From 4522f3ba01f7df6d8a5ef98ba20edb181856f5d4 Mon Sep 17 00:00:00 2001
From: Abdessamad Derraz <3028866+Abdess@users.noreply.github.com>
Date: Tue, 27 Oct 2026 05:55:10 +0100
Subject: [PATCH] fix: keep the paths of a tree no folder lists

A repository that answers 404 (TeknoParrotUI) was listed folder by folder,
every listing failed, and the tree read as holding nothing: a fetch without
its cached tree dropped its recorded paths. It is now unread, so the record
keeps what it held.
---
 scripts/sitesources.py          |  8 ++++++--
 tests/test_site_r4_reference.py | 13 +++++++++++++
 2 files changed, 19 insertions(+), 2 deletions(-)

diff --git a/scripts/sitesources.py b/scripts/sitesources.py
index 5974e3c4e..c0270a5e9 100644
--- a/scripts/sitesources.py
+++ b/scripts/sitesources.py
@@ -272,10 +272,13 @@ def _listed_directories(
     """The cited paths found by listing their directories one by one, for a
     tree the forge truncates (aps3e vendors RPCS3 whole). A directory that
     cannot be listed confirms none of its paths. None off GitHub, the one
-    forge with that call."""
+    forge with that call, and None when no directory could be listed: the
+    tree was not read, so its record is kept (TeknoParrotUI answers 404, and
+    each fetch without its cached tree dropped its four paths)."""
     if repo.family != "github":
         return None
     found: set[str] = set()
+    listed = False
     for folder in sorted({path.rpartition("/")[0] for path in paths} - {""}):
         names = None
         for index, cache_dir in enumerate(cache_dirs):
@@ -286,8 +289,9 @@ def _listed_directories(
                 names = None
             if names is not None:
                 break
+        listed = listed or names is not None
         found.update(f"{folder}/{name}" for name in names or ())
-    return frozenset(found)
+    return frozenset(found) if listed else None
 
 
 def build_source_paths(
diff --git a/tests/test_site_r4_reference.py b/tests/test_site_r4_reference.py
index b11a9cba0..6100cae49 100644
--- a/tests/test_site_r4_reference.py
+++ b/tests/test_site_r4_reference.py
@@ -232,6 +232,19 @@ class SourceLinksOnlyWhereTheFileIs(unittest.TestCase):
         self.assertEqual(held, {"github.com/o/r@" + "e" * 40: ["src/a.c", "src/c.c"]})
         self.assertEqual(unread, [])
 
+    def test_a_tree_no_directory_answers_for_stays_unread(self):
+        # TeknoParrotUI answers 404: listed folder by folder, every listing
+        # failed and the tree read as holding nothing, so a fetch without its
+        # cached tree dropped its four recorded paths.
+        profile = {"source": "https://github.com/o/r", "source_commit": "e" * 40,
+                   "files": [{"source_ref": "src/a.c:5"}]}
+        with mock.patch.object(sitesources, "_cached_tree", return_value=None), \
+                mock.patch.object(sitesources.upstream, "list_directory",
+                                  side_effect=upstream.UpstreamError("HTTP 404")):
+            held, unread = sitesources.build_source_paths({"p": profile}, [TMP])
+        self.assertEqual(held, {})
+        self.assertEqual(unread, ["github.com/o/r@" + "e" * 40])
+
     def test_every_rendered_link_is_recorded(self):
         record = json.loads(sitesources.SOURCE_PATHS.read_text(encoding="utf-8"))
         pages = sorted(_built(DOCS / "emulators").rglob("*.md*")) + sorted(DOCS.glob("gaps*.md*"))