From 88d7ca4814c3bdb59b7ee0e672677b4ccf58e8f4 Mon Sep 17 00:00:00 2001
From: Abdessamad Derraz <3028866+Abdess@users.noreply.github.com>
Date: Sat, 10 Oct 2026 12:08:52 +0200
Subject: [PATCH] fix: stop a retrodeck scrape on a failed fetch

---
 scripts/scraper/retrodeck_scraper.py   | 22 +++++---
 tests/test_retrodeck_scraper_errors.py | 71 ++++++++++++++++++++++++++
 2 files changed, 87 insertions(+), 6 deletions(-)
 create mode 100644 tests/test_retrodeck_scraper_errors.py

diff --git a/scripts/scraper/retrodeck_scraper.py b/scripts/scraper/retrodeck_scraper.py
index e5a580a2..e41bfa48 100644
--- a/scripts/scraper/retrodeck_scraper.py
+++ b/scripts/scraper/retrodeck_scraper.py
@@ -248,7 +248,9 @@ class Scraper(BaseScraper):
             raise ConnectionError(f"Failed to fetch component tree: {e}") from e
 
         if tree.get("truncated"):
-            print("  WARNING: GitHub tree response truncated", file=sys.stderr)
+            # A partial tree lists some components: writing the YAML from it
+            # drops the others' systems without a word.
+            raise ConnectionError("GitHub tree response truncated")
 
         component_dirs = [
             item["path"]
@@ -260,16 +262,24 @@ class Scraper(BaseScraper):
         for comp in sorted(component_dirs):
             url = f"{RAW_BASE}/{comp}/component_manifest.json"
             print(f"  {comp} ...", file=sys.stderr, end="", flush=True)
+            # Only a 404 says the component has no manifest. A 429, a 503 or
+            # a dropped connection is no answer: skipping it wrote
+            # retrodeck.yml without that component's systems, exit code 0.
             try:
                 req = urllib.request.Request(url, headers=headers)
                 with urllib.request.urlopen(req, timeout=15) as resp:
                     data = json.loads(resp.read().decode())
-                manifests.append((comp, data))
-                print(" ok", file=sys.stderr)
-            except (urllib.error.HTTPError, urllib.error.URLError):
-                print(" skip", file=sys.stderr)
+            except urllib.error.HTTPError as e:
+                if e.code == 404:
+                    print(" no manifest", file=sys.stderr)
+                    continue
+                raise ConnectionError(f"{url}: HTTP {e.code}") from e
+            except urllib.error.URLError as e:
+                raise ConnectionError(f"{url}: {e.reason}") from e
             except json.JSONDecodeError as e:
-                print(f" parse error: {e}", file=sys.stderr)
+                raise ValueError(f"{url}: not JSON: {e}") from e
+            manifests.append((comp, data))
+            print(" ok", file=sys.stderr)
         return manifests
 
     def _fetch_local_manifests(self) -> list[tuple[str, dict]]:
diff --git a/tests/test_retrodeck_scraper_errors.py b/tests/test_retrodeck_scraper_errors.py
new file mode 100644
index 00000000..08db2667
--- /dev/null
+++ b/tests/test_retrodeck_scraper_errors.py
@@ -0,0 +1,71 @@
+"""A component manifest that could not be read is not a missing one.
+
+The RetroDECK scraper printed "skip" on any HTTP or network error and wrote
+retrodeck.yml without that component: a 503 on pcsx2's manifest removed the
+PS2 BIOS and the pcsx2 standalone core, exit code 0.
+"""
+
+from __future__ import annotations
+
+import io
+import json
+import sys
+import unittest
+import urllib.error
+from pathlib import Path
+from unittest import mock
+
+REPO_ROOT = Path(__file__).resolve().parent.parent
+sys.path.insert(0, str(REPO_ROOT))
+
+from scripts.scraper.retrodeck_scraper import Scraper  # noqa: E402
+
+TREE = {"tree": [{"path": "pcsx2", "type": "tree"}, {"path": "ppsspp", "type": "tree"}]}
+
+
+class _Response(io.BytesIO):
+    def __enter__(self):
+        return self
+
+    def __exit__(self, *exc):
+        return False
+
+
+def _opener(failure):
+    def urlopen(req, timeout=0):
+        url = req.full_url
+        if "api.github.com" in url:
+            return _Response(json.dumps(TREE).encode())
+        if "pcsx2" in url and failure is not None:
+            raise failure
+        return _Response(json.dumps({"ppsspp": {"bios": []}}).encode())
+    return urlopen
+
+
+class ManifestErrors(unittest.TestCase):
+    def fetch(self, failure):
+        with mock.patch("urllib.request.urlopen", _opener(failure)):
+            return Scraper()._fetch_remote_manifests()
+
+    def test_a_server_error_stops_the_scrape(self):
+        error = urllib.error.HTTPError("u", 503, "unavailable", {}, None)
+        with self.assertRaises(ConnectionError):
+            self.fetch(error)
+
+    def test_a_dropped_connection_stops_the_scrape(self):
+        with self.assertRaises(ConnectionError):
+            self.fetch(urllib.error.URLError("reset"))
+
+    def test_a_404_is_a_component_without_a_manifest(self):
+        error = urllib.error.HTTPError("u", 404, "not found", {}, None)
+        self.assertEqual([name for name, _ in self.fetch(error)], ["ppsspp"])
+
+    def test_a_truncated_tree_stops_the_scrape(self):
+        TREE["truncated"] = True
+        self.addCleanup(TREE.pop, "truncated")
+        with self.assertRaises(ConnectionError):
+            self.fetch(None)
+
+
+if __name__ == "__main__":
+    unittest.main()