feat(blog): add file-based blog with dynamic slugs, MDX content and layout shell

- Introduced blog routing using Next.js App Router
- Implemented dynamic [slug] pages for blog posts
- Added MDX-based content loading via lib/posts
- Integrated shared TopBar layout with navigation
- Established clear content, lib and component separation

apps/public-web/node_modules/next/dist/esm/lib/recursive-delete.js

@@ -0,0 +1,81 @@
+import * as fs from 'node:fs';
+import { join } from 'node:path';
+import isError from './is-error';
+import { wait } from './wait';
+// We use an exponential backoff. See the unit test for example values.
+//
+// - Node's `fs` module uses a linear backoff, starting with 100ms.
+// - Rust tries 64 times with only a `thread::yield_now` in between.
+//
+// We want something more aggressive, as `recursiveDelete` is in the critical
+// path of `next dev` and `next build` startup.
+const INITIAL_RETRY_MS = 8;
+const MAX_RETRY_MS = 64;
+const MAX_RETRIES = 6;
+/**
+ * Used in unit test.
+ * @ignore
+ */ export function calcBackoffMs(attempt) {
+    return Math.min(INITIAL_RETRY_MS * Math.pow(2, attempt), MAX_RETRY_MS);
+}
+function unlinkPath(p, isDir = false, attempt = 0) {
+    try {
+        if (isDir) {
+            fs.rmdirSync(p);
+        } else {
+            fs.unlinkSync(p);
+        }
+    } catch (e) {
+        const code = isError(e) && e.code;
+        if ((code === 'EBUSY' || code === 'ENOTEMPTY' || code === 'EPERM' || code === 'EMFILE') && attempt < MAX_RETRIES) {
+            // retrying is unlikely to succeed on POSIX platforms, but Windows can
+            // fail due to temporarily-open files
+            return (async ()=>{
+                await wait(calcBackoffMs(attempt));
+                return unlinkPath(p, isDir, attempt + 1);
+            })();
+        }
+        if (code === 'ENOENT') {
+            return;
+        }
+        throw e;
+    }
+}
+/**
+ * Recursively delete directory contents.
+ *
+ * This is used when cleaning the `distDir`, and is part of the critical path
+ * for starting the server, so we use synchronous file IO, as we're always
+ * blocked on it anyways.
+ *
+ * Despite using sync IO, the function signature is still `async` because we
+ * asynchronously perform retries.
+ */ export async function recursiveDeleteSyncWithAsyncRetries(/** Directory to delete the contents of */ dir, /** Exclude based on relative file path */ exclude, /** Relative path to the directory being deleted, used for exclude */ previousPath = '') {
+    let result;
+    try {
+        result = fs.readdirSync(dir, {
+            withFileTypes: true
+        });
+    } catch (e) {
+        if (isError(e) && e.code === 'ENOENT') {
+            return;
+        }
+        throw e;
+    }
+    await Promise.all(result.map(async (part)=>{
+        const absolutePath = join(dir, part.name);
+        const pp = join(previousPath, part.name);
+        const isNotExcluded = !exclude || !exclude.test(pp);
+        if (isNotExcluded) {
+            // Note: readdir does not follow symbolic links, that's good: we want to
+            // delete the links and not the destination.
+            let isDirectory = part.isDirectory();
+            if (isDirectory) {
+                await recursiveDeleteSyncWithAsyncRetries(absolutePath, exclude, pp);
+            }
+            return unlinkPath(absolutePath, isDirectory);
+        }
+    }));
+}
+
+//# sourceMappingURL=recursive-delete.js.map