12:29:49 <matthieucan> #startmeeting
12:29:49 <MeetBot> Meeting started Fri Feb 13 12:29:49 2015 UTC.  The chair is matthieucan. Information about MeetBot at http://wiki.debian.org/MeetBot.
12:29:49 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
12:30:02 <matthieucan> #topic OPW debsources - weekly review
12:30:51 <sophiejjj> shall we go on?
12:30:59 <matthieucan> sophiejjj: anything to discuss before the point about c.d.n (where it's better to have zack)?
12:31:35 <matthieucan> what about the item "learn debian.copyright API from python-debian"?
12:31:39 <sophiejjj> matthieucan: the current phase is about refactoring the existing code.
12:31:46 <sophiejjj> yup.
12:31:50 <sophiejjj> I read that.
12:31:55 <matthieucan> yeah that's what I meant
12:31:57 <matthieucan> great!
12:32:03 <matthieucan> did you play with the library?
12:32:19 <sophiejjj> no. read the code and example.
12:32:27 <sophiejjj> I am working on the debsources code base.
12:32:58 <sophiejjj> fyi. https://github.com/sophiejjj/debsources/commits/cr
12:32:58 <matthieucan> alright. you should maybe try to parse a random debian/copyright, to see how things work
12:33:05 <matthieucan> let me see that
12:34:08 <matthieucan> 89 changed files with 2,591 additions and 2,230 deletions
12:34:15 <matthieucan> that's some huge refactoring
12:34:24 <themill> (python-debian's copyright class is brand new and not tested much yet so please do feel free to report bugs and make suggestions about what else it could do)
12:34:32 <sophiejjj> themill: sure.
12:34:40 <sophiejjj> matthieucan: the status is
12:34:53 <sophiejjj> most of the web functionalities are running
12:35:06 <sophiejjj> other then the SourceView which is in progress.
12:35:15 <matthieucan> alright
12:35:18 <matthieucan> how many tests pass?
12:35:42 <sophiejjj> I haven't work on the tests yet.
12:35:51 <sophiejjj> but the api are largely maintained.
12:36:08 <matthieucan> ok
12:36:17 <matthieucan> what's COPYRIGHT-SITE-NOTES and the yaml stuff?
12:36:25 <sophiejjj> my plan is, after all the functionalities looks ok, go to the reviews phase
12:36:38 <sophiejjj> and at that time I fix the tests and change the code per your advice.
12:37:05 <sophiejjj> site-notes are some temp stuff I remind myself of.
12:37:21 <matthieucan> oh alright
12:37:22 <sophiejjj> the yaml stuff is, after some investigation,
12:37:33 <sophiejjj> I think the yaml format is more suitable for the config.
12:37:39 <zack> here I'm
12:37:40 <matthieucan> what amount of work do you think is still needed?
12:37:44 <zack> will catch up with the backlog
12:37:45 <matthieucan> hi zack
12:37:46 <sophiejjj> the source view.
12:37:56 <sophiejjj> and some changes I made in the debsources
12:38:07 <matthieucan> is it possible to fix a deadline for that?
12:38:35 <matthieucan> not sure the yaml format can handle sections with a default section?
12:38:36 <sophiejjj> canont say deadline, but I'd like to have the function been running this weekend.
12:38:49 <matthieucan> this would be nice :)
12:38:53 <sophiejjj> default is a line of code, and I've taken care of it.
12:39:22 <matthieucan> the yaml idea should be discussed/implemented separately from the refactoring
12:39:29 <matthieucan> it's big enough already
12:39:32 <zack> right, so ack to matthieucan's point about the tests, ideally they should all work, at least from the point of view of old APIs and the like
12:39:39 <sophiejjj> zack: yes.
12:39:44 <zack> we shouldn't break the public API due to internal refactoring reasons
12:39:50 <sophiejjj> of course.
12:40:31 <zack> I've a question about the process, how do you imagine the merge in master will work? single patch, multiple patches, or what?
12:40:50 <sophiejjj> single patch I think.
12:41:13 <zack> that would be massive, but better than merging the current long series of "work in progress" commits :)
12:41:24 <matthieucan> I agree about that
12:41:33 <sophiejjj> haha.
12:41:38 <matthieucan> and it keeps the master branch stable between every commit
12:41:43 <sophiejjj> yes.
12:41:47 <sophiejjj> exactly.
12:41:56 <zack> the question is, how hard would it be to review that patch :)
12:42:07 <sophiejjj> very hard, IMHO.
12:42:24 <matthieucan> unless the refactoring can be properly splitted, it would be the same with many patches, right?
12:42:35 <zack> matthieucan: yes, splitting would be best
12:42:52 <zack> but sophiejjj is not very at ease with splitting commits using git, IIRC
12:42:58 <zack> sophiejjj: please correct me if I'm wrong
12:43:03 <zack> so I don't think we really have a choice
12:43:06 <matthieucan> it's not the main problem
12:43:20 <sophiejjj> matthieucan: sadly, that's the truth. I am not really good at that.
12:43:23 <zack> the alternative is spiltting the big patch at merge time
12:43:34 <zack> matthieucan: i.e., you or me do that while reviewing
12:43:35 <matthieucan> the problem would be to split the refactoring part, that is, keep being stable between commits
12:43:45 <sophiejjj> but, there is still some time of work before the merge.
12:44:19 <zack> anyway, so it seems to me that the status for next week is "keep on working, see what happens"
12:44:20 <sophiejjj> AFAIC, step 1. the web functionality work. step2. fix other minor changes, like module path change step3. fix the tests.
12:44:25 <matthieucan> zack: oh I see, to split the reviewing work actually? looks doable with "files in the app"/"infra"
12:44:52 <zack> matthieucan: that's possible too, yes, but I suspect most of the changes will be in the webapp
12:45:08 <sophiejjj> zack: with some minor changes in debsources.
12:45:26 <sophiejjj> say, rename the data.stat to sources_data.stat
12:45:29 <zack> those will be easy targets for a separate commit then
12:45:33 <matthieucan> sophiejjj: what about the conf management? I didn't look in details but I saw you changed things on GH
12:45:36 <sophiejjj> or move out the Location, Directory from models.
12:45:42 <sophiejjj> yes.
12:45:49 <sophiejjj> I changed it to yaml.
12:45:56 <sophiejjj> the webapp can read the conf.
12:46:04 <sophiejjj> but I don't test it with the debsouces update thing.
12:46:15 <zack> sophiejjj: you changed the file configuration syntax to yaml?
12:46:22 <sophiejjj> yes.
12:46:27 <matthieucan> what about that: https://github.com/sophiejjj/debsources/compare/master...cr#diff-e2817b76bc82e8e37ccf811e4cfe886fR92
12:46:30 <zack> without discussing it with us?
12:46:31 <sophiejjj> zack: you don't like yaml?
12:46:35 <zack> that's not the point
12:46:40 <sophiejjj> zack: oh. last week
12:46:45 <sophiejjj> I mentioned changed the mainlib
12:47:03 <sophiejjj> I must forget to mention exactly the yaml
12:47:08 <zack> changing the configuration file format means that the day we deploy on the machine running sources.d.n will *break*
12:47:16 <zack> that's not OK
12:47:30 <sophiejjj> zack: that's what I am curious about recently.
12:47:42 <sophiejjj> How do you deploy the code?
12:47:45 <zack> that's exactly the kind of changes that if not properly discussed/documented with us will block your refactoring work to be integrated
12:47:54 <matthieucan> we git pull
12:48:08 <sophiejjj> en.
12:48:17 <sophiejjj> then while will it break?
12:48:22 <zack> of course it will
12:48:30 <zack> because the configuration file we actually use will not be in yaml
12:48:49 <zack> sophiejjj: is there any concrete advantage in switching configuration file syntax other than "I like it more"?
12:49:02 <sophiejjj> zack: it's human-readable, typebased.
12:49:13 <sophiejjj> so much of the typing stuff in mainlib are omitted.
12:49:26 <zack> that's an advantage, I agree
12:49:30 <sophiejjj> you can directly write list, set,
12:49:55 <zack> do we still have inheritance of values between the "default" section and the other sections?
12:49:56 <sophiejjj> I've spent some time comparing it with other techinques.
12:50:03 <sophiejjj> yes.
12:50:08 <zack> ok
12:50:22 <zack> but please: do not take this kind of decisions without consulting us
12:50:27 <zack> it just takes a mail
12:50:30 <sophiejjj> yes. my fault.
12:50:30 <zack> and we usually respond very quickly
12:50:43 <matthieucan> and don't mix this with the refactoring patch
12:50:55 <zack> also, please keep a list of changes that will need to be changed to fix breakages
12:51:04 <zack> like: "convert the syntax of the configuration file to yaml" :-)
12:51:12 <sophiejjj> got it.
12:51:17 <zack> otherwise we won't know how to deploy the new code
12:51:53 <zack> ok, anything else to report/discuss about the refactoring?
12:52:02 <sophiejjj> The sourceview on the go.
12:52:17 <zack> is that becoming a separate blueprint too?
12:52:21 <sophiejjj> is.
12:52:24 <zack> nice
12:52:25 <sophiejjj> yes.
12:53:06 <zack> please keep us posted next week, before friday, about how things are going
12:53:16 <sophiejjj> sure.
12:53:21 <zack> in particular, given you wanted to have things almost in place for this week-end, it'd be nice to know how it went on monday
12:53:46 <sophiejjj> Probably the webapp is running up.
12:53:54 <sophiejjj> and no link gives an exception.
12:54:05 <zack> great, let us know how the "probably" will end up being ;)
12:54:18 <sophiejjj> sure.
12:54:23 <zack> I've two other requests, on the communication front:
12:54:35 <zack> 1) I think your next blog post is overdue now, please do it
12:54:47 <zack> 2) have a look at https://lists.debian.org/debian-women/2015/02/msg00003.html ; please mail opw@debian.org about your work
12:55:05 <sophiejjj> regarding 2. I just gave a thorough read before the meeting.
12:55:35 <sophiejjj> actually I should have done the SourceView today, but I am busy preparing and applying for the visa to Canada, so...
12:56:22 <zack> sophiejjj: well, you're supposed to work full time on OPW, so that's not really an excuse (especially given you didn't tell us in advance)
12:56:41 <sophiejjj> my bad.
12:56:43 <zack> but let's not go down that path
12:57:00 <zack> sophiejjj: I guess we can move back to backburner the copyright.d.n part, right?
12:57:13 <zack> it's not gonna happen before the refactoring is over anyhow
12:57:28 <sophiejjj> yes.
12:57:38 <zack> ok, please do
12:58:27 <zack> anything else to discuss, before closing the meeting?
12:58:40 <sophiejjj> nothing else.
12:58:57 <zack> ok, so let's talk again on monday via email
12:59:01 <matthieucan> ok
12:59:02 <sophiejjj> sure.
12:59:04 <zack> #endmeeting
12:59:11 <matthieucan> #endmeeting