Hi, We had nice conversation about this at IRC yesterday with Luis and Gary Birkett (lcuk) and I promised to send a short description via mailing list.
The main problem is that CE and Vanilla aren't in sync and CE fixes doesn't work in Vanilla. This may lead to following: Bug reported with Vanilla image: - Might be ignored / marked as a duplicate for a bug existing in the CE. - If bug is fixed for CE, the bug is verified, but the fix is missing from Vanilla. Bug reported with CE image, reproduced with Vanilla: - Fixed and verified for the CE, Vanilla is ignored. So, we need to avoid these. But how? This might happen with other devices also, same hardware but different release, so let's try to find a good method to use not only with N900 but others to come also. One solution is to Clone the bug: - If reported for Vanilla and then reproduced with CE, the CE gets own clone bug and vice versa. - Fixing, releasing and verifying is done separately to both. And the bugs are still linked in a comment level, so nothing important isn't disappearing. One solution is to do as Luis proposed. - Problem: there isn't clear visibility in the bug if it's fixed for the image it has been reproduced later - Problem: if the bug is fixed for the image it's originally reported, but not to the other One solution is keep conversation ongoing in one bug and clone it IF needed - IF the bug is fixed with original image, then bug is cloned for the other image - If the bug is fixed for the image where it has been reproduced later, commenting that this is fixed and keep the bug open to show that originally reported bug isn't fixed/released - Problem: there isn't clear visibility in the bug if it's fixed for the image it has been reproduced later Best solution: - You name it. It must be one when all of us are happy (or at least somehow satisfied) : developers, testers, users, managers. I hope I could add everything needed to this bug. Please be active and tell what you think. Br, Iekku >-----Original Message----- >From: ext Luis Araujo [mailto:luis.ara...@collabora.co.uk] >Sent: 18 August, 2011 03:55 >To: Pylkka Iekku (EXT-Ixonos/Tampere) >Cc: meego-...@meego.com; meego-handset@lists.meego.com >Subject: Re: [Meego-handset] N900 and N900CE related bugs > >Hello Iekku, > >Let me summarise and see if I get this right: > >(*) Bugs reported for N900CE (or other derivatives) that are fixed but still >reproduced in N900 Vanilla images: > >- Set status to RESOLVED/FIXED >- Remove N900CE keyword >- Add comment that it is verified for N900CE > (It could be useful to add a comment stating that it is still reproduced in >Vanilla image too) >- Set status to VERIFIED only once the fix arrives to Vanilla image. > >The current procedures goes like that?, if I follow you right?, This sounds >good >for me. > >Now, my main concern and I also propose here a way to handle such a case is >the following: > >(*) Bugs reported for N900 Vanilla images but the fix is _only_ available so >far >for N900CE or other derivatives: > >- Stay bug status as open (NEW, REOPENED...): This has some benefits in my >opinion, for example, it avoids duplicating bugs, users testing in the Vanilla >images will find this bug report and also add any comment if needed. >- Once there is a fix for N900CE or other derivatives: Add a comment with the >upstream merge or OBS submit request of such a fix. This way, Vanilla users >finding this bug will find the fix is already available for one of the >derivatives >and they can just start using that instead if they want (hey ... project >promotion through bug reporting :P) >- Only set RESOLVED the bug once a specific decision or fix has been applied to >this Vanilla image: >a) Bug fixed: set RESOLVED/FIXED and VERIFIED later. >b) Bug won't be fixed for N900 Vanilla image for X or Y reason (it is not >always >easy to know who takes this decision, Release Managers?, if RM are not sure, >upstream developers could have a call here): the status is set to INVALID or >WONTFIX, probably adding a comment stating the reasons. > >I see two main and important advantages of this approach: > >- Users of the image will still be able to find a bug report for the reproduced >issue, hence avoiding duplicating bugs. >- Users will be able to find where the fix is already available. > >This whole thing is probably a bit complex, like you said Iekku, but it'd be >nice >to make this clear so we can take full advantage of bug reporting for all these >different image derivatives, we could come out with some guidelines not only >for N900 platforms, but also any other ones. > >Hopefully this thread might help to clarify or set some of these guidelines :) >, I >would be glad to add this to the wiki if you think it is good enough. > >Please, comments, suggestions? > >Regards, > >- Luis > >On 08/17/2011 01:36 AM, ext-iekku.pyl...@nokia.com wrote: >> Hi, >> >> You have good point there. It has been agreed that bugs >reproduced/reported only for N900CE can be verified after fix. If the bug is >reproduced with the Vanilla image also, it can be verified for N900CE = add >comment, remove the keyword N900CE. The bug can stay as a resolved fixed, >but can't be verified before the fix is released for Vanilla image. Little bit >complex, but it was the best way to handle the situation. I need to check if >this is commented clear enough in the N900 wiki, need to check and if not, I >will add the information there. >> >> Thanks, >> Iekku >> >>> -----Original Message----- >>> From: meego-handset-boun...@lists.meego.com [mailto:meego- >handset- >>> boun...@lists.meego.com] On Behalf Of ext Luis Araujo >>> Sent: 17 August, 2011 04:45 >>> To: meego-...@meego.com >>> Cc: meego-handset@lists.meego.com >>> Subject: [Meego-handset] N900 and N900CE related bugs >>> >>> Hello, >>> >>> A N900 bug should be resolved as fixed if such a bug is fixed for the >>> N900CE version? >>> >>> I have seen some bugs following this trend, and I am not sure if this >>> is what we want, or if this was discussed somewhere, do we have a >>> consensus about how to go in such a situation where one bug is fixed >>> in one version but not in other? >>> >>> In my opinion a N900 bug should stay open as long as it is reproduced >>> in the vanilla images, or, if Release Managers decide to mark as >>> INVALID or WONTFIX for such a version. Now, I agree that a N900 bug >>> could be used to track N900CE issues too, but closing a bug reported >>> for one image version, because it is fixed in another image version >>> is a bit confusing in my opinion, even if it is platform related. >>> >>> Comments?, I am not sure if this is explained somewhere, but it'd be >>> nice if we could get a consensus about this if there is none yet. >>> >>> Regards, >>> _______________________________________________ >>> MeeGo-handset mailing list >>> MeeGo-handset@lists.meego.com >>> http://lists.meego.com/listinfo/meego-handset _______________________________________________ MeeGo-handset mailing list MeeGo-handset@lists.meego.com http://lists.meego.com/listinfo/meego-handset
