We recently changed the QA workflow on Bugzilla to use a new 'advisory' keyword instead of an 'advisory' Whiteboard entry. This allows us to fix the annoying behaviour of mga-advisories resetting the 'validated_update' keyword when an advisory is malformed - instead, this update candidate lets it reset the 'advisory' keyword from now on. This needs to be pushed as an update to Mageia 5 (and thus also Mageia 6) so that it can be used by sysadmins on the buildsystem. While at it, the update also changes the default version to Mageia 6 in the advisory template, and default CVE prefix to 2017. Advisory: ========= Updated mga-advisories package to match new QA bugzilla workflow This update propagates a workflow change of the QA team on Bugzilla to the tool used by sysadmins to push updates. It also updates the default advisory template to feature Mageia 6 instead of Mageia 5. SRPM in core/updates_testing: ============================= mga-advisories-0.24-1.mga5 mga-advisories-0.24-1.mga6 RPMs in core/updates_testing: ============================= mga-advisories-0.24-1.mga5 mga-advisories-0.24-1.mga6
Whiteboard: (none) => MGA5TOO
To be able to test this, I've uploaded an advisory with a typo on purpose, so that we can check that mga-advisories will properly remove the 'advisory' keyword. I'll mark this as validated_update once the candidate has been installed on the infra, so that we can test it live.
Keywords: (none) => advisory, validated_updateCC: (none) => sysadmin-bugs
Keywords: validated_update => (none)
Testing MGA5/64, mga-advisories-0.24-1.mga5 Confirm that the template now shows '6'. For the rest, need to await pushing the tweaked advisory.
Whiteboard: MGA5TOO => MGA5TOO MGA5-64-OKCC: (none) => lewyssmith
Wanting to test M6/64... This package does not show in M6 Updates Testing. And the bug RPMs link shows empty.
That's weird, it's available in this mirror at least: http://ftp.free.fr/mirrors/mageia.org/distrib/6/x86_64/media/core/updates_testing/
Keywords: (none) => validated_update
(added validated_update for testing)
Update ID assignment failed Checking for QA validation keyword⦠â Checking dependent bugs⦠â (None found) Checking SRPMs⦠â (6/core/mga-advisoriz-0.24-1.mga6) â 'advisory' keyword reset.
Keywords: advisory => (none)
Ok, needs some more work as this second failure message shouldn't have happened.
I need to add a test for the 'advisory' keyword here: http://gitweb.mageia.org/software/infrastructure/mgaadvisories/tree/lib/MGA/Advisories.pm#n258 This way the script will only consider "good" the bugs with both 'advisory' and 'validated_update' defined. If an advisory is malformed and 'advisory' is reset, it will therefore no longer be processed on future runs of the script.
Testing M6/64 (In reply to Lewis Smith from comment #3) > Wanting to test M6/64... > This package does not show in M6 Updates Testing. Well, this time I *was* able to update to: mga-advisories-0.24-1.mga6 Adding an advisory indeed had 6 as the pre-set release in the template. Can do no more, OK for me.
Whiteboard: MGA5TOO MGA5-64-OK => MGA5TOO MGA5-64-OK MGA6-64-OK
(In reply to Rémi Verschelde from comment #9) > I need to add a test for the 'advisory' keyword here: > http://gitweb.mageia.org/software/infrastructure/mgaadvisories/tree/lib/MGA/ > Advisories.pm#n258 I've fixed the srpm for m6 in the advisory for this update, or was that meant to be there for part of the testing? Should we hold this update till the infrastructure has been updated?
CC: (none) => davidwhodgins
(In reply to Dave Hodgins from comment #11) > I've fixed the srpm for m6 in the advisory for this update, or was that meant > to be there for part of the testing? Should we hold this update till the > infrastructure has been updated? As mentioned in comment 1: > To be able to test this, I've uploaded an advisory with a typo on purpose, so that we can check that mga-advisories will properly remove the 'advisory' keyword. :)
Is the advisory as per by Dave's unintentional 'correction' (comment 11)? or reverted to its deliberately incorrect state for testing this update (comment 1)? If the latter, it is best for you Rémi|Nicolas to validate it when you are ready, to test that it does what you want (comment 8); it already did what we want (comments 6,7). Say if you want QA to do anything more with it.
Taking it back for now, QA shouldn't bother with this one until we've confirmed that it works with Nicolas (at which case it should be possible to validate it and push it directly).
Assignee: qa-bugs => rverschelde
Ping on this one... should it be psuhed or dropped ?
CC: (none) => tmbWhiteboard: MGA5TOO MGA5-64-OK MGA6-64-OK => MGA6-64-OK
Thanks for the 1-year reminder :D The current update candidate can be dropped, but I still need to finish the work started in Git and make a proper update then.
CC: lewyssmith => (none)
So this is still "work in progress"? Best regards,
CC: (none) => ouaurelien
Closing as fixed since mga-advisories-0.25-1.mga7 is in core release.
Resolution: (none) => FIXEDStatus: NEW => RESOLVED