fix: publish feeds atomically

repub/postprocessing.py

@@ -0,0 +1,47 @@
+from __future__ import annotations
+
+import os
+from contextlib import suppress
+from pathlib import Path
+from xml.etree import ElementTree
+
+from repub.config import feed_output_path, staged_feed_output_path
+
+
+def publish_staged_feed(*, out_dir: Path, feed_slug: str) -> Path:
+    staged_path = staged_feed_output_path(out_dir=out_dir, feed_slug=feed_slug)
+    public_path = feed_output_path(out_dir=out_dir, feed_slug=feed_slug)
+
+    public_path.parent.mkdir(parents=True, exist_ok=True)
+    _validate_staged_feed(staged_path)
+    _fsync_file(staged_path)
+    os.replace(staged_path, public_path)
+    _fsync_directory(public_path.parent)
+    return public_path
+
+
+def _fsync_file(path: Path) -> None:
+    with path.open("rb") as handle:
+        os.fsync(handle.fileno())
+
+
+def _validate_staged_feed(path: Path) -> None:
+    try:
+        root = ElementTree.parse(path).getroot()
+    except ElementTree.ParseError as error:
+        raise ValueError(f"Staged feed is not well-formed XML: {path}") from error
+
+    if root.tag != "rss":
+        raise ValueError(f"Staged feed is not an RSS document: {path}")
+    if root.find("channel") is None:
+        raise ValueError(f"Staged feed is missing an RSS channel: {path}")
+
+
+def _fsync_directory(path: Path) -> None:
+    flags = os.O_RDONLY | getattr(os, "O_DIRECTORY", 0)
+    with suppress(OSError):
+        fd = os.open(path, flags)
+        try:
+            os.fsync(fd)
+        finally:
+            os.close(fd)