16:32:02 <djmitche> #startmeeting weekly
16:32:02 <bb-supy> Meeting started Tue Mar 22 16:32:02 2016 UTC and is due to finish in 60 minutes.  The chair is djmitche. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:32:02 <bb-supy> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:32:02 <bb-supy> The meeting name has been set to 'weekly'
16:32:07 <djmitche> sorry I was fiddling with my phone :)
16:32:11 <djmitche> #topic Introduction
16:32:34 <djmitche> Agenda: https://titanpad.com/buildbot-agenda
16:32:35 <infobob> http://paste.pound-python.org/show/1YUoHMrWtjNnF2Ipcx52/ (repasted for djmitche)
16:33:11 <djmitche> #topic Bug 2340
16:33:26 <djmitche> we've merged another big PR for the worker
16:33:38 <djmitche> plus some cleanup
16:33:39 <djmitche> what else is new?
16:33:59 <rutsky> its WIP currently, there are around 70 "slave" occuriences in current worker and they mostly significant
16:34:17 <rutsky> I hope to finish worker during this week
16:34:25 <djmitche> ok
16:34:28 <rutsky> tardyp: do you want worker for beta8?
16:34:29 <djmitche> what does "significant" mean here?
16:35:08 <rutsky> not "significant", but more like "semantic" --- they are not just strings. but part of communications API etc
16:35:13 <djmitche> ok
16:35:38 <tardyp> hello
16:35:45 <rutsky> hi, tardyp!
16:35:47 <djmitche> #info work in progress on the details of the worker side - about 70 occurrences of "slave" that need careful attention
16:35:48 <tardyp> yes, I think worker is the reason for beta8
16:36:29 <rutsky> then we definitely should wait till I finish :)
16:36:38 <rutsky> I thought that you want beta8 for some other reason
16:36:45 <tardyp> There have been a bunch of other improvements though
16:37:08 <tardyp> I think we should think of beta8 as rc1
16:37:33 <rutsky> what about blocking bugs for 0.9.0?
16:37:37 <rutsky> they all resolved?
16:38:09 <tardyp> no, but they are not blocker for rc
16:38:12 <tardyp> mostly doc bugs
16:38:26 <tardyp> or very minor stuff
16:38:34 <djmitche> #info beta8 will be released when the worker work is finished
16:38:38 <tardyp> unless I missed something
16:38:45 <djmitche> #info other 0.9.0 blockers are minor and documentation bugs
16:39:25 * sa2ajj waves
16:40:01 <djmitche> #topic CLI Tool Update
16:40:05 <djmitche> verm__: ^^?
16:40:17 <bb-trac> [trac] #3468/defect (closed) updated by rutsky (Max depth 4 was merged in master.) http://trac.buildbot.net/ticket/3468
16:41:06 <djmitche> we'll do the week-in-review and can re-visit
16:41:11 <djmitche> #topic Development Week In Review
16:41:22 <bb-github> [13buildbot] 15sa2ajj commented on issue #2058: @rutsky I do it manuall :)  I use ReST quite a lot and these kinds of problems stand out for me and my fingers automatically fix them :) 02https://git.io/vaQuE
16:41:22 <sa2ajj> what is this cli tool?
16:41:40 * sa2ajj went to check agenda...
16:41:54 <djmitche> Amar's working on a command-line tool that consumes the Data API over websockets and can show build progress
16:42:15 <sa2ajj> oh, good
16:42:16 <rutsky> tardyp: in the mean time can you take a look at http://trac.buildbot.net/ticket/3469?
16:42:30 <djmitche> so looks like we're about even on pull requests opened and close this week - about 20 each
16:42:45 <sa2ajj> you can merge my cmake step :)
16:42:48 <djmitche> mostly bugfixes
16:42:55 <sa2ajj> new feature!
16:43:05 <sa2ajj> i did some cleanup re python 2.6
16:43:15 <sa2ajj> btw, newest twisted does not support 2.6
16:43:43 <djmitche> I think we have a special-case in place in setup.py for the twisted compatibility
16:43:46 <rutsky> sa2ajj: one of the next topics is dropping 2.6 :)
16:43:52 <anish> cleanup that is different from how this is handled along with autobahn ?
16:44:09 <bb-trac> [trac] #3469/defect (closed) updated by tardyp (empty comment) http://trac.buildbot.net/ticket/3469
16:44:25 <djmitche> what's the status on foreign keys?
16:44:34 <djmitche> anish: I think the idea is the saem
16:44:34 <rutsky> buildbot on 2.6 installs latest Twisted release that still support 2.6, same with autobahn (it's dropped 2.6 too)
16:44:35 <djmitche> *same
16:44:45 <rutsky> tardyp: thanks!
16:44:50 <tardyp> I've disabled fk for sqlite as we had strange reports
16:44:59 <tardyp> I dont want to delay nine because of this
16:45:05 <djmitche> anish: https://github.com/buildbot/buildbot/blob/master/master/setup.py#L393
16:45:15 <djmitche> ok
16:45:30 <djmitche> #info foreign keys for sqlite are disabled for now -- strange issues, and not worth delaying nine
16:45:31 <tardyp> I spent a lot of time trying to make it work, but I am not confident that it is really stable or usuable with sqlite
16:45:35 <djmitche> yeah
16:45:36 <djmitche> that's fair
16:45:42 <anish> djmitche: that's my code :)
16:45:50 <djmitche> oh, right heh
16:45:59 <anish> (hence checking)
16:46:02 <djmitche> ah, I misunderstood
16:46:15 <tardyp> ion has been working on enabling innodb for our prod
16:46:25 <djmitche> #info anish and sa2ajj working on python-2.6 compatibility
16:46:30 <tardyp> which for me will be a good test for any fk bugs
16:46:52 <tardyp> innodb has fk support while myisam does not
16:46:56 <djmitche> #info ion is landing support for MySQL innodb tables, at the cost of shrinking some columns from 256 to 255 characters
16:46:58 <djmitche> yep
16:47:06 <rutsky> is innodb has FK with yet another behaviour? :)
16:47:15 <djmitche> of course!
16:47:17 <tardyp> probably, yes
16:47:19 <gracinet> anyone testing with postgresql ?
16:47:28 <djmitche> gracinet: yes, we test in the metabuildbot
16:47:30 <djmitche> against postgres
16:47:32 <gracinet> (hi, people)
16:47:34 <gracinet> great
16:47:36 * djmitche waves
16:47:36 <tardyp> gracinet: at the moment no, but you are welcome to start this
16:47:53 <djmitche> http://buildbot.buildbot.net/builders/db-postgres/builds/420
16:47:54 <gracinet> I'm just starting to experiment with nine actually
16:47:56 <tardyp> pg is supported via unit tests
16:48:06 <tardyp> but we have no known prod users of pg
16:48:07 <gracinet> sounds good enough, I'll report any pbms
16:48:11 <sa2ajj> what tardyp means is that beside testing in meta bb we do not know of many active users of postgresql setup...
16:48:20 <sa2ajj> :)
16:48:22 <gracinet> ok tardyp, count me (anybox) then
16:48:27 <djmitche> there are a few, although I couldn't name them off the top of m head
16:48:30 <djmitche> hehe
16:48:40 <djmitche> ok
16:48:41 <tardyp> djmitche: on nine?
16:48:44 <djmitche> #topic Beta 8 release
16:48:49 <djmitche> tardyp: oh, probably not
16:49:04 <tardyp> nine has much more complex db setup
16:49:08 <djmitche> indeed
16:49:26 <djmitche> #info Beta 8's main focus will be the new worker support, and any additional bugfixes by then
16:49:38 <djmitche> #info Beta 8 will not aim to fix all open 0.9.0 bugs
16:49:39 <tardyp> and also probably the http reporters
16:49:46 <djmitche> #info ..and also HTTP reporters
16:50:06 <rutsky> and maybe drop python 2.6... :)
16:50:12 <sa2ajj> what're the criteria for beta 8?
16:50:25 <djmitche> sa2ajj: basically, bug 2340 for worker complete
16:50:42 <tardyp> which is a big change and a big reason to test it
16:50:45 <sa2ajj> aha. that's good. btw, do we need to keep that branch at github?
16:50:50 <sa2ajj> (it was merged, was it not?)
16:50:54 <tardyp> yes
16:50:57 <djmitche> the bug2340 branch?
16:51:02 <sa2ajj> yes, that one
16:51:09 <djmitche> I think that could be removed - rutsky?
16:51:10 <rutsky> at this moment we can delete it
16:51:14 <djmitche> ok
16:51:21 <bb-trac> [trac] #3371/defect (closed) updated by rutsky (I'll close this, as it seems that this issue is resolved.    @packadal feel free to ...) http://trac.buildbot.net/ticket/3371
16:51:45 <djmitche> #topic reStructuredText formatting tools
16:52:04 <djmitche> sa2ajj has done a lot of work on the formatting of the docs
16:52:08 <bb-github> [13buildbot] 15sa2ajj 04deleted 06bug2340 at 14783f31c: 02https://git.io/vaQ28
16:52:11 <djmitche> but I understand a lot of that was fairly manual?
16:52:15 <djmitche> thanks sa2ajj :)
16:52:20 <sa2ajj> yes, it was manual
16:52:33 <rutsky> I haven't heard of RST formatting/checking tools, do you?
16:52:46 <sa2ajj> as i mentioned in a comment above: since i do a lot of ReST documentation, it's automatic for my fingers.
16:52:57 <djmitche> #info looking for tools to help format, maintain the documentation
16:53:05 <rutsky> sa2ajj: it's hard to review such changes :(
16:53:09 <sa2ajj> and then i have a tendency to review the files i am about to change....
16:53:19 <anish> rutsky: https://pypi.python.org/pypi/restructuredtext_lint ?
16:53:20 <sa2ajj> rutsky: in what way?
16:53:24 <rutsky> if they would be done like "run autopep8" it would be much easier
16:53:31 * sa2ajj agrees
16:53:48 <sa2ajj> ReST is quite special format :/
16:54:02 <djmitche> agreed :/
16:54:20 <sa2ajj> e.g. asciidoc is much more formal (i think it's much more formal than markdown as well)
16:54:22 <anish> that linter does give basic pep8 like functionality though
16:54:31 <rutsky> anish: thanks! This looks interesting
16:54:35 <sa2ajj> i'll look at that
16:54:48 <djmitche> #info anish points to https://pypi.python.org/pypi/restructuredtext_lint
16:54:48 <anish> + in active dev
16:54:51 <sa2ajj> maybe i could change my workflow :)
16:54:56 <djmitche> cool
16:55:27 <djmitche> #topic Droping of Python 2.6 support - http://trac.buildbot.net/ticket/3488
16:55:29 <rutsky> I'll create ticket to try it in our tests workflow
16:55:36 <djmitche> awesome
16:56:05 <sa2ajj> what would be the reasons for dropping 2.6 support?
16:56:13 <bb-trac> [trac] #3495/undecided (v:master) created by rutsky (run linter on RestructuredText) http://trac.buildbot.net/ticket/3495
16:56:15 <sa2ajj> (it would not affect me, but it would our users...)
16:56:34 <anish> twisted + authbahn no longer support it on latest
16:56:35 <rutsky> 2.6 is obsolete and dropped by many libraries
16:56:43 <anish> others probably coming soon
16:56:44 <djmitche> I'm more worried about compatibility on workers - lots of folks ship software that still runs on old enterprise OS's
16:56:54 * anish waves
16:56:56 <djmitche> so maybe just require 2.7 on master
16:57:00 <rutsky> I think we can still support 2.6 on worker
16:57:03 <djmitche> but leave 2.6 support on worker
16:57:04 <sa2ajj> then let's make it explicit: it's about master
16:57:09 <rutsky> and 2.7+ on master
16:57:17 <djmitche> well, 2.7 - no +
16:57:22 <djmitche> yet
16:57:27 * sa2ajj nods :)
16:57:37 <rutsky> >= 2.7 :)
16:57:39 <sa2ajj> bfdls are bfdls
16:57:49 <djmitche> #agreed will drop support for 2.6 on master
16:57:56 <sa2ajj> or were they bdfls?...
16:58:00 <djmitche> #agreed worker & slave will maintain 2.6 support for now
16:58:02 <rutsky> who uses 2.6 on master?
16:58:05 <tardyp> +1
16:58:06 <rutsky> anish: do you?
16:58:07 * anish raises hand
16:58:09 <djmitche> bDFl is correct :)
16:58:10 <skelly> we do
16:58:10 <anish> yeah
16:58:11 <InitHello> who uses 2.6?
16:58:16 <anish> RHEL6
16:58:23 <rutsky> anish: is it will problematic for you to upgrade?
16:58:26 <rutsky> only master
16:58:26 <InitHello> all the cool cats are on 3 already
16:58:32 <InitHello> rutsky: yes, RHEL6
16:58:43 <anish> rutsky: you can't with rhel6 packages
16:58:48 <anish> security is unhappy with epel
16:58:50 * sa2ajj is happily on CentOS7 (with aim at RHEL7)
16:58:54 <InitHello> I have many words for red hat based distros. All of them are obscene.
16:58:55 <anish> that said, master works on 2.6 right now
16:59:06 <anish> it would be nice if 2.6 is dropped when it breaks next
16:59:07 <InitHello> some are also profane
16:59:08 <anish> instead of now
16:59:28 <anish> because it basically means I can't update beta->final
16:59:32 <InitHello> some are profane and profane accessories
16:59:48 <rutsky> so we can leave 2.6 support on 2.6 users, and try to not break it
16:59:51 <anish> I have pocket sand for you InitHello
17:00:07 <InitHello> we're stuck with centos6 on some of our hosts, because reasons
17:00:14 * djmitche right there with you
17:00:22 <djmitche> we install python2.7 in /tools to run buildbot
17:00:27 <djmitche> despite running buildbot 0.8.2 :/
17:00:33 <rutsky> we run BB in Docker :)
17:00:38 <anish> same
17:00:43 <anish> except only slaves in docker
17:00:46 <djmitche> rutsky: if it's not too difficult, it'd be best to keep 2.6 support
17:00:51 <anish> testing is harder in docker
17:00:56 <djmitche> the nine upgrade is going to be "fun" enough without having to do an OS ugprade at the same time
17:01:08 * sa2ajj nods
17:01:10 <cmouse> anish: it is?
17:01:14 <cmouse> anish: i've found it easier
17:01:20 <sa2ajj> fun it is not
17:01:53 <anish> cmouse: harder to examine contents after a failed build
17:02:04 <cmouse> anish: use volumes? =)
17:02:12 <anish> cmouse: lotsa bugs in volumes
17:02:14 <anish> fixed them
17:02:21 <cmouse> anish: but yes, that is a difficulty
17:02:22 <anish> but too late
17:02:35 <cmouse> anish: but then again i've found it nice that you can manually repeat stpes
17:02:37 <cmouse> steps
17:02:38 <rutsky> do we have some features of 2.7 that we _really_ want to use?
17:02:47 <djmitche> not that I know of
17:02:47 <cmouse> anish: in the *same* environment
17:02:53 <cmouse> anish: because you get the exact same image
17:03:01 <anish> cmouse: also volumes can't be mounted in buld dir, plus builds are sandboxed. are you doing some magic to get around this
17:03:02 <sa2ajj> i do not think it's about features, it's more about libraries bb depends on...
17:03:08 <anish> (also we're derailing the meeting)
17:03:08 <djmitche> rutsky: we kept 2.5 support for a long time, too, for rhel5 users
17:03:11 <cmouse> anish: ?
17:03:23 <cmouse> anish: you can easily map a volume to a build dir if you know what you are doing
17:03:31 <djmitche> I don't see a big forcing factor right now for dropping 2.6
17:03:40 <rutsky> djmitche: me too
17:03:41 <cmouse> anish: you can even preseed f.ex. git repos there
17:03:44 <anish> cmouse: pass me your config please
17:03:45 <djmitche> keeping 2.6 support isn't blocking anything we must have by 0.9.0
17:03:47 <rutsky> it's not like 2.7 vs 3 :)
17:03:51 <cmouse> anish: i have no magic up my sleeves
17:04:01 <djmitche> so given this feedback, I"m tending to say, let's keep 2.6 support for both master and worker?
17:04:05 <anish> yayy !
17:04:26 <djmitche> tardyp: ok?
17:04:28 <rutsky> lets keep 2.6 in "compatibility" mode with support mainly from 2.6 users.
17:04:39 <djmitche> that might be too nuanced
17:04:41 <cmouse> anish: you can pass volume mount info in docker latent worker
17:04:41 <sa2ajj> djmitche: i'd like to continue fixing those "legacy" things that are not relevant for 2.6 though :)
17:04:49 <rutsky> anish very quickly solved issue with autobahn - thanks!
17:04:54 <djmitche> I think what we have is OK - tested on 2.6 so we find out if we break compatibility
17:04:54 <cmouse> anish: and you can premake the build dir in your image
17:04:59 <djmitche> sa2ajj: that's fine :)
17:05:08 * sa2ajj is relieved... :P
17:05:13 <cmouse> anish: in fact, i put the git repos i need in the build image during image build
17:05:17 <djmitche> #agreed (after more discussion) to not change compatibility at this point
17:05:22 <cmouse> anish: 'cos it takes up to a minute to clone all them fresh
17:05:28 <cmouse> anish: and few seconds to do git pull
17:05:37 <anish> cmouse: I think I never looked up the option to not clobber build dir at the start of every build. that is proly what needs changing
17:05:48 <anish> clobber + mount = failure
17:06:02 <cmouse> well, yes.
17:06:03 <bb-trac> [trac] #3488/enhancement (closed) updated by dustin (A lot of users are still running RHEL6 / CentOS 6, where running Python-2.7 is ...) http://trac.buildbot.net/ticket/3488
17:06:04 <cmouse> don't clobber?
17:06:06 <djmitche> ok
17:06:14 <cmouse> anish: it's bit pointless to clobber on docker image =)
17:06:22 <anish> yeah
17:06:52 <djmitche> other topics? it looks like the last two (http://trac.buildbot.net/ticket/3469 and foreign keys) have been covered
17:07:08 <rutsky> yes
17:07:15 <rutsky> take a look at agenda please
17:07:28 * sa2ajj would suggest to keep AOB at the end of the agenda...
17:07:38 <gracinet> well, what can I do to help ? Have some time to do buildbot this week (not much afterwards)
17:08:07 <rutsky> djmitche, tardyp: what is this https://github.com/buildbot/buildbot/pull/2060?
17:08:24 <rutsky> sa2ajj: what do you mean?
17:08:34 <djmitche> rutsky: I suspect someone saw "revert-me" in a commit message and thought, "hey, I can do that"
17:08:37 <rutsky> it's already at the end...
17:08:46 <djmitche> #topic Foreign Keys
17:08:51 <rutsky> hm...
17:08:54 <djmitche> https://github.com/buildbot/buildbot/pull/2060
17:09:19 <rutsky> lets wait for response from @Rjaylyn
17:09:34 <djmitche> ok
17:09:38 <rutsky> another topic is PDR: http://trac.buildbot.net/ticket/3493
17:09:43 <rutsky> *PBR
17:09:44 <djmitche> #topic Using PBR
17:09:59 <djmitche> http://trac.buildbot.net/ticket/3493
17:10:02 <sa2ajj> rutsky: the way how agenda is structured, it's not clear what is AOB and what is not :/
17:10:06 <rutsky> ping @stibbons
17:10:13 <djmitche> I've played with PBR a little for other projects, but it suffers from two problems with OpenStack software:
17:10:27 <djmitche> very much tied to OpenStack, and rather inelegantly and poorly implemented
17:11:27 <rutsky> sa2ajj: ok :)
17:11:35 <djmitche> rutsky: I think we could do a lot to improve setup.py, but that is the sort of change that creates unexpected issues with releases, packaging, etc.
17:11:41 <skelly> pbr is a four^Wthree-letter word here
17:11:42 <djmitche> I think at this point in the beta process, that would be a bad idea
17:11:48 <djmitche> haha
17:11:54 <djmitche> pbr is a cheap beer here ;)
17:11:59 * sa2ajj proposes to clean up setup before 0.9 *release*
17:12:21 <djmitche> no
17:12:29 <rutsky> IMO moving to pdr is not "cleanup", but more like refactoring to use another tool/library
17:12:30 <djmitche> if anything *can* be done after the 0.9.0 release, it shoudl be
17:12:46 <sa2ajj> on a related note, i do not really like `pip install` in `master/docs/Makefile` :/
17:12:55 <djmitche> we need to ship 0.9.0, preferably in mid-2014
17:13:01 <djmitche> sa2ajj: I think that got fixed
17:13:08 <sa2ajj> nope
17:13:08 <rutsky> it may be even impossible to use PBR in our workflow
17:13:13 <rutsky> 2014?
17:13:20 <djmitche> as in, we're two years late already
17:13:26 <rutsky> is 2014 is PR or ticket number? :)
17:13:28 <djmitche> so loading more work, bugs, and time into 0.9.0 is a really abd idea
17:13:31 <djmitche> it's a year
17:13:40 <djmitche> *bad idea
17:13:48 <sa2ajj> re setup.py and 0.9 => maybe it's too strong; not during betas for sure :/
17:14:11 <rutsky> djmitche: what we will gain from 0.9.0?
17:14:12 <djmitche> I'm happy to take more bugfixes, and small features (new step is fine)
17:14:19 <rutsky> more users will come from 0.8?
17:14:30 <djmitche> yes
17:14:47 <djmitche> and users will start taking us seriously
17:14:54 <djmitche> so far it's been 3-4 years of vaporware
17:14:57 <sa2ajj> more users -- it's not clear since there'll be breaking changes
17:15:12 <djmitche> yes, but right now we say "yes, eight is broken, wait for nine" which isn't really practical
17:15:13 <gracinet> definitely, for instance I may have an install at a customer's of mine, and it'd be hard to argue for beta
17:15:14 <sa2ajj> and some projects do not use latest 0.8.x...
17:15:26 <rutsky> 0.9.0 should be usable also
17:15:58 <rutsky> when I started hacking 0.9.0 in December 2015 it took me few days just to get pyflakes demo working
17:16:05 <rutsky> *master branch
17:16:07 <djmitche> yes
17:16:16 <djmitche> and I think once we release 0.9.0, we'll find a slew of issues and cleanup to do
17:16:34 <djmitche> and we need to focus on that
17:16:47 <sa2ajj> i think that's worth writing down: 0.9 lots of breaking stuff
17:16:53 <rutsky> oh, another topic: how about some kind of BB bug-day or install-day?
17:16:58 <sa2ajj> 0.9+ fixing what's found
17:17:27 <bdbaddog> Can I sugguest the next version number be 1.0 (after 0.9.0 is released and stable)?
17:17:49 <bdbaddog> selling some customers on anything versioned < 1.0 can be tough to impossible (for ingorance reasons)
17:18:02 <rutsky> announce meeting dedicated to installation of latest nine release and try to run/implement some common tasks with it
17:18:06 <djmitche> bdbaddog: the longer-term plan there is to iron out wrinkles so that we can commit to semantic versioning in 1.0.0
17:18:18 <djmitche> so 0.9.x won't last the >10yr that 0.8.x did
17:18:26 <bdbaddog> :)
17:18:39 <djmitche> rutsky: I like that
17:18:44 <bdbaddog> I had a similar discussion with SCons many years ago. we're now on 2.4.2.. ;)
17:19:07 <rutsky> does anyone have experience with organization/particioation in such meetings?
17:19:32 <djmitche> not me
17:19:34 <bdbaddog> Maybe screencast would be useful? (for upgrade)
17:20:05 <djmitche> gracinet: that may be a good use of your time?
17:20:16 <djmitche> gracinet: especially if you can save some time to fix issues you find
17:20:39 <rutsky> afk
17:20:44 <gracinet> djmitche: sorry, what would be a good use ?
17:20:56 <djmitche> doing an install of nine and finding some bugs
17:21:06 <gracinet> that's what I'm doing right now :-)
17:21:07 <djmitche> I mean, finding out how easy it is and that there are no bugs
17:21:10 <djmitche> ok!
17:21:17 <gracinet> but it's hard to tell apart my bugs and buildbot's
17:21:22 <djmitche> yeah :/
17:21:47 <djmitche> so to sum up
17:21:55 <bdbaddog> once we're ready for 0.9.0 release I'll give a wack at some sort of screen cast and share with the group.
17:22:00 <djmitche> we should be focused on (a) fixing already-filed bugs for 0.9.0
17:22:02 <gracinet> I can try and make a report on the ML if you like
17:22:12 <djmitche> (b) finding (and fixing) new bugs in 0.9.0
17:22:23 <djmitche> (c) bug 2340
17:22:32 <djmitche> (d) continuing to merge external contributions and fixes as they arrive
17:23:06 <djmitche> a big advantage of shipping 0.9.0 is that *new* users will be using it, so we'll get to fix a lot of things for them that are not involved with backward-compatibility
17:23:23 <djmitche> if we find bugs in 0.9.0 that block upgrades from 0.8.x, those users can stay on 0.8.x until we fix them
17:23:59 <djmitche> does that all make sense?
17:24:10 <djmitche> I know I can't tell y'all what to work, but I can exhort :)
17:25:05 <djmitche> >>crickets<<
17:25:22 <rutsky> djmitche: there is one drawback in 0.9.0 release
17:25:47 <rutsky> djmitche: afaik you plan to step-down as Buildbot maintainer after 0.9.0 :(
17:26:04 <djmitche> well, that's not really public information yet
17:26:11 <djmitche> at least it wasn't until now
17:26:30 <bdbaddog> This is not the droid you're looking for...
17:26:32 <rutsky> djmitche: I gathered it from one of the public pull requests of yours
17:26:49 <rutsky> djmitche: sorry if I ruined your plans :(
17:26:53 <djmitche> :)
17:27:06 <djmitche> I don't think that should enter into the decision of when to release
17:27:28 * sa2ajj ponders the idea...
17:27:56 <djmitche> I would like to get Buildbot to a stable state
17:28:07 <djmitche> I think that's good for everyone
17:28:16 <rutsky> agree
17:28:38 <rutsky> lets focus on 0.9.0 release
17:28:47 <djmitche> ok :)
17:28:52 <rutsky> then maybe next beta will be rc1 as tardyp suggested?
17:28:59 <djmitche> hopefully!
17:29:08 <djmitche> sa2ajj: does this sound OK to you?
17:29:19 * sa2ajj nods
17:29:22 <djmitche> ok
17:29:52 <djmitche> #agreed (through roundabout discussion) to focus development work on 0.9.0 release as much as possible
17:29:57 <djmitche> #endmeeting