17:12:49 <tflink> #startmeeting fedoraqa-devel
17:12:49 <zodbot> Meeting started Mon Jan 13 17:12:49 2014 UTC.  The chair is tflink. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:12:49 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
17:12:49 <tflink> #meetingname fedoraqa-devel
17:12:49 <zodbot> The meeting name has been set to 'fedoraqa-devel'
17:12:50 <tflink> #topic Roll Call
17:13:01 * Viking-Ice here
17:13:02 <danofsatx> aqui
17:13:29 * kparal is here
17:13:38 * mkrizek is here
17:14:24 <kparal> pschindl: ping
17:14:31 * roshi is here
17:14:36 * pschindl is here
17:14:47 <tflink> let's wait a few minutes for folks to show up since my schedule-fu seems to be completely off today
17:14:52 <adamw> ahoyhoy
17:17:12 <tflink> ok, I think we have most of the folks I was expecting
17:17:23 <tflink> kparal: any idea if jskladan is planning to attend?
17:17:36 <kparal> I was about to ask the same :)
17:17:40 <kparal> I don't know
17:18:43 <tflink> I suspect not at this point, I'll have to sync up with him later, I guess
17:19:29 <tflink> #topic Taskotron Status
17:19:46 <tflink> pretty much picking up where we left off last week
17:19:51 <tflink> as a reminder
17:20:07 <tflink> #info Taskotron Phase 1 is due to be delivered 2014-03-01
17:20:35 <tflink> #link https://fedoraproject.org/wiki/User:Tflink/taskotron_development_plan#Phase_1:_AutoQA_Replacement
17:21:38 <tflink> phase 0 is mostly complete, tickets need a bit of updating, still
17:21:46 <tflink> #info phase 0 is mostly complete, tickets need a bit of updating, still
17:21:52 <kparal> so, what about depcheck?
17:21:54 <tflink> #link https://phab.qadevel.cloud.fedoraproject.org/T7
17:22:40 * tflink is almost done with the notifications stuff, needs to get a couple of fedmsg things figured out - will pester threebean about it later today
17:22:40 <kparal> I guess I could try to write something dead simple based on repoclosure
17:22:55 <tflink> kparal: handsome_pirate has been working on it
17:23:50 <handsome_pirate> How long has this been going?
17:23:54 <tflink> a few minutes
17:24:00 <handsome_pirate> So, I haven't missed much?
17:24:03 <tflink> nope
17:24:07 <kparal> handsome_pirate: do you have some depcheck-ng news?
17:24:08 <handsome_pirate> Coolio
17:24:14 <Viking-Ice> handsome_pirate, just everything important
17:24:20 <handsome_pirate> lol
17:24:26 <handsome_pirate> kparal:  Give me a sec
17:24:34 <tflink> so much for my plan
17:24:35 <Viking-Ice> handsome_pirate, oh we also signed you up for bunch of stuff
17:24:37 <tflink> but we can do this now
17:24:47 <tflink> #topic Depcheck rewrite status
17:25:00 <kparal> tflink: feel free to organize the meeting according to you, not us :)
17:25:30 <handsome_pirate> https://bitbucket.org/fedoraqa/depcheck-mk-2
17:25:38 <handsome_pirate> depcheck mk 2 lives there
17:25:45 * handsome_pirate needs to do another push
17:26:02 <handsome_pirate> It's based on libsolv for the actual dependency solving back end, same as dnf
17:26:21 <tflink> handsome_pirate: have you gone through the depcheck scenarios yet to make sure it passes?
17:26:30 <handsome_pirate> Unlike the original depcheck, I'm not twisting it to do my bidding, but, rather, using it the way it is meant to be used
17:26:38 <handsome_pirate> tflink:  About half of them
17:26:53 <handsome_pirate> tflink:  Most of them it passes; three did not
17:27:08 <handsome_pirate> I'm working on those three before I continue
17:27:17 <handsome_pirate> Two were conflicts related
17:27:20 <tflink> any guesses on eta?
17:27:20 <kparal> do we have a link for those scenarios?
17:27:36 <handsome_pirate> kparal:  https://bitbucket.org/fedoraqa/depcheck_scenarios
17:27:38 <tflink> #info initial depcheck scenarios have been coded up
17:27:44 <tflink> #link https://bitbucket.org/fedoraqa/depcheck_scenarios
17:27:50 <handsome_pirate> tflink:  Friday is my goal for my next major git push
17:28:32 <handsome_pirate> tflink:  One of the failures looks to be with libsolv; I'm working on that, as well
17:28:44 <handsome_pirate> The other two need to be handled with some logic in depcheck
17:28:53 <tflink> did you ever attempt integration with the runner demo code?
17:29:19 <handsome_pirate> I have started, but need to finish
17:29:41 <handsome_pirate> I'll try to get that done by Friday, as well
17:29:58 <tflink> #info current depcheck-mk-2 code is passing some of the depcheck scenarios, rest are being worked on
17:30:27 <kparal> better progress than I expected, good :)
17:30:42 <tflink> as a side note, f18 goes eol tomorrow IIRC
17:30:55 <tflink> current autoqa depcheck doesn't run on anything newer than f18
17:31:11 <handsome_pirate> Yeah, I had twelve hours riding a train this past weekend to work on this
17:31:13 <tflink> but we should be OK leaving the test clients at f18 for a little while
17:31:29 <kparal> tflink: I guess, or disable the test completely
17:31:30 <tflink> er, autoqa test clients
17:31:40 <tflink> eh, it works as well as it's ever worked
17:32:34 <tflink> handsome_pirate: any additional things that you want to add for depcheck?
17:32:37 <handsome_pirate> Once I get through all the scenarios and get it so it runs as a taskotron task, I'm going to work on getting proper output
17:33:04 <handsome_pirate> My plan for next week is to work on outputting tap
17:33:14 <tflink> handsome_pirate: unit tests and documentation?
17:33:28 <handsome_pirate> I guess I should start working on them, as well
17:33:51 <handsome_pirate> Docs at this point are a somewhat outdated README stub and if you run depcheck -h
17:34:17 <handsome_pirate> I do run it through pylint every once in a while and fix the biggest complaints
17:34:45 <tflink> good to hear
17:34:57 <tflink> any other updates?
17:35:05 <handsome_pirate> That's it for right now
17:35:12 <tflink> ok, let's move on
17:35:15 <handsome_pirate> I have the next two weeks of work fairly well mapped out
17:35:32 <tflink> #topic Documentation
17:36:11 <tflink> I think it became pretty clear to me and everyone else that I'm a bit farther behind on documentation (design docs and usage) than I thought I was
17:36:40 <tflink> I wanted to cover the things that aren't waiting on me for docs first, to make sure they got covered and this is one
17:36:54 <tflink> roshi has volunteered to look into documentation systems for taskotron
17:37:09 * roshi waves
17:37:11 <tflink> #link https://phab.qadevel.cloud.fedoraproject.org/T39
17:37:20 <tflink> #chair kparal roshi
17:37:20 <zodbot> Current chairs: kparal roshi tflink
17:37:42 <handsome_pirate> tflink:  If this is the case, should I hold off on too much docs for depcheck so I can be sure to match the rest of taskotron?
17:37:51 <tflink> the idea is to have a system for building docs that can be run from CI and can work for all types of docs we need
17:38:02 <danofsatx> I'll help with doc writing where I can...just point me to a task
17:38:03 <Viking-Ice> yeah we need to keep this update for at least baseWG which probably will be the only valid WG effort
17:38:07 <tflink> handsome_pirate: if you write it in generic ReSt, it should be OK
17:38:07 <roshi> handsome_pirate: if you write them in md, rst then there shouldn't be an issue
17:38:17 <handsome_pirate> Roger
17:38:20 <roshi> I'll ping you danofsatx :)
17:38:31 <tflink> roshi: I realize you just started, but any updates?
17:38:47 <tflink> or guesses on when you might have something to show?
17:38:57 <roshi> the Docs team is also addressing this issue of how to get shorter tutorials and whatnot - so I'm hoping we can integrate with them easy
17:39:15 <roshi> looking into dexy (http://dexy.it) right now
17:39:26 <Viking-Ice> outside wiki content?
17:39:31 <roshi> it seems to fit what we're looking for
17:39:32 <tflink> yeah
17:39:42 <tflink> wiki is great for some things, but not all
17:39:43 <Viking-Ice> it says sexy there on top good enough tool for me
17:39:49 <roshi> haha
17:40:01 <tflink> #info roshi is looking into using dexy
17:40:09 <tflink> #link http://dexy.it
17:40:15 <Viking-Ice> has it been packaged for the distro?
17:40:19 <handsome_pirate> lol
17:40:40 <roshi> it's in pip
17:41:02 <roshi> if we need a fedora package for it - I would be interested in that (never done that before)
17:41:02 <Viking-Ice> I thought we had to ship that ourselves
17:41:07 <tflink> eventually, we will be populating qadevel with docs for qadevel projects
17:41:32 <roshi> tflink: perhaps it's a good idea to go over the vision for qadevel docs and whatnot
17:41:45 <tflink> yeah, we'd eventually need to have any tool packaged
17:41:51 <tflink> any tool we use
17:41:56 <Viking-Ice> I would prefer we have unified documentation for QA community in general
17:42:11 <tflink> roshi: you want to go over that or should I?
17:42:12 <Viking-Ice> so if doxy is the future I think we should move all doc from the wiki to that
17:42:21 <roshi> as would I Viking-Ice
17:42:32 <roshi> sure, as far as I know it :)
17:42:43 <tflink> we can discuss that, but I'm not sure that we really need to be moving everything out of the wiki
17:42:53 <tflink> we -> wider qa community, not just the folks here
17:43:10 <Viking-Ice> right but fragmented way is not the way to go
17:43:14 <roshi> with taskotron and other projects coming down the pipe for qa devel we will have a need for more and more "getting started" and tutorial documentation for all the tools
17:43:21 <Viking-Ice> so one landing place one documentation place etc
17:43:40 <roshi> I'll get there in a second Viking-Ice - but you're right
17:43:48 <tflink> Viking-Ice: I don't see the problem with using different tools for different purposes if that ends up making sense
17:44:17 <roshi> we need a tool that docs can be written along with the code for the project, something that's easy to write and easy to integrate into CI
17:44:19 <Viking-Ice> tflink, you dont want to have our community members running around half the internet to start participating
17:44:26 <roshi> dexy fits that pretty well
17:44:36 <Viking-Ice> tflink, but anyway we will cross that bridge when we get there
17:44:44 <tflink> agreed
17:44:54 <roshi> the Docs team is also trying to find a means of doing similar things with the Guide
17:45:40 <roshi> so in a perfect world, any docs we write for QA in general (tutorials, etc) they would also be able to use
17:45:54 * handsome_pirate is +1 for this
17:46:08 <roshi> this documentation for qadevel is a good proving ground for testing a new documentation system
17:46:16 <Viking-Ice> right
17:46:26 <roshi> does that about cover it tflink? or am I off?
17:46:30 <tflink> roshi: where they overlap, sure
17:46:34 <roshi> right
17:46:59 <roshi> it would be along the lines of "Hey, we have these tuts we use. You can easily use them if yo ulike."
17:47:02 <tflink> just a couple of things that I may not have mentioned when we talked about this earlier
17:48:35 <tflink> within the scope of qadevel docs, there are three main types that I want to cover: user docs (tutorials, triage guide), task maintainer docs (how to use the bits and pieces) and devel docs (how to write code, extensions etc.)
17:49:03 <tflink> docs live with code, separate them and they tend not to get updated
17:49:15 <tflink> have a higher tendancy to not get updated, rather
17:49:58 <tflink> the idea for qadevel stuff is to have a method for taking those docs, building them and hosting them in a common place
17:50:29 <tflink> and use as common of a toolset as we can so that we're not learning 5 different things
17:51:02 <tflink> so on any change, CI would pull in changes from the main docs repo and any child repos, build and deploy those docs on qadevel
17:51:21 <tflink> roshi: does that match what you understood from earlier?
17:51:26 <roshi> yep
17:51:34 * handsome_pirate is +1 for this
17:51:34 <tflink> any questions within the scope of docs for qadevel?
17:52:22 * tflink is going to start using the system to write up design docs and the other bits that need to be documented now while roshi is working on the method for making those docs into a more human understandable format (ie, html)
17:52:44 <roshi> when you say "system" what are you meaning?
17:52:56 <tflink> rst -> html
17:53:07 <tflink> if we deploy manually at first, I'm OK with that
17:53:22 <roshi> ok
17:53:52 <tflink> roshi: anything else in the way of updates or questions?
17:54:21 <roshi> not at this point - fleshing out dexy atm, looking for anything else that might be better
17:54:32 <roshi> if anyone has suggestions, let me know :)
17:55:01 <roshi> then I'll write up a blog post with a comparison and send it to the list
17:55:12 <roshi> right now I think we've talked about dexy and sphinx
17:55:38 <tflink> #action roshi to write up comparison between dexy and sphinx, send out to list when complete
17:56:10 <tflink> if there's nothing else on this topic, let's move on to tasks/checks
17:56:32 <tflink> #topic Taskotron Tasks/Checks
17:56:49 <tflink> This is one large area of work that I don't think is stuck waiting on me to write stuff up
17:57:18 <tflink> #info initial tasks will be the checks from AutoQA PUATP (minus depcheck, which is being rewritten)
17:57:38 <kparal> should we distribute the tasks now?
17:57:49 <tflink> #info these checks are: rpmlint, upgradepath and possibly repoclosure
17:57:52 <kparal> I might be a good candidate for upgradepath rewrite
17:58:07 * handsome_pirate is already busy with depcheck
17:58:20 <tflink> kparal: yeah, I was planning to ask for volunteers and hoping you would be willing
17:58:37 <tflink> rpmlint is mostly done - I've been using that as the demo task
17:59:21 <tflink> the basic idea is to extract the checks and any needed common funtions from autoqa so that it'll run in taskotron
17:59:41 <tflink> bah, I missed one
17:59:47 <tflink> do we want to include rpmguard?
18:00:00 <tflink> it's not being reported in autoqa, but it is implemented
18:00:09 <kparal> I think not for the initial run, let's keep it simple
18:00:50 <tflink> good point
18:00:52 <tflink> #undo
18:00:52 <zodbot> Removing item from minutes: <MeetBot.items.Info object at 0x2f4150d0>
18:01:30 <tflink> #info the target checks for phase 1 are: rpmlint, upgradepath, depcheck (mk 2)
18:01:51 <kparal> the wiki also says repoclosure
18:02:08 <tflink> yeah, I was tempted to leave that one off for phase 1, though
18:02:08 <kparal> but we don't report it right now
18:02:27 <tflink> #action tflink to move repoclosure to later phase in wiki docs
18:02:30 * handsome_pirate thinks we should leave off repoclosure for now
18:02:41 * kparal nods
18:03:33 <tflink> kparal: would you also be interested in/willing to document the scenarios for upgradepath and create some scenarios like jskladan did for depcheck?
18:04:02 * shisatum waves
18:04:08 <shisatum> tflink: is there a simple task I can do?
18:04:57 <kparal> tflink: I'll try
18:05:02 <tflink> shisatum: depends on what your background is (in terms of what you can do or are willing to learn), to be honest
18:05:14 <tflink> #link https://phab.qadevel.cloud.fedoraproject.org/T15
18:05:26 <tflink> #link https://phab.qadevel.cloud.fedoraproject.org/T14
18:06:12 <tflink> for now, let's keep the task git repos in bitbucket
18:06:28 <tflink> I'm not convinced that's going to be their final place, but it works for now
18:06:44 <tflink> and that seems to be where everything else has ended up
18:08:33 <tflink> shisatum: are you OK with talking about tasks for you after the meeting?
18:08:45 <shisatum> tflink: I have novice/intermediate Python skills and a bit of HTML, mostly. My biggest project (in GWBASIC) was a control/macroing program (via keyboard) for a simple robot. I'm still becoming familiar with the Linux landscape, but would like to learn as much as I can as quickly as possible.
18:08:55 <shisatum> tflink: That is perfect.
18:09:26 <tflink> for the record, I'm referring to "tasks" as in the unit of work for taskotron here
18:09:38 <tflink> not just assignments and things for people to work on
18:10:25 <tflink> #info possible checks for a future phase: repoclosure, rpmguard, abicheck
18:10:58 <tflink> alrighty, moving on to the next topic
18:11:10 <handsome_pirate> tflink:  Also, I was asked about depcheck potentially checking to see if all required deps are listed in the rpm spec file
18:11:11 <tflink> #topic Taskotron Runner
18:11:34 <tflink> handsome_pirate: that seems to be rather out of scope for depcheck
18:11:47 <handsome_pirate> tflink:  That's what I thought
18:11:50 <tflink> if I'm understanding correctly, the build system should catch that
18:11:57 <handsome_pirate> Indeed
18:12:05 <handsome_pirate> And, that's what I said
18:12:27 <tflink> at least we had the same answer :)
18:12:35 <handsome_pirate> LOL
18:12:48 <tflink> this still needs some brain-dumping from me, but there are things which can be done
18:12:59 <tflink> #info demo code is public, but incomplete
18:13:15 <tflink> #link https://bitbucket.org/fedoraqa/libtaskotron-demo
18:13:52 <tflink> outside of any common methods that we need to pull in for upgradepath or depcheck, there are 4 major components which need work
18:14:12 <tflink> resultsdb integration, bodhi integration, logging and moving to dynamically loaded modules
18:14:53 <tflink> my goal is to have the runner be distro agnostic by only loading fedora specific stuff if it's called for in the task
18:15:16 <tflink> which also makes virtualenv creation easier if you're not dealing with code for koji or bodhi
18:15:26 <tflink> which are so much fun to install in virtualenvs
18:15:58 <tflink> oh, variable substitution as well - that makes 5 big parts
18:16:42 <tflink> mkrizek: is this something that you'd be interested in helping out on?
18:17:05 <tflink> #action tflink to move repo from libtaskotron-demo to libtaskotron
18:17:16 <mkrizek> tflink: yes, I would
18:18:10 <tflink> mkrizek: ok, I'll get some tickets written up today and send you the list of ones that are complete enough to start working on now
18:18:26 <mkrizek> tflink: sounds good, thanks
18:18:38 <tflink> I think that's about it for the runner right now
18:19:14 <mkrizek> tflink: seems like we compete who's going to create more tickets in phab today :)
18:19:35 <tflink> we have ~ 40 minutes left, I was planning to cover beaker quickly and move on to devops/infra, short term plans and some strategy stuff
18:20:07 <tflink> that leaves out some status stuff, but that's all waiting on me for documentation, anyways so I figured that emails/blog posts/docs could work just as well
18:20:11 <tflink> any objections?
18:20:33 <tflink> mkrizek: yeah, I noticed all the emails to qadevel@ :)
18:20:38 <kparal> no
18:20:43 <tflink> ok, moving on
18:20:56 <tflink> #topic Fedora Beaker Installation
18:21:35 <tflink> I had been originally planning to put this off a bit but after talking to kparal last week, it makes sense to get this moving sooner than later
18:22:17 <tflink> there are multiple groups within RH who want to start using the installation to run their test cases against fedora
18:22:44 <tflink> since this would benefit us, it seems like a good idea to get this more done
18:22:55 <tflink> most of the todo right now is infra-ish work,
18:22:57 <tflink> though
18:23:30 <kparal> I think we should do just the necessary work to allow them to work on this stuff and fix what's necessary. I don't want to imply we should invest heavily into this right now
18:23:37 <tflink> writing an ansible playbook for beaker deployment, getting FAS groups set up for ACLs and getting FAS integration
18:23:54 * kparal saw tflink's emails to infrastructure list
18:23:57 <tflink> kparal: good point. I'm just thinking of getting it to a usable state
18:24:02 <kparal> right
18:24:42 <tflink> #link https://lists.fedoraproject.org/pipermail/infrastructure/2014-January/013885.html
18:24:54 <tflink> that's the email I sent out to the infra list
18:25:29 <tflink> the second part about isolation is going to take some time and I don't see as absolutely needed right now if we control access to the system to folks who are known and trusted
18:26:29 <tflink> any volunteers to work on getting stuff set up and working?
18:26:59 <tflink> the beaker devs have volunteered to help maintain the installation
18:27:11 <handsome_pirate> tflink:  I may have some cycles next week
18:27:19 <handsome_pirate> tflink:  But, I couldn't get to it until then
18:27:27 <tflink> handsome_pirate: I'd rather you stay focused on depcheck, to be honest
18:27:32 <handsome_pirate> tflink:  Roger
18:27:40 <tflink> I'm not saying that you couldn't do it, just that depcheck is a much higher priority
18:28:07 <handsome_pirate> Indeed
18:28:14 <tflink> I assume that silence means no volunteers?
18:28:33 <mkrizek> the question is, does anyone have any experience with ansible
18:28:57 <tflink> other than me? not that I'm aware of
18:29:07 <tflink> which is someting that we need to fix at some point
18:29:11 <mkrizek> yeah
18:29:27 * kparal tries to stay away from infrastructure stuff :)
18:30:22 <tflink> ok, I'll work on this as I have time unless someone comes forward as a volunteer
18:30:43 <tflink> but documentation and getting everything else to the point that they can be handed off is a higher priority
18:31:16 <tflink> at some point, I'll start pestering someone to learn ansible and start helping with infra stuff
18:31:27 <tflink> bus rule and all
18:31:35 <mkrizek> :)
18:31:46 <tflink> if there's nothing else, I'll move on to the next topic
18:32:08 <tflink> #topic Devops and Infrastructure
18:32:11 <danofsatx> I'm about to apply to the infra team, I could take this on
18:32:46 <tflink> danofsatx: awesome, I want to start making things more infra-compliant
18:32:56 <tflink> instead of the current stand-alone stuff that I've been using
18:33:01 <danofsatx> I have to see if they'll accept me though ;)
18:33:39 <tflink> they probably will, especially if you have a mentor lined up
18:34:16 <tflink> there are a few general things I wanted to touch on here
18:35:10 <tflink> first, as I think everyone is aware, we're using phabricator to track tickets and do code reviews
18:35:31 <tflink> #link http://phabricator.org/
18:35:46 <tflink> #link https://phab.qadevel.cloud.fedoraproject.org/
18:36:09 <tflink> it's not perfect, but I couldn't find anything better and it's worlds better than trac+reviewboard
18:36:50 <tflink> after a snafu with jenkins for blockerbugs, I started setting up a buildbot for qadevel projects
18:37:11 <tflink> it's still in staging, though
18:37:16 <tflink> #link https://qadevel-stg.cloud.fedoraproject.org/builds/
18:37:43 <tflink> if we weren't using buildbot as the core of taskotron, I would have just continued to use infra's jenkins instance
18:38:03 <tflink> but it seems stupid to become proficient in both systems
18:38:13 <kparal> what was the problem in jenkins?
18:38:33 <tflink> it was registering passes when the whole build has been failing since september
18:39:00 <tflink> which wasn't a jenkins issue - it had to do with how the tests were being setup and executed
18:39:46 <tflink> but this way, we have some more flexibility in our CI system and we can re-use the experience we'll get with buildbot when we eventually start hacking on buildsteps
18:40:40 <tflink> at the moment, most of the qadevel and taskotron stuff is deployed with ansible playbooks
18:41:22 <tflink> #link https://bitbucket.org/tflink/fedoraqa-ansible
18:41:46 <tflink> it's mostly up to date but at some point, we should move over to infra's playbooks instead of a standalone system
18:42:18 <tflink> if you're interested in working on those playbooks, let me know and I'll get you the needed details to work on it
18:42:46 <tflink> I'm hoping that we'll get some new servers next year - our current infra is all at least 3 years old. most of it is 4 years old
18:43:10 <handsome_pirate> tflink:  Are these complete enough to use to set up a full taskotron system localy for devel?
18:43:22 <tflink> so there will be some planning on that front, but I that's mostly an internal RH issue since they'd be paying for it
18:43:46 <tflink> handsome_pirate: the playbooks? not quite. they're missing variables and the private stuff
18:43:57 <handsome_pirate> tflink:  Roger
18:44:08 <tflink> the docs allude to the variables but IIRC, that doc is out of date
18:44:22 <tflink> #action tflink to update playbooks and docs so that folks can setup dev systems
18:44:52 <tflink> the last infra-ish thing is task git repositories
18:44:57 <handsome_pirate> tflink:  post meeting, mind if I badger you for the info on setting up a devel taskotron instance?
18:45:23 <tflink> the design of taskotron is such that each task has it's own repo and that repo is cloned every time the task is run
18:46:00 <tflink> which means that I'd rather not pound the snot out of bitbucket but I also don't want to move the repos around every time the system is redeployed
18:46:17 <tflink> so the working plan is to have a local mirror of the bitbucket task repos for now
18:46:55 <tflink> so the canonical upstream will be bitbucket and the taskmaster will mirror changes to those upstreams
18:47:04 <tflink> any concerns or questions about that?
18:47:07 <tflink> handsome_pirate: sure
18:47:34 <tflink> ok, moving on
18:47:59 <tflink> #topic Coding Policy/Strategy Going Forward
18:48:19 <tflink> one of the things that I want to start is improving our code quality
18:48:43 <tflink> I'm not placing any blame (I'm as much to blame as anyone) but there was some suboptimal code in autoqa
18:49:01 <kparal> nicely put ;)
18:49:12 <tflink> to keep that from happening again, I'd like to propose some rules going forward
18:49:29 <tflink> 1) all code must be reviewed before being used
18:49:48 <tflink> 2) all projects must be in CI for style checking (pylint) and unit test running
18:50:15 <tflink> 3) all projects (with the exception of tasks) should use the git-flow branching strategy
18:50:25 <tflink> #link http://nvie.com/posts/a-successful-git-branching-model/
18:50:35 <tflink> #link https://github.com/nvie/gitflow
18:50:49 * kparal needs to study that again
18:50:55 <tflink> 4) all code should adhere to a common style guide
18:51:10 <tflink> and the obvious part about exceptions where needed
18:51:40 <kparal> the CI will happen pre-commit or post-commit?
18:51:47 <mkrizek> ad 1), are we going to have minimal number of reviewers that needs to review the code before pushing to develop?
18:52:04 <tflink> kparal: that'll depend on how we set things up
18:52:22 <tflink> mkrizek: ideally, yes but we have a limited number of folks to draw on for reviews
18:52:33 <mkrizek> true
18:53:03 <tflink> this'll need hashing out for the details and some docs on how to do reviews etc. but I wanted to bring this up before drafting anything for the lists
18:53:06 <kparal> we can set up multiple accounts <wink> <wink>
18:53:51 <tflink> oh, another thing that I'll send out to the list (we're short on time) is minimum python version, eventual python3 work and possibly using python-future
18:54:07 <tflink> #link http://python-future.org/index.html
18:54:08 <mkrizek> kparal: :))
18:54:09 <kparal> I just find some pylink requirements stupid. so it depends whether we just check it, or also enforce it
18:54:17 <kparal> *pylint
18:54:26 <tflink> yeah, I suspect that we'll want to have some custom pylint configuration
18:54:52 <tflink> but it sounds like there is general support for the idea?
18:54:54 <kparal> e.g. 80 chars on a line should be enough for everyone :)
18:55:04 <kparal> tflink: yes
18:55:06 <handsome_pirate> tflink:  I'm +1
18:55:13 <tflink> wasn't the standard updated for 120?
18:55:19 <tflink> to 120, rather
18:55:19 <kparal> tflink: dunno
18:55:30 <handsome_pirate> pylint still requires 80
18:55:33 <tflink> either way, it'll be a topic for discussion
18:56:03 <tflink> since we have 5 minutes left, any objections to moving on?
18:56:10 <kparal> no
18:56:12 <tflink> #topic Short Term Plans
18:56:21 <tflink> in a nutshell, this is what I have in mind:
18:56:39 <tflink> #info get a new demo system up and running to start testing taskotron as a whole
18:56:48 <tflink> #info get docs and planning done
18:57:07 <tflink> #info continue with the work described earlier (checks, runner etc.)
18:57:23 * tflink will be locked in a dungeon until the docs are more complete :)
18:57:51 <tflink> any questions here?
18:58:31 <tflink> I realize that I've left out major components but the summary of those is mostly "tflink needs to braindump and write stuff up"
18:58:32 <handsome_pirate> tflink:  At some point, a devel insance for running tasks against koji?
18:58:49 <tflink> handsome_pirate: see the line from 2 minutes ago :-P
18:58:55 <handsome_pirate> DOH
18:59:04 * handsome_pirate goes back into his hole
18:59:10 <tflink> :-D
18:59:19 <tflink> otherwise, it's time for ...
18:59:22 <tflink> #topic Open Floor
18:59:34 <tflink> Any topics that folks want to bring up?
19:00:58 * handsome_pirate is good
19:01:01 * tflink assumes that silence == no
19:01:30 <tflink> I'll be working on docs and tickets, please find me if you have questions
19:01:42 <kparal> thanks
19:01:46 <tflink> we'll schedule another qadevel meeting if/when appropriate
19:02:18 <tflink> if there's nothing else,
19:02:27 * tflink sets fuse for [0,5] minutes
19:02:43 * kparal thinks it's time to have some bacon
19:02:51 <tflink> thanks for coming, everyone. I hope it wasn't too boring
19:03:29 <shisatum> ++
19:04:08 * tflink will send out minutes shortly
19:04:11 <tflink> #endmeeting