Bug#711833: Preparing for RFS
Hei Yoann,
after sending the previous mail. I talked to pere and asked if he could look at
the package. The package looked good according to him. pere encouraged me to
talk about postr on #debian-python. Which I did and pere offered to sponsor the
package, but the PAPT team did not seem to accept packages maintained in git
currently. So I decided to change the maintainer instead of going back to the
past and using SVN. Below is a transcript from IRC to get the context:
17:13 < pere> ccscanf: hi. what is the status of the postr package you are working on?
17:14 < ccscanf> I think it is ready for a sponsor.
17:18 < ScottK> barry: Here's another problem of that kind I ran across recently: https://github.com/bro/pysubnettree/pull/3/files
17:18 < ScottK> If I hadn't actually be regenerating the files during the build, I probably wouldn't have noticed.
17:20 < ccscanf> If anyone here can try the package, that would be nice. http://mentors.debian.net/debian/pool/main/p/postr/postr_0.13-1.dsc
17:22 < pere> ccscanf: did you ask the python team for git repo hosting?
17:23 < paultag> hahaha
17:23 < paultag> no git for python team :)
17:23 < paultag> svn :)
17:23 < pere> aha.
17:23 < pere> I guess that is cleared, then. :)
17:24 < pere> ccscanf: I guess I'll push your github repo to the collab-maint repo, then?
17:25 < pere> ccscanf: btw, refresh my memory. what is the status of the current postr maintainer David Paleino?
17:26 < ccscanf> pere: No. I have cc-ed David on most of my postr email, but he has not replied any.
17:26 < pere> ok.
17:26 < ScottK> FTR, it's no git without some work being done that no one seems to want to do.
17:26 < ccscanf> pere: Pushing to collab-maint would be nice. Thanks.
17:27 < paultag> ScottK: olasd did it
17:27 < paultag> ScottK: but no one commented or anything, I don't think
17:27 < paultag> git-buildpackage + mergeWithUpstream
17:27 < paultag> which is how I keep repos anyway
17:27 < ScottK> Really.
17:27 < paultag> so nommmm
17:27 < paultag> Really!
17:27 < ScottK> I don't think that was all that was needed.
17:28 < ScottK> There was also the issue of how to make changes across a stack of repos or do partial branches of one large one.
17:28 < ScottK> p1otr: What was the list of stuff before we thought git was OK?
17:30 < pere> ccscanf: git pushed.
17:31 < pere> should http://anonscm.debian.org/gitweb/?p=collab-maint/postr.git or your github repo be the packaging repo?
17:32 < ccscanf> I think it is better to revert back to collab-maint for the Vcs-* fields.
17:33 < p1otr> ScottK: https://lists.debian.org/debian-python/2013/02/msg00123.html
17:34 < paultag> olasd: where's the set of git packages?
17:34 < paultag> if the consensus is try it out and get mindshare, I'll just start updating stuff in git.
17:35 < ccscanf> pere: I can make the change and push to the github repo.
17:35 < ScottK> p1otr: Thanks.
17:36 < ansgar> paultag: The last time Git was mentioned it ended in a flamefest. And the time before too...
17:36 < paultag> I love svn-buildpackage's workflow
17:36 < paultag> I hate git-buildpackage's default workflow
17:36 < ScottK> For git, I like git-dpm.
17:36 < paultag> I love git-buildpackage emulating svn-buildpackage's workflow
17:36 < paultag> so that's what i'll use.
17:37 < ansgar> paultag: What are the workflows? I use neither of them.
17:37 < paultag> git-buildpackage has this stupid idea upstream source should be in the repo
17:37 < paultag> svn-buildpackage just tracks debian, and merges that with pristine upstream
17:38 < ansgar> paultag: I just use git-sbuild which doesn't care if upstream source is present or not :>
17:38 < ccscanf> pere: git pushed.
17:39 < paultag> :)
17:49 < pere> ccscanf: time to upload?
17:50 < ccscanf> pere: If anyone is willing to sponsor, yes.
17:51 < barry> yeah, i've had better luck with git-dpm than git-buildpackage
17:51 < pere> perhaps someone from the python app team would like to do it?
17:52 < ScottK> If it's in Git, (ATM) it's not a team issue.
17:52 < pere> if no-one volunteer quickly, I'll upload it myself. :)
17:52 < pere> ScottK: what do you mean? what is "a team issue"?
17:52 < ScottK> Team packages aren't in git.
17:53 < pere> is that a statement meaning packages in git are unwanted as team packages, or a statement meaning that so far no team maintained packages have been in git?
17:54 < ScottK> It's a statement that the team only has one repo. For now that's in svn. If we move to git, we'll (as described in the mail p1otr linked) move everything.
17:54 < ScottK> We don't want a split.
17:57 < pere> ccscanf: this confuses me, given that the team is listed as the maintainer for postr in git.
17:58 < ccscanf> ScottK: How is that related? postr is a application not a module?
18:00 < pere> ccscanf: did you discuss postr on the team mailing list?
18:02 < ccscanf> pere: I mentioned it https://lists.debian.org/debian-python/2014/05/msg00125.html
18:03 < ScottK> pere: I think that's an error then.
18:03 < ScottK> It should be in svn if it's team maintained.
18:04 < ScottK> It's the same issue for modules and applications teams. It's the ~same people.
18:04 < pere> so far postr have been in the collab-maint git, and ccscanf based his updates on this git repo.
18:05 < ScottK> OK. Then it's not really team maintained.
18:06 < pere> so I guess there are three options. (1) the team accept to put a package in collab-maint git under the team umbrella, (2) the package is moved to the python-apps svn or (3) the package is maintained without the team listed as maintainer.
18:07 < ScottK> Right. #1 has been discussed many times and rejected.
18:08 < pere> if the team as ScottK expresses it refuses to touch packages in git, (1) is out of the question. options 2 and 3 can be selected by ccscanf, while option 1 need team accept. ccscanf, I guess it is up to you. :)
18:08 < ScottK> p1otr: ^^^ am I wrong?
18:09 < ScottK> pere: We don't want multiple repositories and a git transition isn't going to do itself. That's not really the same thing as "refuses to touch packages in git".
18:09 < ccscanf> pere: I am going for 3.
18:09 -!- achadwick [~achadwick@nat-alpha.ceu.ox.ac.uk] has quit [Quit: Ex-Chat]
18:09 < pere> in practice, I am covinced the nice people on this channel is willing to provide advice even if the team is not listed in the control file, so the practical difference should not be too big. :)
18:12 < p1otr> ScottK is right. Show us the code or go away
18:13 < pere> p1otr: does it mean I am wrong, and the nice people on this channel want us to go away if the package stay in git?
18:14 < pere> (the code is in <URL: http://git.debian.org/?p=collab-maint/postr.git >, if you want to look at it. :)
18:14 < p1otr> I didn't read everything. "go away" is for people who want to force me to do more work
18:14 < p1otr> I will move at least PAPT to git at some point
18:14 < p1otr> I love git
18:15 < p1otr> I hate people who want to force me to do something they like
18:15 < p1otr> ... and do nothing themselfs
18:15 < pere> p1otr: I fail to understand why you bring this up in this discussion. why do you believe it is relevant?
18:16 < p1otr> relevant to what?
18:16 < paultag> It's not, we're just a disfunctional team :)
18:17 < pere> that seem like a nice summary of the situation. ccscanf, so what is next?
18:18 < p1otr> if it's not "you guys should do this and this but I will not do it myself because you don't like me", then I'm sorry
18:19 < pere> p1otr: I am sorry to, to have learned that this is not the place I believed it was, and that I suggested to ccscanf that he should get in touch with the python team.
18:19 < ScottK> pere: I think you are misunderstanding.
18:19 < ccscanf> pere: I set myself as the maintainer, the changes are on github.
18:20 < p1otr> pere: I don't have time to process what we have right now, I don't have time to prepare an transition
18:20 < pere> ScottK: could be.
18:22 -!- gudjon [~quassel@31-209-201-19.dsl.dynamic.simnet.is] has quit [Read error: Connection reset by peer]
18:23 -!- gudjon [~quassel@31-209-201-19.dsl.dynamic.simnet.is] has joined #debian-python
18:24 < pere> ccscanf: pushed to collab-maint.
18:24 < ccscanf> Thanks.
18:24 -!- simonft [~simonft@castro.tenthcircle.net] has joined #debian-python
18:24 < pere> ccscanf: I guess we are done here. :) Very illuminating session. :)
18:25 < p1otr> pere: I'm sorry, I didn't want to be rude
PS: The log is longer, but I tried to make it short enough to get the context. Sorry if it is too long to read.
Reply to: