From f40d4c817bb46cd95c2b1d52b96dda3a2512c6b5 Mon Sep 17 00:00:00 2001
From: Spark <spark@hatch.local>
Date: Sun, 11 Oct 2026 11:42:23 -0400
Subject: [PATCH] =?UTF-8?q?test(r=C3=A9gression):=20le=20fake=20reproduit?=
 =?UTF-8?q?=20le=20payload=20Gitea=20r=C3=A9el=20(pull=5Frequest:=20null)?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Le faux omettait la clé pull_request sur les issues, ce qui a
masqué le bug de la 0.5.1 pendant que 84 tests restaient verts.
Clé désormais toujours présente (null = issue, objet = PR) +
4 tests dédiés dans test_regression_gitea_payload.py.
---
 tests/conftest.py                      |  5 ++--
 tests/test_regression_gitea_payload.py | 34 ++++++++++++++++++++++++++
 2 files changed, 37 insertions(+), 2 deletions(-)
 create mode 100644 tests/test_regression_gitea_payload.py

diff --git a/tests/conftest.py b/tests/conftest.py
index 4abbf2c..9742982 100644
--- a/tests/conftest.py
+++ b/tests/conftest.py
@@ -55,9 +55,10 @@ def _issue(number: int, title: str, state: str = "open", pr: bool = False) -> di
         "created_at": "2026-09-01T09:00:00Z",
         "updated_at": "2026-09-02T09:00:00Z",
         "html_url": f"http://gitea.test/Projets/demo/issues/{number}",
+        # Faithful to real Gitea 1.27 payloads: the key is ALWAYS
+        # present — null for a plain issue, an object for a PR.
+        "pull_request": {"url": "http://gitea.test/pr"} if pr else None,
     }
-    if pr:
-        issue["pull_request"] = {"url": "http://gitea.test/pr"}
     return issue
 
 
diff --git a/tests/test_regression_gitea_payload.py b/tests/test_regression_gitea_payload.py
new file mode 100644
index 0000000..c513366
--- /dev/null
+++ b/tests/test_regression_gitea_payload.py
@@ -0,0 +1,34 @@
+"""Regression: real Gitea returns "pull_request": null on plain issues.
+
+The 0.2.0–0.5.0 filter tested key *presence* ("pull_request" not in
+item), which classified every real issue as a PR and silently emptied
+every board against a real Gitea — while the test fake omitted the
+key entirely, so the suite stayed green. Found 2026-10-11 after a
+user report ("new issue does not appear on the board"): the live
+flowdeck board rendered 0 cards for 6 open issues.
+"""
+
+from app.main import _is_issue
+
+
+def test_plain_issue_with_null_pull_request_is_an_issue():
+    assert _is_issue({"number": 1, "pull_request": None}) is True
+
+
+def test_pull_request_object_is_not_an_issue():
+    assert _is_issue({"number": 2, "pull_request": {"url": "http://x/pr/2"}}) is False
+
+
+def test_missing_key_is_an_issue():
+    assert _is_issue({"number": 3}) is True
+
+
+def test_board_renders_cards_from_realistic_payload(client):
+    # The shared fake now mirrors the real payload (key always
+    # present, null on plain issues): if the filter regresses to a
+    # presence test, these cards vanish and this test goes red.
+    r = client.get("/board/Projets/demo")
+    assert r.status_code == 200
+    assert "Tâche un" in r.text
+    assert "Tâche focus" in r.text
+    assert "PR déguisée" not in r.text