Bug 9942

Summary:	patchwork: ensure patches are not marked as Accepted when a newer Series revision is modified
Product:	[Yocto Project Subprojects] Patchwork/Patchtest	Reporter:	Jose Lamego <jose.a.lamego>
Component:	Patchwork	Assignee:	Jose Lamego <jose.a.lamego>
Status:	RESOLVED FIXED	QA Contact:
Severity:	normal
Priority:	Medium+	CC:	bluelightning, leonardo.sandoval.gonzalez, sgw
Version:	unspecified
Target Milestone:	2.3 M1
Hardware:	All
OS:	Multiple
Whiteboard:
OS type for building Yocto:	---	Type of Regression:	---
Verified:		Documentation change:	No (bug/feature does not impact docs)
Bug Depends on:	10153
Bug Blocks:

Description Jose Lamego 2016-07-14 15:16:42 UTC

Patches from a previous revision (v1) were marked as Accepted after two patches in a newer series revision (v2) were changed to Accepted state:


https://patchwork.openembedded.org/series/271/

Wrongly marked patches: 126549, 126551, 126553, 126555, 126557, 126559.

Find root cause and verify that there are no other affected series.

Comment 1 Jose Lamego 2016-08-25 21:57:35 UTC

Attempted to replicate the issue at a staging instance with the following conditions:

- Same Patchwork commit as in OE-Patchwork (fc3c66a1ac3e94ce26ca29b4296f1196b869cbac)
- Created an initial patch Series from patch email messages created with create-pull-request script
- Created both a subsequent (v2) patch Series and a second revision (for original series) from patch email messages created with create-pull-request script
- Updated some patches from second revision/series

But reported behavior was not detected.
 
This seems to be a symptom and result of a bad series creation, which is fixed in bug 9904. After this fix is implemented, further testing will support this assumption.

Comment 2 Leonardo Sandoval Gonzalez 2016-09-08 17:07:59 UTC

This wont be ready for M3 due to the dependency (not met). Moving to M4.

Comment 3 Leonardo Sandoval Gonzalez 2016-10-19 13:54:29 UTC

This bug is still depending on 10153 but I believe we can move both to 2.3 M1. 10153 is launching a PW testing instance so this is definitely something can be done post 2.2.

Comment 4 Jose Lamego 2016-12-09 17:07:49 UTC

Couldn't replicate issue at patchwork-staging instance with series 3895 [1], adding to the assumption that the reported issue was caused by a wrongly-created series, which is already fixed in production.

[1] https://patchwork-staging.openembedded.org/series/3895/