16:29:58 <djmitche> #startmeeting weekly 16:29:58 <bb-supy> Meeting started Tue Sep 27 16:29:58 2016 UTC and is due to finish in 60 minutes. The chair is djmitche. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:29:58 <bb-supy> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:29:58 <bb-supy> The meeting name has been set to 'weekly' 16:30:18 * sa2ajj waves 16:30:18 <djmitche> #topic introduction 16:30:20 <djmitche> https://titanpad.com/ep/pad/view/buildbot-agenda/v3hIm0CMOM 16:30:26 * djmitche waves to sa2ajj -- who else is here? 16:30:38 <tomprince> -ish 16:30:51 * tardyp waves as well 16:31:21 <djmitche> hi all -- please tack items into AOB if you wish to discuss :) 16:31:31 <djmitche> #topic Development week in review 16:31:48 <bdbaddog> Here.. 16:31:58 <bdbaddog> have client call at 12:45, but it'll be short. 16:32:19 <djmitche> hm, I can't load trac.. 16:32:38 <djmitche> #info Bug 3612 - failing to send usage reports makes the master fail to start 16:33:11 <djmitche> that sounds like something straightforward to fix, since it was intended to not cause failure anyway? 16:33:26 <skelly> (I can load trac but it's very slow) 16:33:45 <djmitche> I can't even SSH 16:34:07 <skelly> same 16:34:09 <tardyp> trac was working 5 hours ago 16:34:09 <tardyp> re bug 3612: I missed that 16:34:09 <tardyp> my master managed to start when events.buildbot.net was down last weel 16:34:33 * sa2ajj . o O (loading...) 16:34:38 <tardyp> but there was a traceback in the log, so maybe was confusing 16:34:43 <djmitche> well we can address teh downtime in a bit 16:35:25 <djmitche> tardyp: other highlights of the week? 16:36:28 <tardyp> I have fixed 3614, and another old one 16:36:45 <tardyp> but last week, I did n't manage to do much 16:36:46 <djmitche> awesome :) 16:37:12 <sa2ajj> for some reason -- no luck :/ 16:37:21 <verm__> it's up now 16:37:22 <verm__> it's their networking 16:37:25 <tardyp> I'll look at the other bugs, and probably release rc4 this week 16:37:37 <djmitche> #info bug 3614 (responsive design not working) is fixed now 16:37:37 <djmitche> cool! 16:37:40 <verm__> look at all the timeouts here 16:37:44 <djmitche> verm__: thanks for looking! their = OSUOSL 16:37:57 <verm__> i think it's not the OSL some kind of big network flap happened 16:38:11 * sa2ajj does not seem to be able to change the agenda :/ 16:38:16 <skelly> no longer just Krebs now? 16:38:32 <djmitche> sa2ajj: might the same effect 16:38:52 <djmitche> #info skelly also did some updates to the infra machines 16:39:00 <djmitche> ok, let's talk releases 16:39:05 <skelly> service3 is pretty much done 16:39:08 <djmitche> #topic Releases 16:39:11 <sa2ajj> anyway, i wonder if moving trac (tickets, wiki) to github is a feasible option 16:39:13 <skelly> the jails need to be updated still 16:39:14 <djmitche> awesome :) 16:39:18 <djmitche> #undo 16:39:18 <bb-supy> Removing item from minutes: <ircmeeting.items.Topic object at 0x80692b510> 16:39:28 <skelly> I rebuilt world and probably now kernel on vm1 16:39:29 <djmitche> #info service3 is almost done, but jails need to be updated 16:39:40 <sa2ajj> bb-supy: thanks for the informative message 16:39:40 <bb-supy> sa2ajj: Error: "thanks" is not a valid command. 16:39:45 * sa2ajj knows 16:39:46 <djmitche> hahaha 16:39:51 <skelly> so if anyone ( verm__ ??) wants to install and reboot and pray ;) 16:40:15 <djmitche> do you guys want to schedule a time (when the network is back up..) to work on that? 16:40:32 <djmitche> #info vm1 has a new kernel and needs a (careful) reboot 16:40:47 <verm__> skelly: we can do that tonight possibly 16:41:11 <verm__> i will be dismanteling my lab thursday morning for my move i won't have access to any computers for a day or two 16:41:18 <skelly> I haven't done any install* targets yet 16:41:23 <skelly> okay 16:41:39 <djmitche> cool! 16:41:51 <djmitche> well, not cool moving.. that's no fun 16:41:52 <skelly> /done 16:42:08 <verm__> yeah heh it's a lot of work 16:42:15 <djmitche> #agreed amar and skelly will work together to do that reboot 16:42:19 <djmitche> #topic Releases 16:42:46 <djmitche> #info tardyp planning an rc4 this week after a few other bugs are fixed 16:42:52 <djmitche> sa2ajj: updates on 0.8.x? 16:43:03 <sa2ajj> not yet, sorry :( 16:43:50 <djmitche> ok :) 16:44:11 <djmitche> are you hearing demand for those releases? Are folks suffering on 0.8.12? 16:44:31 <djmitche> and is the release issue with 0.8.14 causing pain? 16:44:44 <skelly> haven't seen any questions about it in a few weeks 16:44:48 <skelly> it did pop up a few times 16:45:00 <djmitche> ok 16:45:10 <djmitche> #info no update on 0.8.x yet 16:45:18 <djmitche> #topic Infra downtime 16:46:27 <djmitche> #info it appears today's downtime was due to an OSUOSL-level issue, or perhaps something even bigger 16:46:44 <djmitche> #topic move tickets/wiki to github 16:46:48 <skelly> just got a global that freenode is under attack too 16:46:50 <djmitche> ^^ sa2ajj take it away 16:46:54 <djmitche> ahh 16:47:59 <sa2ajj> i've been going through a major update of 0.8.x local installation 16:48:00 <sa2ajj> there's a number of things that i should submit upstream (missing items for plugin classes) 16:48:00 <sa2ajj> and that goes to master first (as they can :)) 16:49:09 <djmitche> ah, cool 16:49:22 <djmitche> I meant "take it away" regarding tickets/wiki on github :) 16:49:26 <skelly> oh, that reminds me 16:49:26 <tardyp> good 16:49:39 <skelly> there was something that came up about deprecated loggers in twisted 16.2+ 16:50:12 <skelly> don't see anything filed for it 16:50:21 <sa2ajj> 0.8.14 was taken away 16:50:33 <sa2ajj> demand -- yes 16:51:05 <tomprince> skelly: The old logging system likely isn't going away anytime soon. 16:51:16 <tomprince> twisted itself isn't entirely ported over, even. 16:51:18 <djmitche> I feel like I'm in a time-warp :) 16:51:19 <sa2ajj> even thought it was explicitly stated: master first, eight next 16:51:49 * sa2ajj is [slowly] moving internal tools from py2 to py3 16:51:55 <djmitche> sa2ajj: I think it's OK to land fixes on eight 16:53:54 <djmitche> I think maybe we should cut this meeting short due to the freenode issues.. 16:54:04 <skelly> seems to better now 16:54:12 <djmitche> ok 16:54:13 <skelly> haven't seen a disconnect in four minutes 16:54:24 <djmitche> sa2ajj: do you want to talk about github issues/tickets? 16:55:45 <sa2ajj> yes, it is OK, it's just not meeting people's expectations :/ 16:56:01 <sa2ajj> yes, github thingie :) 16:56:54 <tardyp> what is the context? 16:57:30 * djmitche wondering the same 16:57:36 <sa2ajj> we have issue that are closed by PRs 16:57:57 <sa2ajj> and those PRs do not always refer properly to trac tickets 16:58:27 <sa2ajj> if we have issues on github, then we have good linkage 16:58:33 <sa2ajj> (that's first thing) 16:59:00 <sa2ajj> 2nd. we have lots of spam fighting engines in place for wiki updates 16:59:12 <tardyp> well, its the same, we have linkage only if people are actually refering to the issue they are fixing 16:59:53 <sa2ajj> true, however, not everybody puts a link to the fixed trac issue 17:00:10 <sa2ajj> they quite often put "Fixes #trac-ticket", which is not helpful 17:00:40 <tardyp> even my self, I put fixes when I can, but the ticket dont get closed 17:01:05 <skelly> it worked for a little while but broke again 17:01:06 <tardyp> which is not that annoying for me . ticket closing is more like nice to have 17:01:08 <djmitche> #info pro: github PRs and trac tickets do not interlink well; using github only would take advantage of github's integration 17:01:35 <djmitche> #info pro: github takes care of fighting spam so we don't have to 17:02:01 <djmitche> I'll add, just about everyone has a github account, and many are reluctant to create a new account just to file or comment on a bug 17:02:02 * sa2ajj nods 17:02:21 <djmitche> #info pro: everyone has a github account, whereas many do not want to sign up for Trac for a small issue 17:02:32 <skelly> not using trac means a bit less effort needed for infra 17:02:35 <sa2ajj> our contributors have to have a github account 17:02:37 <tardyp> especially that the spam filter often prevent them to create accoun 17:02:50 <sa2ajj> and that as well 17:02:57 <djmitche> #info (regarding spam, this also prevents people from creating accounts quite frequently) 17:03:08 <djmitche> #info pro: less infra requirements if we're not running our own Trac 17:03:14 <djmitche> so what are the cons? 17:03:14 <tardyp> we have lots of wiki pages that depend on trac's integration between wiki and bugs 17:03:34 <sa2ajj> the cons are: different markup language 17:03:35 <tardyp> we have to figure out if it is possible to achieve the same 17:03:42 <djmitche> tardyp: I don't get the feeling many people look at those pages 17:03:49 <sa2ajj> the issues are not on github 17:04:01 <sa2ajj> (this can be fixed *easily*) 17:04:06 <tardyp> I can only think of gsoc page, indeed 17:04:11 <djmitche> #info con: different wiki markup language, so porting wiki pages would be tricky 17:04:25 <djmitche> sa2ajj: what could be fixed easily? 17:04:29 <tardyp> there is a wiki translation engine I think 17:04:32 <djmitche> just enabling issues on the repo? 17:04:35 <djmitche> oh cool 17:04:46 <sa2ajj> and we'd force our users to get github account (those who just report issues) 17:05:20 <sa2ajj> not just enabling issues, but transferring the existing one 17:05:22 <bdbaddog1> pandoc is your friend here. 17:05:25 <__aalvz> Hi guys. i Don't get how to pass properties via "sendchange" I'm trying "buildbot sendchange .... --property=test:something" and in the steps I have "steps.ShellCommand(command=['echo', 'buildername:',util.Property('test')]" 17:05:34 <tardyp> bdbaddog1: yes, that's pandoc 17:05:39 <__aalvz> Is there something I'm doing wrong? =/ 17:05:48 <djmitche> #info (although pandoc would help there) 17:05:55 <djmitche> sa2ajj: so, how would we transfer existing issues? 17:06:01 <tardyp> __aalvz: hi. we are in the middle of a meeting. please ask your question after 17:06:13 <__aalvz> sorry! thanks 17:06:15 <sa2ajj> i *think* there are tools. i can look into that. 17:06:40 <djmitche> sa2ajj: that's my major concern is that the history currently in trac be preserved and findable somehow 17:06:40 <tomprince> Moving issues would keep the existing issue numbers. Probably not a big deal, although any references in the code would probably want to be updated. 17:06:43 <sa2ajj> __aalvz: we are about to finish :) 17:06:52 <djmitche> tomprince: we have PR's with the same #'s as bugs 17:07:01 <djmitche> and github shares the namespace for PRs and issues 17:07:20 <tomprince> Gah. I meant wouldn't. 17:07:23 <bdbaddog1> https://github.com/mavam/trac-hub 17:07:29 <djmitche> ah, ok 17:07:46 <bdbaddog1> looks like trac -> github is a fairly common path. 17:07:54 <skelly> yeah, multiple tools 17:07:56 <bdbaddog1> http://stackoverflow.com/questions/6671584/how-to-export-trac-to-github-issues 17:07:59 <djmitche> I think I'd be satisfied if we had a bunch of 302's from trac to buildbot, or if trac was "frozen" with pointers to github issues where necessary 17:08:21 <bdbaddog1> u still need to maintain trac then though.. 17:08:27 <djmitche> verm__: ^^ what do you think? 17:08:41 <djmitche> bdbaddog1: yeah, although maybe it could be snapshotted with wget or something and just served as static content? 17:08:48 <tardyp> indeed, the namespace will change, so we need to update the doc 17:08:49 <skelly> it could be put in read-only first 17:09:01 <tardyp> we have plenty of links to trac in the doc 17:09:16 <djmitche> yes :) 17:10:09 <djmitche> so I'm hearing a lot in favor of this 17:10:16 <djmitche> I'm kind of surprised :) 17:10:41 <bdbaddog1> seems like it would unload the need for a bunch of infra support and you wouldn't lose much (if anything) in the process. 17:10:43 <skelly> but someone still needs to actually do all of this 17:10:44 <djmitche> proposal: those in favor here draft a plan for discussion next week 17:10:46 <djmitche> yeah 17:11:38 <djmitche> I figure drafting a plan is a good measurement of the desire to actually act :) 17:11:47 <bdbaddog1> sure.. 17:12:26 <tardyp> +1 17:12:36 <djmitche> ok 17:12:38 <skelly> freenode failing so seems like a good time to wrap it up 17:12:56 <djmitche> #agreed interested parties will draft a plan, circulate on mailing list, and discuss in next week's meeting 17:12:57 <djmitche> indeed 17:13:03 <tardyp> lets close the meeting and wait freenode is back online... 17:13:03 <djmitche> also, taskcluster's broken so I should go 17:13:06 <djmitche> :) 17:13:12 <djmitche> #endmeeting