10738 2016-11-28 18:11:42 +0000 patchtest: Patches not sent as reply to their respective series/root message should trigger a warning message 2017-12-06 17:18:50 +0000 1 1 6 Yocto Project Subprojects Patchwork/Patchtest Patchtest Testsuite unspecified x86 Multiple RESOLVED INVALID Medium normal 2.5 1 jose.a.lamego leonardo.sandoval.gonzalez bluelightning patrick.ohly --- --- No (bug/feature does not impact docs) oldest_to_newest 68636 0 jose.a.lamego 2016-11-28 18:11:42 +0000 Patchtest should have a test to detect patches that may belong to a series (due to the "x/n" prefix in subject), but that were not sent as a reply to any root message. This should not cause the test to fail, but to trigger a warning. An example of such a patch email is in http://lists.openembedded.org/pipermail/openembedded-core/2016-November/129348.html 74253 1 leonardo.sandoval.gonzalez 2017-06-15 19:39:50 +0000 (In reply to comment #0) > Patchtest should have a test to detect patches that may belong to a series > (due to the "x/n" prefix in subject), but that were not sent as a reply to > any root message. This should not cause the test to fail, but to trigger a > warning. > > An example of such a patch email is in > http://lists.openembedded.org/pipermail/openembedded-core/2016-November/ > 129348.html Not sure if we way to trigger a warning/fail on this case. People usually send their n revision without indicating the in-reply-to message-id of the n-1 revision. Also, for the moment, patchtest just trigger failures, not warnings, but at the end if patchwork is sending a email to the mailing list on behalf of patchtest, we want actions to be done from the developer, so I am not sure if failures/warnings would make sense. Paul, thoughts about it? 74870 2 leonardo.sandoval.gonzalez 2017-07-07 17:01:20 +0000 comment 1 basically resumes the decision to mark it as invalid. 74872 3 leonardo.sandoval.gonzalez 2017-07-07 17:07:26 +0000 (In reply to comment #2) > comment 1 basically resumes the decision to mark it as invalid. s/resumes/summarizes