Hello, Dnia środa, 10 grudnia 2014 10:42:18 Dominig ar Foll pisze: > Hello, > > I see today in Gerrit a patch pushed in an unacceptable way and I would > like to see that type of behaviour changed. > I speak about the review > https://review.tizen.org/gerrit/#/c/31701/
What does it mean pushed? It's not merged yet, so I don't see the problem. > The push is justified on the comment : > > "Patch Set 3: > As of now, I feel adding these credentials here is the right way to do > because: > We want kdbus to be part of official images, so it is not going to > be optional, although we may disable kdbus at boot time by default. > Until we enable sysusers feature of systemd we'd like to avoid > adding and removing users with useradd/userdel." I think, that this comment is biggest issue here, as it suggest, that we're trying to push kdbus to official Common images. This is not true. As I stated in other discussions, we're preparing separate images with kdbus. We don't want to integrate kdbus with Tizen as long as it isn't stable. Łukasz motivation was to have small patch, but now he is changing it to fit Stéphane comments. > The "We want Kdbus ..." is not an agreed general feature but can only > be, as today, a valid profile specific request. > So the patch as it is shall be rejected :-( Hmm? According to official presentations[1] and wiki[2] kdbus is official Tizen 3.0 feature. I haven't seen any information about transition to Yocto in such documents and changing build-system is bigger task then preparation for kdbus. > In order to get it in, the same person pushed it, verified it, and > accept it in less than 2 hours of working day time in Europe where are > located the Common reviewers. > Sorry but that is not acceptable. 2014-12-10 - 2014-12-08 = 2 days (not 2 hours, BTW change owner is also located in Europe). During these 2 days there was only positive feedback. After +2, there were some -1 comments (and these are taken into account, new patchset is prepared), so still I don't see a problem. > 1) Same person submitting, verifying and accepting a patch (even simple) > is not in line with Tizen review model, Yes, but when there's no review from maintainers, person who submitted the patch may do this AFAIK. > 2) A decent time should be left for reviewer to voice their concerns. > Typical 24h to cover multiple time zones. 2 days = 2*24 hours > 3) Architecture change shall be agreed by architecture team to get in > Common. systemd is adding quite a few users for its daemons. Do we have to ask architecture team for approval for every such user during systemd update? Maybe even every systemd upgrade should be consulted, because upstream frequently changes features which have bigger impact on architecture then adding a daemon user. > So please do not force changes any more. > In that specific case, if you need Kdbus in a profile before than it is > agree to be generalised in Common, please make an agreement with Common > to enable it with a clean model. > In that specific case the use of a "%bcond_with kdbus" seems a viable model. > > Please sync with Common RE (Stéphane) to see if that model would work > for all of us. > > Please accept my apologies for having to be a bit rude. I'm also sorry for few sharp words, but I'm feeling my work isn't treated fairly. > Regards [1] <http://download.tizen.org/misc/media/conference2014/slides/tdc2014-tizen3.0-update.pdf> [2] <https://wiki.tizen.org/wiki/Tizen_3.0> regards, -- Maciej Wereski Samsung R&D Institute Poland Samsung Electronics [email protected] _______________________________________________ Dev mailing list [email protected] https://lists.tizen.org/listinfo/dev
