patchtest: warn if Upstream-Status is after patch scissors
PATCH v3 1/1 patchtest: warn if Upstream-Status is after patch scissors
[YOCTO #15940]
Detect cases where Upstream-Status is placed after the patch scissors ("---") and emit a warning instead of failing.
Such placement may be intentional in some workflows to avoid including the tag in upstream commits.
v3:
- Fix indentation issues from previous version
- Keep changes minimal and avoid unrelated formatting changes
- Retain intended warning behavior
--- meta/lib/patchtest/tests/test_patch.py | 24 +++++++++--------------- 1 file changed, 9 insertions(+), 15 deletions(-)
meta/lib/patchtest/tests/test_patch.py +9 -15
@@ -67,13 +67,10 @@ class TestPatch(base.Base): if not header_has_upstream: if body_has_upstream: - self.fail( - 'Upstream-Status is present only after the patch scissors. ' - "It must be placed in the patch header before the scissors line.", - data=[ - ("Standard format", self.standard_format), - ("Valid status", self.valid_status), - ], + print( + "WARNING: Upstream-Status is present only after the patch scissors. " + "It may be ignored by git and lost during patch refresh. " + "Consider placing it before the scissors line." ) else: self.fail(
@@ -92,15 +89,12 @@ class TestPatch(base.Base): continue if scissors_index is not None and idx > scissors_index: - self.fail( - 'Upstream-Status must be placed in the patch header before the scissors line, ' - "but was found afterwards.", - data=[ - ("Current", line.lstrip("+")), - ("Standard format", self.standard_format), - ("Valid status", self.valid_status), - ], + print( + "WARNING: Upstream-Status found after patch scissors. " + "It may be ignored by git and lost during patch refresh. " + "Consider placing it before the scissors line." ) + continue if patchtest_patterns.inappropriate.searchString(line): try:
Signature · 1 line
2.34.1
Quoted 8 lines · Aditya GS via lists.openembedded.org
[YOCTO #15940]
Detect cases where Upstream-Status is placed after the patch scissors ("---") and emit a warning instead of failing.
Such placement may be intentional in some workflows to avoid including the tag in upstream commits.
I'm afraid there has been a confusion here, and the patch is not correct.
The bug was already addressed in oe-core earlier:
commit 2156ef9e6defa3ec9087789fcea25fb4fee7b83c
Author: Luka Krstic <[email protected]> Date: Thu Dec 18 10:11:38 2025 +0100
patchtest: reject Upstream-Status after scissors
but the bugzilla ticket was never closed.
Then https://bugzilla.yoctoproject.org/show_bug.cgi?id=15940#c1 was interpreted as a request to downgrade it to a warning, but we should not be doing that: it means patchtest will not longer actually notify of the issue on the oe-core mailing list.
I understand some other layers intentionally put it after the scissors; if they're using patchtest, they can disable this particular test or configure it to produce some kind of soft fail, but in oe-core it should remain a hard fail.
Alex
Hi Alex,
Thanks for the clarification.
I understand now that the behavior should remain a hard failure in oe-core, and that the earlier interpretation was incorrect.
I'll drop the patch that downgrades the check to a warning.
Thanks for pointing out the existing commit in oe-core as well that helps clear up the confusion.
Regards,
Aditya
Quoted 35 lines · Alexander Kanavin
Quoted 8 lines · Aditya GS via lists.openembedded.org
[YOCTO #15940]
Detect cases where Upstream-Status is placed after the patch scissors ("---") and emit a warning instead of failing.
Such placement may be intentional in some workflows to avoid including the tag in upstream commits.
I'm afraid there has been a confusion here, and the patch is not correct.
The bug was already addressed in oe-core earlier:
commit
2156ef9e6defa3ec9087789fcea25fb4fee7b83cAuthor: Luka Krstic <[email protected]> Date: Thu Dec 18 10:11:38 2025 +0100patchtest: reject Upstream-Status after scissors
but the bugzilla ticket was never closed.
Then https://bugzilla.yoctoproject.org/show_bug.cgi?id=15940#c1 was interpreted as a request to downgrade it to a warning, but we should not be doing that: it means patchtest will not longer actually notify of the issue on the oe-core mailing list.
I understand some other layers intentionally put it after the scissors; if they're using patchtest, they can disable this particular test or configure it to produce some kind of soft fail, but in oe-core it should remain a hard fail.
Alex