16:30:19 <rutsky> #startmeeting weekly
16:30:19 <bb-supy> Meeting started Tue Mar 29 16:30:19 2016 UTC and is due to finish in 60 minutes.  The chair is rutsky. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:30:19 <bb-supy> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:30:19 <bb-supy> The meeting name has been set to 'weekly'
16:30:27 <rutsky> #topic Introduction
16:30:39 <rutsky> Agenda: https://titanpad.com/buildbot-agenda
16:30:40 <infobob> http://paste.pound-python.org/show/1YUoHMrWtjNnF2Ipcx52/ (repasted for rutsky)
16:30:48 <rutsky> #topic Bug 2340
16:31:06 <rutsky> I haven't made any significant progress during this week, sorry
16:31:28 <rutsky> so beta8 should be postponed for at least a week
16:31:37 <rutsky> any questions about bug 2340?
16:32:13 <rutsky> #info worker transition not finished yet, beta8 postponed
16:33:01 <rutsky> ok, looks like no questions for 2340 :)
16:33:02 <rutsky> #topic CLI Tool Update
16:33:08 <rutsky> verm__: ?
16:33:56 <rutsky> tardyp: you here?
16:34:27 <tardyp> hi
16:34:46 <rutsky> hi! glad too see that I'm not alone here :)
16:34:57 <tardyp> hehe :p
16:35:14 <gracinet> I'm here (nothing to say on CLI)
16:35:20 <verm__> yep i'm here now
16:35:28 <verm__> the CLI is going great
16:35:35 <verm__> i got realtime updating working via websockets
16:35:51 <verm__> working on packing it using distutils so it can be installed as a library + binary
16:36:03 <rutsky> verm__: have you published current WIP somewhere?
16:36:09 <verm__> no but that is the next step
16:36:27 <tardyp> verm__: http://docs.openstack.org/developer/pbr/
16:36:47 <tardyp> I recommend this for new and simple packages..
16:36:51 <tardyp> instead of distutils
16:37:01 <verm__> tardyp: looks fantastic i will use this
16:37:01 <rutsky> #info CLI tool works with websockets
16:37:23 <rutsky> #info @verm working on packaging of CLI tool
16:37:43 <rutsky> #info CLI tool will use PBR for packaging
16:37:52 <rutsky> anything else about CLI tool?
16:38:09 <verm__> only 'tail' mode is implemented
16:38:21 <verm__> but there will be a 'top' mode as well
16:38:27 <verm__> that's it otherwise
16:38:43 <rutsky> tail? top? What CLI does again? :)
16:38:58 <verm__> it's a commandline replacement for the web view
16:39:13 <verm__> 'top' mode will show you build activity similar to 'top' but builders or workers
16:39:28 <verm__> 'tail' mode will give you realtime updates as they come in for builders and workers
16:39:55 <rutsky> how this tool will be used and by whom?
16:40:11 <rutsky> e.g. for automating status reporing to other systems?
16:40:14 <verm__> well it's commandline.. so in a terminal?
16:40:23 <verm__> no it's a replacement for the web view
16:40:37 <rutsky> ah, so it's like "top" program, but for builds?
16:40:47 <verm__> yes and 'tail'
16:40:53 <tardyp> looks cool!
16:40:56 <rutsky> sounds interesting!
16:40:59 <verm__> you can interact in other ways, too
16:41:11 <verm__> eg, search for builders, list by tags
16:41:26 <gracinet> I imagine using it over SSH if web view not public and there's no VPN
16:41:36 <verm__> yep, exactly
16:41:56 <verm__> i have a design for a multi-master mode too so you can link in builds from multiple different buildbot instances
16:42:14 <verm__> this way you can keep track of multiple projects at once
16:42:27 <gracinet> sounds good :-)
16:42:37 <verm__> i started it because i had more than a few people find the website view too 'heavy'
16:42:46 <verm__> not my opinon i love it
16:43:05 <rutsky> #info CLI tool is like Web UI, but for terminal. Can be used for example on headless servers in terminal over SSH.
16:43:27 <rutsky> ok, movin to next topic?
16:43:32 <verm__> yep
16:43:39 <rutsky> #topic Beta 8 release
16:43:51 <rutsky> tardyp: Sorry, I haven't finished worker transition yet
16:43:56 <tardyp> waiting for 2340
16:44:02 <rutsky> so we need to postpone beta8
16:44:20 <rutsky> #info Waiting for worker transition completion (bug 2340)
16:44:21 <tardyp> Do you have known issue?
16:44:23 <gracinet> currently, if one installs beta7 and then switches to master, it requires rebuilding JS UI
16:44:29 <tardyp> for me, master as it is work
16:44:41 <rutsky> tardyp: no, just haven't had any time to work on it last week
16:44:54 <tardyp> and beta7 has non-working forceui
16:45:15 <tardyp> so maybe we can do beta8 this week, even if 2340 is not completly done,
16:45:25 <tardyp> and we have some relinquant slave words in the code
16:45:44 <tardyp> rutsky: you are the one who can tell, as I dont know the details
16:45:44 <rutsky> tardyp: this is an option, since I haven't updated docs yet
16:46:35 <tardyp> so we go for beta this week?
16:46:42 <rutsky> I haven't touched master and docs yet --- they should work without buildbot-worker module
16:46:54 <rutsky> yes, let's do it
16:47:02 <rutsky> tardyp: can you do beta8 release?
16:47:04 <tardyp> ok lets write the #decision
16:47:08 <tardyp> I'll do it
16:47:24 <rutsky> #agreed release beta8 without buildbot-worker
16:47:49 <rutsky> #info beta8 will have fixes for Force Build UI
16:48:06 <tardyp> and tons of other stuff
16:48:21 <rutsky> #action @tardyp will run beta8 release
16:48:29 <tardyp> but beta7 is barely unusable because of that bug
16:48:45 <tardyp> ok next topic
16:49:06 <rutsky> #topic InnoDB support
16:49:13 <rutsky> I see this topic in agenda
16:49:18 <rutsky> innodb PR? just a few reduction of varchar size is enough to be compatible
16:49:41 <rutsky> do we have something to discuss here?
16:49:43 <tardyp> this PR was merged today
16:50:09 <tardyp> My collegues have just rebased, and started using buildbot with innodb.
16:50:28 <tardyp> we will see if there are more fixes needed
16:50:34 <rutsky> #info InfoDB support merged. Thanks to @yetanotherion!
16:50:47 <rutsky> tardyp: what kind of fixes?
16:50:52 <tardyp> there might be some FK bugs remaining
16:51:09 <tardyp> I know that gracinet has found one, I've got the time to look at it yet
16:51:29 <tardyp> or any subtleties related to innodb
16:51:43 <rutsky> #info Probably there are some Foreign Key bugs in Buildbot
16:51:46 <gracinet> yes #3504, tried to reproduce it today
16:52:08 <gracinet> had another weird issue in the way
16:52:22 <bb-trac> [trac] #3504/undecided (new) updated by rutsky (empty comment) http://trac.buildbot.net/ticket/3504
16:52:42 <bb-trac> [trac] #3504/undecided (new) updated by tardyp (empty comment) http://trac.buildbot.net/ticket/3504
16:52:57 <tardyp> ahah
16:53:05 <gracinet> you both did it at the same time ?
16:53:18 <rutsky> good teamwork :)
16:53:37 <tardyp> I added miletsone, he fixed the description :)
16:53:51 <gracinet> ok no conflict :-)
16:53:52 <rutsky> gracinet: will you investigate this issue?
16:54:22 <gracinet> yes, I will, but it'll also be a matter of decisions (what we really want to remove, that kind of thing)
16:54:45 <rutsky> #action @gracinet will try to investigate FK bug with PostgreSQL (#3504)
16:54:52 <rutsky> thanks. gracinet!
16:54:59 <gracinet> welcome
16:55:16 <rutsky> anything else about FK/InnoDB?
16:55:50 <tardyp> no
16:55:52 <rutsky> #topic PBR in Buildbot (#3493)
16:56:04 <rutsky> I wrote my opinion in ticket: http://trac.buildbot.net/ticket/3493
16:56:18 <rutsky> We discussed this on last meeting, but haven't changed ticket status.
16:56:23 <tardyp> yes, I agree we have too much needs at least for main packages
16:56:36 <rutsky> but we can try to use PBR for smaller packages
16:56:37 <tardyp> and also it is supporting only the one package per git
16:56:43 <rutsky> oh...
16:57:00 <rutsky> this is serious limitation for BB subprojects
16:57:21 <cmouse> *sigh*
16:57:25 <cmouse> I should followup on my trac account
16:57:27 <cmouse> wonder where it left
16:57:28 <tardyp> it would be good to split our gits, but that is also very practical for maintainance, and also to make sure we merge dependent things together
16:57:57 <tardyp> (I mean having several packages in one git)
16:58:12 <tardyp> so for me we can close wontfix
16:58:29 <rutsky> tardyp: do suggest to split current git on several repos?
16:58:54 <rutsky> #agreed we can't use PBR with main Buildbot parts due to PBR limitations
16:59:01 <gracinet> would several implementations of buildbot-worker make sense at some point ?
16:59:39 <tardyp> gracinet: the original goal of making a better worker/master RPC is to be able to implement simpler worker
16:59:46 <tardyp> in other langage than python
16:59:50 <tardyp> for example for embedded
17:00:17 <tardyp> but as of now, we only have one protocol, which is heavy twisted based
17:00:19 <bb-trac> [trac] #3493/enhancement (closed) updated by rutsky (PBR is good, but we can't use it with main Buildbot parts.) http://trac.buildbot.net/ticket/3493
17:00:26 <gracinet> yeah, in rust, for instance (/me big fan)
17:00:35 <tardyp> I started one protocol based on crossbar.io and wamp, but that is not yet finished
17:00:35 <cmouse> at least trac is still broken
17:01:05 <rutsky> cmouse: do you have problems with your Trac account?
17:01:09 <cmouse> still
17:01:14 <cmouse> i just followed up on the issue
17:01:29 <tardyp> cmouse: can we discuss it after the meeting?
17:01:34 <cmouse> yes, no rush.
17:01:39 <cmouse> it's not going anywhere =)
17:02:26 <tardyp> next topic?
17:03:32 <tardyp> #topic development week
17:04:02 <tardyp> There has been a bunch of PR this week, I'm trying to keep the review quick
17:04:21 <rutsky_> sorry, disconnected
17:04:39 <tardyp> 3 <@tardyp> #topic development week
17:04:39 <tardyp> 19:04 <@tardyp> There has been a bunch of PR this week, I'm trying to keep the review quick
17:05:15 <tardyp> ahh but you changed nick..
17:05:25 <tardyp> so I am not sure the bot will recognise you..
17:05:31 <rutsky_> yep...
17:05:38 <rutsky_> let's wait for timeout...
17:05:39 <rutsky_> #topic Development Week In Review
17:05:47 <cmouse> so anyone who just has nick rutsky is rutsky? =)
17:06:02 <tardyp> there is nickserv for that
17:06:09 <cmouse> ah indeed, carry on.. =)
17:06:18 <rutsky_> I even registered there, but not on this machine...
17:06:33 <tardyp> https://www.irccloud.com
17:06:36 <rutsky_> lets do Week-in-review meantime
17:06:39 <tardyp> its fine service..
17:07:17 <rutsky_> I'm interested in import reorder
17:07:57 <rutsky_> tardyp: what tool you used?
17:07:58 <rutsky> #topic Development Week In Review
17:08:00 <cmouse> ah, there it goes =)
17:08:01 <tardyp> I merged the reorder this morning
17:08:37 <tardyp> sa2ajj__: used its keyboard as a tool
17:08:56 <tardyp> and I used fiximports.py which is automatic
17:09:01 <rutsky> heh)
17:09:12 <rutsky> does pyflakes/pylint happy?
17:09:20 <tardyp> but does not implement some rules that sa2ajj__ cares about (like putting stdlib on top)
17:09:27 <tardyp> I dont personally have big opinion
17:09:48 <rutsky> #info Huge import reordering was merged
17:09:52 <tardyp> appart that if it is not automated, then it will deviate after time
17:10:23 <tardyp> then sa2ajj__ has some python2.6 cleans up that he needs to rebase, so that we can review
17:10:47 <tardyp> I personally would like to discuss BytesIO vs StringIO, because I'm not sure of the exact equivalence
17:10:57 <tardyp> (but in the PR, not in thie meeting)
17:11:27 <tardyp> I also merged the big HttpPush porting PR
17:11:37 <tardyp> which give github support to nine
17:11:41 <tardyp> (and stash)
17:12:12 <tardyp> we still have to port gitlab and other http based reports, but this will be much easier due to factorization in HttpPushBase
17:12:24 <rutsky> #info HttpStatusPush, GithubStatusPush and StashStatusPush were ported to nine by Pierre
17:12:35 <tardyp> (I'll let each community of those tools to handle it)
17:12:53 <gracinet> tardyp: great, I'd be interested in Kallithea integration myself (also a SFC project btw)
17:13:24 <gracinet> yet later
17:13:27 <tardyp> gracinet: cool!
17:14:28 <tardyp> also some optimization of the UI .
17:14:47 <tardyp> the change page for example is merged, and another optimization for the trigger build step is in review
17:15:01 <rutsky> #info Pierre merged bunch of Web UI improvements (Force UI, recent changes optimizations and others)
17:15:24 <tardyp> optimization for scalability
17:15:35 <tardyp> and then next big PR from me will be the raml PR
17:16:01 <tardyp> which rewrites a bunch of data api doc in raml in order to be able to make sure it is coherent with implementation
17:16:20 <tardyp> we have too much obsolete doc
17:16:43 <tardyp> https://github.com/buildbot/buildbot/pull/1949
17:16:50 <rutsky> #info Pierre's RAML PR is upcoming
17:17:45 <rutsky> tardyp: am I correct that we will be able to load RAML specification into Web editor and run requests from it?
17:18:06 <rutsky> like swagger-ui, but for RAML (don't remember how it called)
17:18:19 <tardyp> well the tooling for raml is not really ready for that
17:18:23 <tardyp> but eventually yes
17:18:41 <tardyp> for now, I am concentrating on doc generation
17:19:10 <tardyp> its based on raml1.0, which is not even released, but it is very promising
17:19:34 <rutsky> #info sa2ajj implemented CMake steps
17:19:44 <tardyp> I originally wanted to use swagger, but writting a spec in json is very very annoying for me
17:20:03 <rutsky> current RAML visualizations looks pretty cool (regarding using it for documentation)
17:20:28 <rutsky> tardyp: we write Swagger specs in YAML
17:20:53 <rutsky> and Swagger tools support YAML AFAIK, at least swagger-ui and default editor
17:21:18 <rutsky> anything else for todays meeting?
17:22:03 <tardyp> rutsky: but yet swagger requires jsonschema for the data type specification
17:22:40 <tardyp> and raml has imports, and lot of syntax sugar
17:22:43 <rutsky> tardyp: yep, bu we write jsonschema in YAML too :)
17:23:13 <rutsky> when I checked RAML I decided that it's not yet ready for our project company where I'm working
17:23:24 <tardyp> yes
17:24:12 <tardyp> its not, this is why I only use it for doc generation, and I have a simple raml parser just for our own subset of raml
17:24:29 <tardyp> and bet on the fact that tooling will improve in 2016
17:24:43 <rutsky> yeah
17:24:57 <rutsky> anything else for today meeting?
17:25:46 <gracinet> rustky: in general ?
17:26:07 <rutsky> gracinet: any topic to discuss during meeting?
17:27:22 <rutsky> we can continue general discussion after end of the meeting.
17:27:33 <gracinet> fine
17:27:54 <rutsky> ok
17:27:56 <rutsky> #endmeeting