Date is 2015-05-01, times are UTC+10.

--- Day changed Fri May 01 2015
06:55 < whot> good morning gents
06:56 < jcristau> good evening whot 
06:56 < keithp> four minutes and counting?
06:56 < whot> enough time for me to wake up, I hope
06:57 < keithp> good luck. I had teleconferences starting at 07:00 here, and I'm still not awake at 14:00
06:59 <+egbert> hi guys!
07:00 < agd5f> hi
07:00 <+egbert> keithp: stop bragging. i started to work at 6:15
07:00 -!- alanc [~alanc@inet-hqmc05-o.oracle.com] has quit [Remote host closed the connection]
07:00 < keithp> egbert: I'm not bragging, I'm whinging
07:00 < keithp> completely different
07:01 <+egbert> keithp: ;p
07:01 -!- alanc [~alanc@inet-hqmc05-o.oracle.com] has joined #xf-bod
07:01 -!- mode/#xf-bod [+v alanc] by ChanServ
07:02 -!- marcoz [~oftc-webi@c-67-190-28-99.hsd1.co.comcast.net] has joined #xf-bod
07:02 < marcoz> hi all
07:06 < whot> mupuf: ping
07:06 < marcoz> whot: pang
07:07 -!- robclark [~robclark@nat-pool-bos-t.redhat.com] has joined #xf-bod
07:07 < whot> ok, that's all of us now, I think. danvet mentioned he won't make it this time iirc
07:07 <+egbert> whot: tomorrow is may 1st. in europe this is a holiday (as you may recall ;) ) so some pepole may not be there.
07:08 < whot> agenda: GSoC update, taxes, expo.x.org, SPI
07:08 < whot> yeah, I think it used to be a holiday here until some gov moved it to november (iirc)
07:09 < whot> stuart isn't here, so I'll skip the taxes bit and mail him directly
07:09 <+egbert> so, you want to talk about expo?
07:09 <+egbert> i did copy all the data that i believe is relevant off the machine to make sure.
07:10 <+egbert> it is sitting somewhere here on my disk. i still have to get the usb stick.
07:10 <+egbert> point is: since keithp is here: is there a way to get physical access to the machine?
07:10 <+alanc> jg has a big archive dir on expo of old X consortium docs - it's where I got the historical xc source history from - you want to preserve that if you can
07:11 < whot> i've got sudo access to the box now, so I can copy everything too if you give me a quick list of what you pulled down
07:11 <+egbert> alanc: this is not what i mean, we have more current data on expo that we do want to archive 
07:11 < whot> otherwise it'll try to rsync everything-ish
07:11 <+alanc> okay
07:12 <+egbert> whot: it is pretty much everything but the os. (and i think i didn't copy the home dirs and email spool due to privacy concerns)
07:12 < whot> i'll leave the email alone, but knowing jg he'll have some goodies in his hom :)
07:13 <+egbert> keithp: do you know anyone at mit still we could contact about this machine?
07:13 < whot> in terms of hosting, members and bod run off expo, where do we migrate these to?
07:13 <+egbert> his physical home i believe, not his home on expo.
07:13 < keithp> egbert: I've never known who at MIT manages these machines, I'm afraid.
07:14 <+egbert> keithp: any clue who could help us there? jg?
07:14 < keithp> I thought someone like ajax had visited the hardware at some point
07:14 <+egbert> (i can try and contact jg over linkedin)
07:14 <+egbert> yeah, but his contacts may have dried up.
07:14 <+egbert> by now.
07:15 <+egbert> you need someone connected like jg ;p
07:16 < whot> tbh, phys access is the least of my worries. if we wipe it and leave it switched off everything else is icing
07:17 <+egbert> whot: point is: if the machine goes down we may need access to the management interface.
07:17 <+egbert> someone seemed to have pulled the cable.
07:17 <+egbert> it is not accessible any more.
07:17 < robclark> not sure if it helps at all but imirkin is at least an mit alum.. he might know some people there still..  although I guess it would help if someone knew what dept and/or building it was in (since mit is not a small place)..
07:18 < whot> egbert: wfm
07:18 <+egbert> robclark: maybe ajax would remember that.
07:18 < robclark> yeah, would be worth asking.. although he left already (otherwise would be sitting next to me)
07:19 <+egbert> it is fine if you ask tomorrow or next week ;)
07:19 <+egbert> if he doesn't i could ask jg.
07:20 < robclark> k.. I should just ping him on irc..
07:21 < whot> so again: where do we migrate to? because we may get away without physical access
07:22 < whot> and who wants to do the migration (hint: not me, too many deadlines atm)
07:22 <+egbert> it would be nice we we would have our own machine. some small box (1u) we could rack somewhere.
07:24 < robclark> I thought cloud stuff was all the rage :-P
07:24 < robclark> would at least avoid the problem of loosing the machine :-P
07:24 < whot> hehe
07:24  * keithp imagines small rack mount machines on the rampage
07:25 <+egbert> maybe that would be a solution as well. we just need to remember to pay the bills in time ;p
07:25 < robclark> at any rate, solutions that don't involve me pretending (badly) to be a sysadmin seem like a good thing.. :-P
07:26 < whot> ok, who wants to take that on as homework?
07:26 <+egbert> whot: what exactly?
07:27 <+egbert> finding a location or be the sysadmin?
07:27 < whot> finding a suitable migration strategy
07:27 < whot> once we know where to put the lot we can talk about how much work and what kind of work it'll be
07:28 <+egbert> can we find out what mit thinks about the machine first?
07:28 < robclark> I guess first question is, whether there is someone who can gift us hosting?  Otherwise it is question of how much it costs for someone to host our server + hw, vs cost for some virtual machine off in the cloud..
07:29 <+egbert> they haven't pulled the plug so maybe they are even eager to keep it just don't know wh to talk to...
07:29 < whot> egbert: sounds like you want to take this over? hint, hint, nudge, nudge
07:29 <+egbert> whot: i can give it a try.
07:30 < whot> aaaand sold!
07:30 < robclark> (fyi egbert, I've already pinged ajax.. I'll let you know if he knows anything about the mit situation)
07:30 <+egbert> robclark: thanks!
07:30 < whot> ok, let's move on: GSoC update - agd5f, mupuf - any comments?
07:32 < agd5f> whot, 6 accepted projects
07:33 < agd5f> several of the non-accepted projects have asked about EVoC
07:34 < agd5f> I'm assuming at this point I can discuss what projects have been accepted
07:34 < whot> students have been notified, right?
07:34 < agd5f> yeah
07:34 < whot> so I think it's public then
07:37 < agd5f> in no particular order:
07:37 < agd5f> 1. Convert the BOCHS and CIRRUS drivers to atomic mode-setting 
07:37 < agd5f> 2. Atomic counters support for llvmpipe including computer shader interactions
07:37 < agd5f> 3. Range analysis pass, and porting optimization passes to Mesa's NIR 
07:37 < agd5f> 4. Apitrace: Add support for performance counters in the profiling view
07:37 < agd5f> 5. Freedreno support on Android with DRM/KMS based Hardware Composer
07:37 < agd5f> 6. OpenCL image support for the r600 gallium driver
07:38 < whot> cool, thanks
07:38 < whot> any other comments regarding GSoC?
07:39 < agd5f> nope
07:40 < whot> alright, thanks. let's move to the last item: SPI
07:40 < whot> what are we going to do re re-voting?
07:41 <+egbert> the idea was to get the bylaws into a better shape first i think.
07:41 <+egbert> mupuf isn't around. so we don't know what his plans there are.
07:42 < keithp> egbert: what else do we want to change? Seems like just putting it up for a vote again and prodding people until they actually do vote might work?
07:43 <+egbert> keithp: there were a bunch of thing that were left for another vote as we didn't have enough time.
07:44 <+egbert> if we do a revote now it would look really dump if we did another one in a year to fix those things.
07:44 < keithp> meh?
07:44 < keithp> getting merged with SPI seems important enough to make that not pend yet another long round of editing to me
07:46 < robclark> not sure if it gets awkward for GSoC if who the money goes to changes half way through.. either way sooner rather than later seems like a good idea..
07:47 < marcoz> agreed, sooner rather than later,   if we look dumb in the future, so be it
07:48 < keithp> so, how about just planning on running exactly the same ballot item in a couple of weeks and letting people know well ahead of time that not voting is the same as voting no
07:49 < robclark> (and reminder spam perhaps, with one week or less remaining in voting period..)
07:49 < keithp> hourly spam until enough people vote?
07:49 < robclark> not like anyone procrastinates or anything.. 
07:49 < robclark> heheh
07:50 < robclark> wouldn't hurt to point out # of people who haven't voted yet, or something like that..  anyways.. one way or another more proactive reminders is probably a good thing..
07:50 < whot> right, so the next question is obvious then: who wants to run the vote?
07:50 < keithp> who knows how?
07:51 <+egbert> i know how to run the voting system
07:53 <+egbert> guys, i would like to call it a day pretty soon. the long delays may be acceptable to you but not here.
07:53 < keithp> could you set up a ballot and let me drive the email bits?
07:53 <+egbert> i could do that.
07:54 < keithp> ok, lets vote here on what we should do, and then move on from there.
07:54 < keithp> Who thinks we should re-run the bylaws issue in another election in the next few weeks?
07:54 < keithp> +1
07:54 < robclark> +1 for egbert setting up ballot and keithp playing spam-master
07:55 < agd5f> +1
07:55 < whot> +1
07:55 <+egbert> i would rather get the bylaws fixed.
07:55 <+egbert> first
07:55 < robclark> is there a list of things needing to be fixed?
07:55 <+egbert> mupuf knows best
07:56 < robclark> hmm
07:56 < marcoz> +1
07:56 < robclark> any chance we could take this to bod mailing list to avoid waiting another two weeks (to see if there is something that could be fixed in short order about bylaws)?
07:57 < keithp> THere's absolutely nothing saying we can't work on fixing the bylaws in parallel and calling for another vote in a few months
07:57 < robclark> +1
07:57 <+egbert> right
07:57 <+egbert> but we need to do it and stay with it.
07:57 <+egbert> not drop it again until a few days before the election.
07:57 < keithp> yeah, that didn't work out well
07:58 < robclark> mupuf, maybe when you are around, you could send todo list to mailing list?
07:59 < whot> ok, then let's do this: mupuf will send the todo list to the mailing list, we figure out what to amend and what is appropriate in the scope. once we agree on that, mupuf can push the changes to the git repo, once that's done egbert can set up the ballot
07:59 < robclark> sounds good
07:59 <+egbert> +1
07:59 < whot> and theory goes this should even be possible so that by the next meeting we have a ballot date and keithp can start spamming around
08:00 < keithp> whot: so you'd like to change your vote to pend for a few weeks until we figure out how hard that will be?
08:00 < agd5f> +1
08:00 < whot> keithp: I'd like _us_ at least to be in agreement that the bylaws are good enough, right now it doesn't seem to be the case
08:01 < keithp> that seems completely reasonable to me, and limits the scope to a two week timeframe, which is also reasonable
08:02 < keithp> I will also change my vote then, and we'll revisit this in  two weeks. +1
08:02 < keithp> Are we done for the day? It's been an hour now.
08:02 < whot> I think so, any other urgent items?
08:02 <+egbert> yes, please 
08:02 < keithp> egbert: thanks for staying up late, and also thanks for reminding us that we weren't making progress
08:02 <+egbert> it is past midnight here. i would like to leave
08:03 < keithp> see y'all in two weeks time
08:03 <+egbert> keithp: :)
08:03 <+egbert> see ya!
08:03 < whot> yeah, let's call it a day. see you guys in two weeks