ReviewerMeeting20100602

Not logged in - Log In / Register

ReviewerMeeting20100602

summary

logs

ameu

[15:00] <MootBot> Meeting started at 09:00. The chair is bac.
[15:00] <MootBot> Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE]
[15:00] <bac> hi, and welcome to the launchpad reviewers meeting
[15:00] <bac> who is here today?
[15:00] <EdwinGrubbs> me
[15:00] <sinzui> me but I am not really in my head yet
[15:01] <henninge> me
[15:01] <abentley> me
[15:01] <BjornT> me
[15:01] <deryck> me
[15:02] <gary_poster> me
[15:02] <jelmer_> me
[15:02] <jtv> me
[15:02] <leonardr> me
[15:02] <mars> me
[15:02] <bac> noodles775, jelmer_: ping
[15:02] <bac> sorry jelmer_
[15:03] <bac> gmb: ping
[15:03] <bac> team leads would you round up your strays?
[15:03] <gmb> me
[15:03] <noodles775> me
[15:04] <bac> [topic] agenda
[15:04] <MootBot> New Topic:  agenda
[15:04] <bac>  * Roll call
[15:04] <bac>  * Agenda
[15:04] <bac>  * Outstanding actions
[15:04] <bac>  * Mentoring update
[15:04] <bac>  * New topics
[15:04] <bac>    * Schedule shuffle
[15:04] <bac>    * Reminder about policy for reviews and landings from external contributors. Making sure the "lander" is recorded is a Good Idea. [henninge]
[15:04] <bac> [topic] outstanding actions
[15:04] <MootBot> New Topic:  outstanding actions
[15:05] <bac> * bac to define new doctest policy regarding what is "testable documentation".
[15:05]  * bac didn't do anything last week.  roll.
[15:05] <bac> on that note, i apologize for the very late cancellation of this  meeting last week.
[15:05] <bac> [topic] * sinzui to talk to QA about our QA tracking problem and create a proposal on the mailing list
[15:05] <MootBot> New Topic:  * sinzui to talk to QA about our QA tracking problem and create a proposal on the mailing list
[15:05] <henninge> bac: np
[15:05] <abentley> bac, I could take a run at the "testable documentation" thing if you like.
[15:05] <sinzui> That is stalled
[15:06] <bac> sinzui: ok.
[15:06] <bac> abentley: thanks.  let's talk later today.
[15:06] <abentley> bac, sure.
[15:06] <bac> [topic] * Schedule shuffle
[15:06] <MootBot> New Topic:  * Schedule shuffle
[15:06] <bac> due to some personel changes we've had holes in our schedule
[15:07] <henninge> bac: you skipped the Mentoring update ;-)
[15:07] <bac> henninge has agreed to move to Monday/EU
[15:07] <henninge> yup
[15:07] <bac> thanks henninge
[15:07] <bac> mars and leonardr have also agreed to start doing OCR again, sharing a shift.  they will alternate OCR on tuesdays.
[15:08] <bac> to accomodate them i'm moving to take the friday shift
[15:08] <bac> with those changes we have pretty good daily coverage again
[15:08] <bac> thanks to everyone for being flexible.
[15:08] <bac> [topic] * Ursinha's comments on orphaned commits after last week's meeting. [henninge]
[15:08] <MootBot> New Topic:  * Ursinha's comments on orphaned commits after last week's meeting. [henninge]
[15:09] <henninge> oh, a new topic.
[15:09] <henninge> ;)
[15:09] <henninge> bac: isn't that what sinzui's action item is about?
[15:09] <sinzui> it overlaps
[15:10] <bac> henninge: is that item a leftover?
[15:10] <bac> sinzui, henninge: yes, it looks like it can be rolled into one
[15:10] <henninge> well, she commented on the orphaned commit discussions, I summarized on the agende page.
[15:10] <sinzui> bug numbers, qa tags, and kanban is my broad topic that needs team leads to provide a clear process
[15:10] <bac> [topic]  Reminder about policy for reviews and landings from external contributors. Making sure the "lander" is recorded is a Good Idea. [henninge]
[15:10] <MootBot> New Topic:   Reminder about policy for reviews and landings from external contributors. Making sure the "lander" is recorded is a Good Idea. [henninge]
[15:11] <henninge> ok, I prepared that ...
[15:11] <henninge> Because it just came up, I want to remind everybody to please require external contributors (community, but also non-LP afaik) to have had a pre-imp with somebody from the relevant team.
[15:11] <henninge> Unless, of course, you are on that team. ;-)
[15:11] <henninge> If not, you could either ask the contributor to do that before you start the review or you could pass it off to be reviewed by somebody from that team if he/she agrees.
[15:11] <henninge> Also, please make sure that your name is mentioned when you land a branch for somebody else. In the case at hand that was not done and the reviewer was not really sure if he did the landing.
[15:11] <henninge> In an earlier discussion two possible ways to land were mentioned.
[15:11] <henninge> 1. Create your own branch, merge in the contributor's branch, land that. The commit will obviously mention your name. Here it makes sense to add "landed on behalf of ..." to the commit message.
[15:11] <henninge> 2. Land the contributor's branch directly, e.g. "ec2 land <mp-url>". Here the landing goes through under the contributor's name and you need to add "landed by ..." to the commit message. (and not "on behalf of")
[15:11] <henninge> That's all. ;-)
[15:11] <mars> henninge, for 2., should we also use the --author flag?
[15:12] <henninge> mars: I don't know about that flag, tbh
[15:12] <mars> henninge, ok, nm :)
[15:12] <bac> i'd like to discourage using 2. as it allows for unreviewed changes to slip in
[15:12] <henninge> oh, right, I remenber that discussion
[15:12]  * henninge always uses 2.
[15:13] <Ursinha> gary_poster and I had a conversation yesterday about the orphaned commits
[15:13] <bac> henninge: thanks for the reminder about pre-imp calls for contributed work.  it really is vital.
[15:13] <gary_poster> (Yeah, I didn't know when or if to bring that up)
[15:13] <bac> anything else on henninge's topic?
[15:13] <bac> gary_poster: wait for peanut galllery...
[15:14] <gary_poster> :-) +1
[15:14] <bac> [topic] mentoring update.
[15:14] <MootBot> New Topic:  mentoring update.
[15:14] <bac> jelmer_: are you getting a sufficient number of non-soyuz branches?
[15:14] <jelmer_> bac: Yep
[15:14] <jelmer_> bac: Last weeks shift was quite busy with other code.
[15:15] <bac> great.
[15:15] <bac> [topic] peanut gallery
[15:15] <MootBot> New Topic:  peanut gallery
[15:15] <bac> other issues?  gary?
[15:15] <abentley> I have one.
[15:15] <gary_poster> thanks
[15:15] <gary_poster> ok
[15:15] <gary_poster> go ahead :-)
[15:16] <bac> abentley: go ahead with yours
[15:16] <abentley> I'm noticing we're getting some stuff under lp/ that isn't teams.
[15:16] <abentley> ie lp/buildmaster and lp/poppy.
[15:16] <abentley> I thought items like that were supposed to go under lp/services.  Am I misunderstanding?
[15:17] <sinzui> If they are not user visible, I think they are services
[15:17] <jtv> There's also answers & blueprints of course.
[15:17] <jtv> So maybe that should be "that aren't tabs"
[15:17] <sinzui> They are user visible and are applications so they are not services
[15:17] <abentley> jtv, I could say "applications", I guess.
[15:18] <jtv> Not that that diminishes your point.
[15:18] <bac> abentley: well, then that would move registry
[15:18] <sinzui> Services are expected to be requires by one or applications
[15:18] <jtv> Didn't we go over this on one of the mailing lists?  ISTRM "coop" as the proposed solution.
[15:18] <abentley> jtv, isn't that for shared code that is part of applications?
[15:19] <jtv> True.
[15:19] <abentley> poppy is an sftp server, buildmaster part of the build farm.
[15:19] <sinzui> coop is for application-to-application code. cross domain code
[15:19] <abentley> sinzui, I'm not sure what "they" meant" in "they are user-visible".
[15:19] <jtv> What I see under lib/lp is all nicely stand-alone stuff.  Maybe we just shouldn't worry about this until it becomes a problem again.
[15:20] <abentley> jtv, well, it seems weird that the jobs system and the build system are in different places.
[15:20] <sinzui> is users cannot see poppy, no api, web, rss, I think it may be a service
[15:21] <jtv> Well, the jobs system is much more generic isn't it?  For instance the build system makes use of it.  So for that, the current structure makes sense to me.
[15:22] <jtv> I don't know about poppy but buildmaster is a beast that may not fit comfortably in the same cage as worlddata or the jobs system.
[15:22] <abentley> jtv, "makes use of it" is a bit strong.  There is a bit of model/interface sharing.
[15:23] <mars> sinzui, abentley, maybe punt to BjornT?  Those systems are interfaces to Launchpad, just not HTTP-based ones.  He can make the call as to which interfaces are at the root of the tree.
[15:23] <jtv> abentley: Job plays a pretty central role in buildmaster.
[15:24] <abentley> jtv, it plays a role, but it's an entirely different role to the one it plays in the jobs system.  The equivalent of Job is Build.
[15:24] <bac> mars: i agree, in discussion with bigjools.
[15:25] <bac> BjornT: will you take that as an action item?
[15:26] <BjornT> bac: sure. if someone sends a mail to the list summarizing the issue
[15:27] <bac> well the issue seems pretty straightforward.  we need a decision about what belongs in lib/lp and if poppy and buildmaster are inappropriate for there where should they move.
[15:28] <gary_poster> OK, I'm taking that as a "move to next item" marker.
[15:28] <gary_poster> Ursula and I got an action item both from the reviewers meeting and from the team lead meeting to make orphan branches (those merged to PQM without a bug) "illegal" somehow.  In the team lead call we also talked about handling branches that are incremental branches towards a bug that can't be QA'd individually.
[15:28] <gary_poster> Ursula and I came up with a solution that we think will address these concerns.  We will send an email to launchpad-dev about this solution.
[15:28] <bac> [action] Bjornt to set a policy on what can live in lib/lp, lib/services, and lib/coop
[15:28] <MootBot> ACTION received:  Bjornt to set a policy on what can live in lib/lp, lib/services, and lib/coop
[15:28] <bac> thanks abentley for bringing it up
[15:28] <gary_poster> ok, was slightly early :-)
[15:28] <abentley> bac, np
[15:29] <bac> gary_poster: done?
[15:29] <gary_poster> bac, yes, unless you want me to summarize.  The summary takes a while.
[15:29] <bac> nope
[15:29] <bac> thanks
[15:30] <bac> that seems to be it.
[15:30] <bac> thanks for coming everyone.
[15:30] <bac> #endmeeting
[15:30] <MootBot> Meeting finished at 09:30.

asiapac

[22:53] <bac> rockstar, thumper: meeting?
[22:53] <rockstar> bac, sure.
[22:54] <bac> how's it going rockstar?  progress on the van?
[22:55] <rockstar> bac, yea, enough progress to know I want to pay someone to do some of the progress...
[22:55] <bac> :)
[22:55] <rockstar> bac, body work is hard, so I'm going shopping.
[22:55] <bac> probably a wise move
[22:55] <bac> is tim around?  thumper?
[22:56] <bac> well, i guess not.
[22:56] <bac> not much happened at the AMEU meeting today.
[22:57] <bac> abentley brought up the fact that lib/lp is getting lots of new stuff in it and he doesn't think it is the proper location.
[22:57] <bac> poppy and buildfarmmaster were his two examples
[22:57] <wgrant> Well, buildmaster needs to become lib/lp/services/buildfarm eventually.
[22:57] <wgrant> Not sure about poppy.
[22:57] <bac> wgrant: yeah, moving to services was the idea
[22:58] <rockstar> wgrant, have you a branch that does that?
[22:58] <wgrant> rockstar: No. There was also a proposal to remove it from the tree entirely.
[22:58] <thumper> hi
[22:58] <bac> we assigned bjornt the task of finding homes for those things and creating a policy for what should live in lib/lp lib/lp/services and lib/lp/coop
[22:58] <bac> hi thumper
[22:58] <rockstar> wgrant, what was the reasoning for removing it from the tree?
[22:58] <thumper> I hate lp.coop
[22:58] <thumper> it is a stupid name
[22:59] <thumper> that is why I've not put things there
[22:59] <bac> yes, i think francis blessed us with that one
[22:59] <rockstar> thumper, you are so rational.  :)
[22:59] <wgrant> rockstar: The tree is big, and lp.buildmaster shouldn't really depend on too much.
[22:59] <bac> the tree is too big.  no more code.
[22:59] <thumper> I get sad when I see new files in lib/canonical/launchpad
[23:00] <bac> thumper: yep
[23:00] <thumper> which I saw a lot of last cycle
[23:00] <thumper> I've started moving the code security adapters into lp.code.security
[23:00] <thumper> other teams should do the same
[23:00] <wgrant> Well, but a bit of stuff has been removed from lib/c/l the last couple of cycles. It's looking better.
[23:00] <thumper> I also started moving some tales adapters into lp.app.browser
[23:00] <bac> yeah, we've moved a lot of the JS out, thanks to rockstar's example
[23:00]  * rockstar does a little dance
[23:01] <bac> that bug is highly abused, though
[23:01] <thumper> which bug?
[23:01] <bac> the one for moving JS.  it gets the full cycle of QA tags each release.
[23:01] <thumper> ah
[23:01] <bac> hope two teams never tackle it the same release
[23:02] <bac> so, y'all have anything new, review-wise?
[23:02] <thumper> yep
[23:02] <thumper> during review we should strongly suggest moving code to lp namespace from canonical.launchpad
[23:02] <bac> it doesn't affect you, but the ameu schedule has been shuffled about to ensure better coverage
[23:02] <thumper> even if it makes the branch bigger
[23:04] <bac> [action] bac to discuss moving code out of c/l in AMEU reviewers meeting per thumper's suggestion
[23:04] <thumper> ta
[23:04] <bac> oops, i forgot to wake up mootbot
[23:05] <bac> anything else?
[23:05] <thumper> I don't think so
[23:05] <bac> me neither.  you guys have a good day/evening.
[23:06] <thumper> ciao

ReviewerMeeting20100602 (last edited 2010-06-09 13:05:13 by bac)