eta", {
+      repo,
+      pages: [
+        {
+          name: "Home",
+          href: "/ariya/phantomjs/wiki",
+          updatedAt: "2018-02-08T18:57:56Z",
+        },
+      ],
+    });
+
+    assert.match(html, /href="\/ariya\/phantomjs\/wiki">Home/);
+    assert.match(html, /<time title="2018-02-08T18:57:56Z">2018-02-08<\/time>/);
+  });
+
+  // The editor and the timestamp are scraped independently, so a page missing
+  // one of them should still show the other.
+  it("should render each part of the header meta it has", () => {
+    assert.doesNotMatch(eta.render("wiki.eta", { ...page, updatedAt: null }), /<time/);
+    assert.match(eta.render("wiki.eta", { ...page, updatedAt: null }), /pixiuPL edited this page/);
+    assert.doesNotMatch(eta.render("wiki.eta", { ...page, updatedBy: null }), /edited this page/);
+  });
+
+  // GitHub reserves underscore-prefixed names for its own actions, so a page
+  // can never be one; the rest of them are for signed-in users only.
+  it("should not scrape a reserved wiki name as a page", async () => {
+    const reserved = ["/ariya/phantomjs/wiki/_new", "/ariya/phantomjs/wiki/Home/_edit"];
+    for (const path of reserved) {
+      const res = await app.request(`http://cgithub.example${path}`);
+
+      assert.strictEqual(res.status, 200);
+      assert.match(await res.text(), new RegExp(`url=https://github.com${path}"`));
+    }
+  });
+
+  it("should only read a revision off a path that names one", async () => {
+    const res = await app.request("http://cgithub.example/ariya/phantomjs/wiki/Home/latest");
+
+    assert.match(
+      await res.text(),
+      /url=https:\/\/github.com\/ariya\/phantomjs\/wiki\/Home\/latest"/,
+    );
+  });
+});
+
 describe("redirects to GitHub", () => {
   // An HTTP redirect would be caught by the extension's declarativeNetRequest
   // rules and bounced straight back here.
diff --git a/cgithub/src/app.ts b/cgithub/src/app.ts
index 33b408e..956a8cb 100644
--- a/cgithub/src/app.ts
+++ b/cgithub/src/app.ts
@@ -16,6 +16,8 @@ import {
   getGitHubReleases,
   getGitHubRepo,
   getGitHubTree,
+  getGitHubWiki,
+  getGitHubWikiPages,
   GitHubHTTPError,
   RedirectError,
 } from "./scraper.ts";
@@ -146,6 +148,36 @@ export function createApp(eta: Eta) {
     return new Response(response.body, { headers: { "Content-Type": contentType } });
   });
 
+  // GitHub's own wiki URLs, in the order it resolves them: names beginning with
+  // an underscore are reserved for its actions (_pages, _new, _edit, ...), so a
+  // page can never be mistaken for one.
+  app.get("/:owner/:repo/wiki", async (c) => {
+    const { owner, repo } = c.req.param();
+    return tryRender(c, "wiki.eta", getGitHubWiki(owner, repo));
+  });
+
+  app.get("/:owner/:repo/wiki/_pages", async (c) => {
+    const { owner, repo } = c.req.param();
+    return tryRender(c, "wikiPages.eta", getGitHubWikiPages(owner, repo));
+  });
+
+  // An older revision of a page renders exactly like its current one.
+  app.get("/:owner/:repo/wiki/:page/:oid", async (c) => {
+    const { owner, repo, page, oid } = c.req.param();
+    if (page.startsWith("_") || !/^[0-9a-f]{40}$/.test(oid)) {
+      return redirectToGitHub(c, githubUrlFor(c));
+    }
+    return tryRender(c, "wiki.eta", getGitHubWiki(owner, repo, page, oid));
+  });
+
+  app.get("/:owner/:repo/wiki/:page", async (c) => {
+    const { owner, repo, page } = c.req.param();
+    if (page.startsWith("_")) {
+      return redirectToGitHub(c, githubUrlFor(c));
+    }
+    return tryRender(c, "wiki.eta", getGitHubWiki(owner, repo, page));
+  });
+
   app.get("/:owner/:repo/issues", async (c) => {
     const { owner, repo } = c.req.param();
     const q = c.req.query("q");
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index 88e9271..372772b 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -14,6 +14,8 @@ import {
   getGitHubRepo,
   getGitHubOwner,
   getGitHubTree,
+  getGitHubWiki,
+  getGitHubWikiPages,
   GitHubHTTPError,
   RedirectError,
 } from "./scraper.ts";
@@ -384,6 +386,84 @@ describe("GitHub scraper", () => {
     });
   });
 
+  describe("wikis", () => {
+    it("should fetch the home page a wiki serves off /wiki itself", async () => {
+      const data = await getGitHubWiki("microsoft", "vscode");
+
+      assert.strictEqual(data.repo.owner, "microsoft");
+      assert.strictEqual(data.repo.name, "vscode");
+      assert.strictEqual(data.title, "Home");
+      assert.strictEqual(data.updatedBy, "Greg Van Liew");
+      assert.strictEqual(data.updatedAt, "2019-11-22T03:08:32Z");
+
+      assert.match(data.bodyHtml, /Welcome to the Visual Studio Code Wiki/);
+      // The _Sidebar and _Footer pages this wiki defines.
+      assert.match(data.sidebarHtml ?? "", /href="\/microsoft\/vscode\/wiki\/Roadmap"/);
+      assert.match(data.footerHtml ?? "", /Want to contribute to this Wiki\?/);
+    });
+
+    it("should fetch a page by its slug", async () => {
+      const data = await getGitHubWiki("microsoft", "vscode", "Commit-Signing");
+
+      assert.strictEqual(data.title, "Commit Signing");
+      assert.match(data.bodyHtml, /GPG/);
+    });
+
+    // The slug arrives from our route already decoded, so it has to be encoded
+    // again rather than interpolated into the URL as-is.
+    it("should fetch a page whose slug is not path-safe", async () => {
+      const data = await getGitHubWiki(
+        "microsoft",
+        "vscode",
+        "[DEV]-Perf-Tools-for-VS-Code-Development",
+      );
+
+      assert.strictEqual(data.title, "[DEV] Perf Tools for VS Code Development");
+      assert.ok(data.bodyHtml.length > 0);
+    });
+
+    it("should fetch an older revision of a page", async () => {
+      const oid = "f1c60035ba38b68ba383f58e482c312218daf86c";
+      const data = await getGitHubWiki("microsoft", "vscode", "Commit-Signing", oid);
+
+      assert.strictEqual(data.title, "Commit Signing");
+      assert.strictEqual(data.updatedBy, "Joaqu\u00edn Ruales");
+      assert.strictEqual(data.updatedAt, "2025-10-17T15:07:14Z");
+    });
+
+    // The sidebar list is complete; its "Show N more pages…" button only
+    // unhides the rest, so no second request is needed to link every page.
+    it("should collect every page from the sidebar of any page", async () => {
+      const data = await getGitHubWiki("microsoft", "vscode");
+
+      assert.ok(data.pages.length > 15);
+      assert.deepStrictEqual(data.pages[0], { name: "Home", href: "/microsoft/vscode/wiki" });
+      assert.ok(data.pages.some((page) => page.href === "/microsoft/vscode/wiki/Commit-Signing"));
+    });
+
+    // An archived repository, so neither its wiki nor these dates can move.
+    it("should fetch the page index with its last-updated dates", async () => {
+      const data = await getGitHubWikiPages("ariya", "phantomjs");
+
+      assert.strictEqual(data.repo.owner, "ariya");
+      assert.ok(data.pages.length > 1);
+      assert.ok(data.pages.every((page) => page.updatedAt));
+
+      const home = data.pages.find((page) => page.name === "Home");
+      assert.ok(home);
+      assert.strictEqual(home.href, "/ariya/phantomjs/wiki");
+      assert.strictEqual(home.updatedAt, "2018-02-08T18:57:56Z");
+    });
+
+    it("should redirect to the repository when it has no wiki", async () => {
+      await assert.rejects(getGitHubWiki("git", "git"), (err) => {
+        assert(err instanceof RedirectError, "error should be a RedirectError");
+        assert.strictEqual(err.location, "/git/git");
+        return true;
+      });
+    });
+  });
+
   describe("owner pages", () => {
     it("should fetch user profile", async () => {
       const data = await getGitHubOwner("torvalds");
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index c02cb5e..6461b19 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -1,5 +1,5 @@
 import { parseDocument } from "htmlparser2";
-import { type AnyNode, type Element } from "domhandler";
+import { isText, type AnyNode, type Element } from "domhandler";
 import { getAttributeValue, getInnerHTML, textContent } from "domutils";
 import * as cssSelect from "css-select";
 
@@ -352,6 +352,26 @@ export interface GitHubRelease extends GitHubCommon {
   release: Release;
 }
 
+export interface WikiPage {
+  name: string;
+  href: string;
+  updatedAt?: string;
+}
+
+export interface GitHubWiki extends GitHubCommon {
+  title: string;
+  updatedBy: string | null;
+  updatedAt: string | null;
+  bodyHtml: string;
+  sidebarHtml: string | null;
+  footerHtml: string | null;
+  pages: WikiPage[];
+}
+
+export interface GitHubWikiPages extends GitHubCommon {
+  pages: WikiPage[];
+}
+
 interface OwnerRepo {
   owner: string;
   name: string;
@@ -1109,3 +1129,96 @@ export async function getGitHubOwner(owner: string): Promise<GitHubOwner> {
   // Not a profile page (marketing pages like github.com/features, etc.).
   throw new RedirectError(`https://github.com/${owner}`);
 }
+
+// The text directly inside an element, ignoring any nested elements' text.
+function directText(elem: Element): string {
+  return elem.children
+    .filter(isText)
+    .map((node) => node.data)
+    .join("")
+    .replace(/\s+/g, " ")
+    .trim();
+}
+
+// The sidebar holds every wiki page, which are revealed by JavaScript.
+function parseWikiSidebarPages(document: AnyNode): WikiPage[] {
+  const pages: WikiPage[] = [];
+  const anchors = cssSelect.selectAll(
+    ".js-wiki-sidebar-page-container > div > a[href]",
+    document,
+  ) as unknown as Element[];
+  for (const anchor of anchors) {
+    const href = getAttributeValue(anchor, "href");
+    const name = textContent(anchor).replace(/\s+/g, " ").trim();
+    if (href && name) pages.push({ name, href });
+  }
+  return pages;
+}
+
+function wikiHtml(selector: string, document: AnyNode): string | null {
+  const elem = cssSelect.selectOne(selector, document);
+  if (!elem) return null;
+  const html = getInnerHTML(elem).trim();
+  return html || null;
+}
+
+export async function getGitHubWiki(
+  owner: string,
+  repo: string,
+  // Empty for the wiki's home page, which GitHub serves off /wiki itself.
+  page = "",
+  // A revision SHA renders that older version of the page.
+  oid = "",
+): Promise<GitHubWiki> {
+  // Page names keep characters that are not path-safe (C#, [DEV] ...), so the
+  // slug has to be re-encoded rather than interpolated the way a ref is.
+  const path = [`${owner}/${repo}/wiki`, page && encodeURIComponent(page), oid]
+    .filter(Boolean)
+    .join("/");
+  const html = await fetchGitHubPage(path);
+  const document = parseDocument(html);
+
+  const heading = cssSelect.selectOne("h1.gh-header-title", document) as Element | null;
+  if (!heading) {
+    throw new Error("Could not find the wiki page heading");
+  }
+
+  // "<author> edited this page <date> · <n> revisions", where the author is the
+  // only part not wrapped in an element of its own.
+  const meta = cssSelect.selectOne(".gh-header-meta", document) as Element | null;
+  const updatedBy = meta ? (/^(.+?) edited this page\b/.exec(directText(meta))?.[1] ?? null) : null;
+  const time = meta ? cssSelect.selectOne("relative-time", meta) : null;
+
+  return {
+    repo: { owner, name: repo },
+    title: directText(heading),
+    updatedBy,
+    updatedAt: time ? getAttributeValue(time, "datetime") || null : null,
+    bodyHtml: wikiHtml("#wiki-body .markdown-body", document) ?? "",
+    // Wikis can define _Sidebar and _Footer pages, rendered alongside every page.
+    sidebarHtml: wikiHtml(".wiki-custom-sidebar", document),
+    footerHtml: wikiHtml("#wiki-footer .markdown-body", document),
+    pages: parseWikiSidebarPages(document),
+  };
+}
+
+export async function getGitHubWikiPages(owner: string, repo: string): Promise<GitHubWikiPages> {
+  const html = await fetchGitHubPage(`${owner}/${repo}/wiki/_pages`);
+  const document = parseDocument(html);
+
+  const pages: WikiPage[] = [];
+  for (const row of cssSelect.selectAll(
+    "#wiki-content li.Box-row",
+    document,
+  ) as unknown as Element[]) {
+    const anchor = cssSelect.selectOne("a[href]", row) as Element | null;
+    const href = anchor ? getAttributeValue(anchor, "href") : undefined;
+    const name = anchor ? textContent(anchor).replace(/\s+/g, " ").trim() : "";
+    if (!href || !name) continue;
+
+    const time = cssSelect.selectOne("relative-time", row) as Element | null;
+    pages.push({ name, href, updatedAt: time ? getAttributeValue(time, "datetime") : undefined });
+  }
+
+  return { repo: { owner, name: repo }, pages };
+}
diff --git a/cgithub/views/_wikiNav.eta b/cgithub/views/_wikiNav.eta
new file mode 100644
index 0000000..cc177c7
--- /dev/null
+++ b/cgithub/views/_wikiNav.eta
@@ -0,0 +1,4 @@
+<nav>
+  <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/wiki">Wiki home</a>
+  <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/wiki/_pages">Pages</a>
+</nav>
diff --git a/cgithub/views/layout.eta b/cgithub/views/layout.eta
index 6d1b052..1ac82cc 100644
--- a/cgithub/views/layout.eta
+++ b/cgithub/views/layout.eta
@@ -17,6 +17,7 @@ git clone git@github.com:<%= it.repo.owner %>/<%= it.repo.name %>.git</code></pr
       <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/issues">Issues</a>
       <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/pulls">Pull requests</a>
       <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/releases">Releases</a>
+      <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/wiki">Wiki</a>
       <% if (it.branch) { %>
         <a href="/<%= it.repo.owner %>/<%= it.repo.name %>/commits/<%= it.branch %><%= it.path && it.path !== '/' ? '/' + it.path : '' %>">Commits</a>
       <% } %>
diff --git a/cgithub/views/wiki.eta b/cgithub/views/wiki.eta
new file mode 100644
index 0000000..d6047e1
--- /dev/null
+++ b/cgithub/views/wiki.eta
@@ -0,0 +1,37 @@
+<% layout("./layout.eta"); it.path = "wiki/" + it.title %>
+
+<%~ include("./_wikiNav.eta", it) %>
+
+<div class="box wiki">
+  <header>
+    <span><%= it.title %></span>
+    <% if (it.updatedBy || it.updatedAt) { %>
+      <span class="timestamp">
+        <% if (it.updatedBy) { %><%= it.updatedBy %> edited this page<% } %>
+        <% if (it.updatedAt) { %><time title="<%= it.updatedAt %>"><%= it.updatedAt.split("T")[0] %></time><% } %>
+      </span>
+    <% } %>
+  </header>
+  <%~ it.bodyHtml %>
+  <% if (it.footerHtml) { %>
+    <footer><%~ it.footerHtml %></footer>
+  <% } %>
+</div>
+
+<% if (it.sidebarHtml) { %>
+  <div class="box wiki">
+    <header><span>Sidebar</span></header>
+    <%~ it.sidebarHtml %>
+  </div>
+<% } %>
+
+<% if (it.pages.length > 0) { %>
+  <details class="box wiki-pages">
+    <summary><%= it.pages.length %> page<%= it.pages.length === 1 ? "" : "s" %></summary>
+    <ul>
+      <% for (const page of it.pages) { %>
+        <li><a href="<%= page.href %>"><%= page.name %></a></li>
+      <% } %>
+    </ul>
+  </details>
+<% } %>
diff --git a/cgithub/views/wikiPages.eta b/cgithub/views/wikiPages.eta
new file mode 100644
index 0000000..58d6d21
--- /dev/null
+++ b/cgithub/views/wikiPages.eta
@@ -0,0 +1,16 @@
+<% layout("./layout.eta"); it.path = "wiki/_pages" %>
+
+<%~ include("./_wikiNav.eta", it) %>
+
+<table>
+  <% for (const page of it.pages) { %>
+    <tr>
+      <td><a href="<%= page.href %>"><%= page.name %></a></td>
+      <td class="timestamp">
+        <% if (page.updatedAt) { %>
+          <time title="<%= page.updatedAt %>"><%= page.updatedAt.split("T")[0] %></time>
+        <% } %>
+      </td>
+    </tr>
+  <% } %>
+</table>
-- 
cgit v1.3.1


From 5bfe8b345d2ee92111a2363321a248adb4f5195a Mon Sep 17 00:00:00 2001
From: Joe Mou <dev@mou.fo>
Date: Thu, 3 Sep 2026 23:08:08 -0400
Subject: Handle route errors in middleware

Every page route wrapped its scrape in tryRender, which awaited the
promise, rendered the template on success, and turned a failure into a
redirect or an error page. Hono's onError does that mapping for all of
them at once, so handlers now await their scrape inline and both
wrappers are gone. Fragments under /api/ still fail as an empty body
rather than a page; onError checks for them first, since a fragment
turns a redirect into a 404 instead of following it.

RedirectError becomes an abstract base with InternalRedirectError and
ExternalRedirectError under it, so the handler dispatches on the type
instead of sniffing the location for an https:// prefix.

One behavior change: a thrown value that is not an Error no longer
renders error.eta, because Hono only routes Error instances to onError.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01QERwoRJhuqMhKzzCUW4Skv
---
 cgithub/src/app.ts          | 104 +++++++++++++++++---------------------------
 cgithub/src/scraper.test.ts |  13 +++---
 cgithub/src/scraper.ts      |  12 +++--
 3 files changed, 56 insertions(+), 73 deletions(-)

(limited to 'cgithub/src/scraper.ts')

diff --git a/cgithub/src/app.ts b/cgithub/src/app.ts
index bbe07fc..2fa5832 100644
--- a/cgithub/src/app.ts
+++ b/cgithub/src/app.ts
@@ -18,7 +18,9 @@ import {
   getGitHubTree,
   getGitHubWiki,
   getGitHubWikiPages,
+  ExternalRedirectError,
   GitHubHTTPError,
+  InternalRedirectError,
   RedirectError,
 } from "./scraper.ts";
 
@@ -64,73 +66,53 @@ export function createApp(eta: Eta) {
     return redirectTo(c, githubUrlFor(c));
   }
 
-  app.get("/", async (c) => {
-    return c.render("home.eta");
-  });
-
-  async function tryRender<T extends object>(c: Context, template: string, promise: Promise<T>) {
-    let data: T;
-    try {
-      data = await promise;
-    } catch (e) {
-      if (e instanceof RedirectError) {
-        if (e.location.startsWith("https://")) {
-          return redirectTo(c, e.location);
-        }
-        return c.redirect(e.location);
-      } else if (e instanceof GitHubHTTPError) {
-        c.status(e.status as StatusCode);
-        const message = `GitHub responded with HTTP ${e.status} ${e.message}`;
-        return c.render("error.eta", { title: e.message, message });
-      } else {
-        c.status(500);
-        return c.render("error.eta", { message: "" + e });
-      }
-    }
-    return c.render(template, data);
-  }
-
-  // For fragments fetched asynchronously by client-side JS (see public/static/refs.js):
-  // errors render as an empty body rather than the full error.eta page.
-  async function tryRenderFragment<T extends object>(
-    c: Context,
-    template: string,
-    promise: Promise<T>,
-  ) {
-    let data: T;
-    try {
-      data = await promise;
-    } catch (e) {
-      const status =
+  app.onError((e, c) => {
+    // Fragments fetched asynchronously by client-side JS (see public/static/refs.js)
+    // fail as an empty body rather than the full error.eta page.
+    if (c.req.path.startsWith("/api/")) {
+      c.status(
         e instanceof GitHubHTTPError
           ? (e.status as StatusCode)
           : e instanceof RedirectError
             ? 404
-            : 500;
-      c.status(status);
+            : 500,
+      );
       return c.body(null);
     }
-    return c.render(template, data);
-  }
+
+    if (e instanceof InternalRedirectError) return c.redirect(e.location);
+    if (e instanceof ExternalRedirectError) return redirectTo(c, e.location);
+    if (e instanceof GitHubHTTPError) {
+      c.status(e.status as StatusCode);
+      const message = `GitHub responded with HTTP ${e.status} ${e.message}`;
+      return c.render("error.eta", { title: e.message, message });
+    }
+    c.status(500);
+    return c.render("error.eta", { message: "" + e });
+  });
+
+  app.get("/", async (c) => {
+    return c.render("home.eta");
+  });
 
   app.get("/:owner", async (c) => {
     const { owner } = c.req.param();
-    return tryRender(c, "owner.eta", getGitHubOwner(owner));
+    return c.render("owner.eta", await getGitHubOwner(owner));
   });
 
   app.get("/:owner/:repo", async (c) => {
     const { owner, repo } = c.req.param();
-    return tryRender(c, "repo.eta", getGitHubRepo(owner, repo));
+    return c.render("repo.eta", await getGitHubRepo(owner, repo));
   });
 
   app.get("/:owner/:repo/tree/:branch/:path{.*}?", async (c) => {
     const { owner, repo, branch, path = "" } = c.req.param();
-    return tryRender(c, "tree.eta", getGitHubTree(owner, repo, branch, path));
+    return c.render("tree.eta", await getGitHubTree(owner, repo, branch, path));
   });
 
   app.get("/:owner/:repo/blob/:branch/:path{.*}", async (c) => {
     const { owner, repo, branch, path } = c.req.param();
-    return tryRender(c, "blob.eta", getGitHubBlob(owner, repo, branch, path));
+    return c.render("blob.eta", await getGitHubBlob(owner, repo, branch, path));
   });
 
   app.get("/:owner/:repo/raw/:branch/:path{.*}", async (c) => {
@@ -168,12 +150,12 @@ export function createApp(eta: Eta) {
   // page can never be mistaken for one.
   app.get("/:owner/:repo/wiki", async (c) => {
     const { owner, repo } = c.req.param();
-    return tryRender(c, "wiki.eta", getGitHubWiki(owner, repo));
+    return c.render("wiki.eta", await getGitHubWiki(owner, repo));
   });
 
   app.get("/:owner/:repo/wiki/_pages", async (c) => {
     const { owner, repo } = c.req.param();
-    return tryRender(c, "wikiPages.eta", getGitHubWikiPages(owner, repo));
+    return c.render("wikiPages.eta", await getGitHubWikiPages(owner, repo));
   });
 
   // An older revision of a page renders exactly like its current one.
@@ -182,7 +164,7 @@ export function createApp(eta: Eta) {
     if (page.startsWith("_") || !/^[0-9a-f]{40}$/.test(oid)) {
       return redirectToGitHub(c);
     }
-    return tryRender(c, "wiki.eta", getGitHubWiki(owner, repo, page, oid));
+    return c.render("wiki.eta", await getGitHubWiki(owner, repo, page, oid));
   });
 
   app.get("/:owner/:repo/wiki/:page", async (c) => {
@@ -190,7 +172,7 @@ export function createApp(eta: Eta) {
     if (page.startsWith("_")) {
       return redirectToGitHub(c);
     }
-    return tryRender(c, "wiki.eta", getGitHubWiki(owner, repo, page));
+    return c.render("wiki.eta", await getGitHubWiki(owner, repo, page));
   });
 
   app.get("/:owner/:repo/issues", async (c) => {
@@ -203,13 +185,13 @@ export function createApp(eta: Eta) {
       return c.redirect(`/${owner}/${repo}/issues?q=${encodeURIComponent(`is:issue ${q}`)}`);
     }
 
-    return tryRender(c, "issues.eta", getGitHubIssues(owner, repo, { q }));
+    return c.render("issues.eta", await getGitHubIssues(owner, repo, { q }));
   });
 
   app.get("/:owner/:repo/pulls", async (c) => {
     const { owner, repo } = c.req.param();
     const q = c.req.query("q");
-    return tryRender(c, "pulls.eta", getGitHubPulls(owner, repo, { q }));
+    return c.render("pulls.eta", await getGitHubPulls(owner, repo, { q }));
   });
 
   app.get("/:owner/:repo/search", async (c) => {
@@ -229,7 +211,7 @@ export function createApp(eta: Eta) {
 
   app.get("/:owner/:repo/commits/:branch/:path{.*}?", async (c) => {
     const { owner, repo, branch, path = "" } = c.req.param();
-    return tryRender(c, "commits.eta", getGitHubCommits(owner, repo, branch, path));
+    return c.render("commits.eta", await getGitHubCommits(owner, repo, branch, path));
   });
 
   // Internal endpoints fetched asynchronously by client-side JS, not meant to
@@ -246,32 +228,28 @@ export function createApp(eta: Eta) {
     if (view !== "tree" && view !== "blob" && view !== "commits") {
       return c.notFound();
     }
-    const refs = getGitHubRefs(owner, repo, type).then((data) => ({ ...data, view, path }));
-    return tryRenderFragment(c, "_refs.eta", refs);
+    const refs = await getGitHubRefs(owner, repo, type);
+    return c.render("_refs.eta", { ...refs, view, path });
   });
 
   app.get("/api/:owner/:repo/latest-commit/:branch/:path{.*}?", async (c) => {
     const { owner, repo, branch, path = "" } = c.req.param();
-    return tryRenderFragment(
-      c,
-      "_latestCommit.eta",
-      getGitHubLatestCommit(owner, repo, branch, path),
-    );
+    return c.render("_latestCommit.eta", await getGitHubLatestCommit(owner, repo, branch, path));
   });
 
   app.get("/api/:owner/:repo/sidebar", async (c) => {
     const { owner, repo } = c.req.param();
-    return tryRenderFragment(c, "_sidebar.eta", getGitHubSidebar(owner, repo));
+    return c.render("_sidebar.eta", await getGitHubSidebar(owner, repo));
   });
 
   app.get("/:owner/:repo/releases/tag/:tag", async (c) => {
     const { owner, repo, tag } = c.req.param();
-    return tryRender(c, "release.eta", getGitHubRelease(owner, repo, tag));
+    return c.render("release.eta", await getGitHubRelease(owner, repo, tag));
   });
 
   app.get("/:owner/:repo/releases", async (c) => {
     const { owner, repo } = c.req.param();
-    return tryRender(c, "releases.eta", getGitHubReleases(owner, repo));
+    return c.render("releases.eta", await getGitHubReleases(owner, repo));
   });
 
   app.all("*", async (c) => {
diff --git a/cgithub/src/scraper.test.ts b/cgithub/src/scraper.test.ts
index f6be2b6..8dcb0b8 100644
--- a/cgithub/src/scraper.test.ts
+++ b/cgithub/src/scraper.test.ts
@@ -16,8 +16,9 @@ import {
   getGitHubTree,
   getGitHubWiki,
   getGitHubWikiPages,
+  ExternalRedirectError,
   GitHubHTTPError,
-  RedirectError,
+  InternalRedirectError,
 } from "./scraper.ts";
 
 describe("GitHub scraper", () => {
@@ -215,7 +216,7 @@ describe("GitHub scraper", () => {
       const oid = "b39c421b98f49d83ae50ec502c3ddfc3bf28f2c6";
 
       await assert.rejects(getGitHubTree("actions", "deploy-pages", oid, "src/index.js"), (err) => {
-        assert(err instanceof RedirectError, "error should be a RedirectError");
+        assert(err instanceof InternalRedirectError, "error should be an InternalRedirectError");
         assert.strictEqual(err.location, `/actions/deploy-pages/blob/${oid}/src/index.js`);
         return true;
       });
@@ -490,7 +491,7 @@ describe("GitHub scraper", () => {
 
     it("should redirect to the repository when it has no wiki", async () => {
       await assert.rejects(getGitHubWiki("git", "git"), (err) => {
-        assert(err instanceof RedirectError, "error should be a RedirectError");
+        assert(err instanceof InternalRedirectError, "error should be an InternalRedirectError");
         assert.strictEqual(err.location, "/git/git");
         return true;
       });
@@ -571,7 +572,7 @@ describe("GitHub scraper", () => {
 
     it("should redirect to GitHub for non-profile pages", async () => {
       await assert.rejects(getGitHubOwner("features"), (err) => {
-        assert(err instanceof RedirectError, "error should be a RedirectError");
+        assert(err instanceof ExternalRedirectError, "error should be an ExternalRedirectError");
         assert.strictEqual(err.location, "https://github.com/features");
         return true;
       });
@@ -599,7 +600,7 @@ describe("GitHub scraper", () => {
       await assert.rejects(
         getGitHubBlob("earendil-works", "pi", "main", "packages/agent"),
         (err) => {
-          assert(err instanceof RedirectError, "error should be a RedirectError");
+          assert(err instanceof InternalRedirectError, "error should be an InternalRedirectError");
           assert.strictEqual(err.location, "/earendil-works/pi/tree/main/packages/agent");
           return true;
         },
@@ -610,7 +611,7 @@ describe("GitHub scraper", () => {
       await assert.rejects(
         getGitHubTree("earendil-works", "pi", "main", "packages/agent/package.json"),
         (err) => {
-          assert(err instanceof RedirectError, "error should be a RedirectError");
+          assert(err instanceof InternalRedirectError, "error should be an InternalRedirectError");
           assert.strictEqual(
             err.location,
             "/earendil-works/pi/blob/main/packages/agent/package.json",
diff --git a/cgithub/src/scraper.ts b/cgithub/src/scraper.ts
index 6461b19..fb48700 100644
--- a/cgithub/src/scraper.ts
+++ b/cgithub/src/scraper.ts
@@ -13,16 +13,20 @@ export class GitHubHTTPError extends Error {
   }
 }
 
-export class RedirectError extends Error {
+export abstract class RedirectError extends Error {
   location: string;
 
   constructor(location: string) {
     super(`Redirect to ${location}`);
-    this.name = "RedirectError";
+    this.name = this.constructor.name;
     this.location = location;
   }
 }
 
+export class InternalRedirectError extends RedirectError {}
+
+export class ExternalRedirectError extends RedirectError {}
+
 interface TreeItem {
   contentType: "directory" | "file";
   name: string;
@@ -423,7 +427,7 @@ async function fetchGitHubResponse(path: string, accept: string): Promise<Respon
     if (location) {
       const url = new URL(location);
       if (url.hostname === "github.com") {
-        throw new RedirectError(url.pathname);
+        throw new InternalRedirectError(url.pathname);
       }
     }
   }
@@ -1127,7 +1131,7 @@ export async function getGitHubOwner(owner: string): Promise<GitHubOwner> {
     return parseOrganization(document, owner);
   }
   // Not a profile page (marketing pages like github.com/features, etc.).
-  throw new RedirectError(`https://github.com/${owner}`);
+  throw new ExternalRedirectError(`https://github.com/${owner}`);
 }
 
 // The text directly inside an element, ignoring any nested elements' text.
-- 
cgit v1.3.1