From 0198445ee47aa651336e8816220554c36ca3e220 Mon Sep 17 00:00:00 2001
From: Abdessamad Derraz <3028866+Abdess@users.noreply.github.com>
Date: Fri, 16 Oct 2026 15:56:50 +0200
Subject: [PATCH] fix: stop the core-info scrape on a failed fetch

---
 scripts/scraper/coreinfo_scraper.py | 13 +++++++++----
 tests/test_scraper_contract.py      | 13 +++++++++++++
 2 files changed, 22 insertions(+), 4 deletions(-)

diff --git a/scripts/scraper/coreinfo_scraper.py b/scripts/scraper/coreinfo_scraper.py
index 4a33ff12..d9e34925 100644
--- a/scripts/scraper/coreinfo_scraper.py
+++ b/scripts/scraper/coreinfo_scraper.py
@@ -228,7 +228,12 @@ class Scraper(BaseScraper):
             raise ConnectionError(f"Failed to list core-info files: {e}") from e
 
     def _fetch_info_file(self, filename: str) -> dict:
-        """Fetch and parse a single .info file."""
+        """Fetch and parse a single .info file.
+
+        A file the listing names and the fetch cannot read is an error: read
+        as empty, the core dropped out of the result and --compare-db
+        counted none of its firmware missing.
+        """
         url = f"{RAW_BASE}/{filename}"
         try:
             req = urllib.request.Request(
@@ -236,9 +241,9 @@ class Scraper(BaseScraper):
             )
             with urllib.request.urlopen(req, timeout=15) as resp:
                 content = resp.read().decode("utf-8")
-            return _parse_info_file(content)
-        except (urllib.error.URLError, urllib.error.HTTPError):
-            return {}
+        except (urllib.error.URLError, TimeoutError) as e:
+            raise ConnectionError(f"Failed to fetch {url}: {e}") from e
+        return _parse_info_file(content)
 
     def fetch_requirements(self) -> list[BiosRequirement]:
         """Fetch firmware requirements from all core .info files."""
diff --git a/tests/test_scraper_contract.py b/tests/test_scraper_contract.py
index 5abff128..304e64a4 100644
--- a/tests/test_scraper_contract.py
+++ b/tests/test_scraper_contract.py
@@ -284,6 +284,19 @@ class UnreadableReleaseStopsTheScrape(unittest.TestCase):
         ):
             libretro_scraper.Scraper()._fetch_core_metadata()
 
+    def test_a_core_info_file_the_forge_refuses_raises(self):
+        """Read as empty, the core left the result and --compare-db counted
+        none of its firmware missing."""
+        from scraper import coreinfo_scraper  # noqa: PLC0415
+
+        scraper = coreinfo_scraper.Scraper()
+        with mock.patch.object(
+            scraper, "_fetch_info_list", return_value=["mednafen_psx_libretro.info"]
+        ), mock.patch("urllib.request.urlopen", self._refuse), self.assertRaises(
+            ConnectionError
+        ):
+            scraper.fetch_requirements()
+
 
 
 class EveryGitHubApiCallerSendsTheToken(unittest.TestCase):