From Fedora Project Wiki

Fedora Release Engineering Meeting :: Monday 2008-06-23

Discussion

  • See IRC log

IRC Log

< f13> ping: notting jeremy spot wwoods warren jwb rdieter poelcat lmacken
* jeremy is around
< rdieter> yo
* warren exists.
* notting is here
< f13> spot is on a beach somewhere
* wwoods here
* poelcat here
< f13> hrm, I wonder if jwb is still dealing with relocation stuff
< jeremy> I believe he said he would be
< f13> k
< f13> well I was going to go over open tickets, but his is the first one, spins for F9
* lmacken is around
* nirik would like to see those done... anyway to let someone else make them?
< f13> nirik: I think we were actually held up on the spins SIG on how they were going to be produced, where stored, sources, etc...
< nirik> ah, ok. fair enough.
< f13> jwb went to their meeting, where this was going to be one of the topics.
< f13> Hopefully we can touch base with jwb before next week's releng meeting
< f13> .rel 23
< zodbot> f13: #23 (Fedora 10 Naming) - Fedora Release Engineering - Trac - https://fedorahosted.org/projects/rel-eng/ticket/23
< f13> Josh got the list back from the board and I've forwarded those on to RH Legal.
< f13> No response yet from legal.
< notting> hm. what happens if legal comes back with one or two? do we have the full results?
< f13> notting: I'd hope that the full results of names are kept by the board, so that the board can dole out a few more names to legal clear
< jeremy> they're in the voting app db, so should be able to be grabbed
< f13> nod
< f13> .rel 105
< zodbot> f13: #105 (Vette F10 schedule against community events, holidays, and upstream release schedules) - Fedora Release Engineering - Trac - https://fedorahosted.org/projects/rel-eng/ticket/105
< f13> sadly, there has been no input here. And we're over 1 month past the F9 release, we /really/ need an F10 schedule set.
< notting> then it's fine as is!
< f13> alright, mark as such and close the ticket (:
< notting> done
< f13> The rest of the tickets are all Alpha/Beta+ related.
-!- f13 changed the topic of #fedora-meeting to: Fedora releng - Future!
< f13> poelcat had posted a 'What is rawhide' mail a while back, and a number of us spent time at FUDCon going over that
* poelcat needs to get the notes out :)
< jeremy> can we do something before that
< jeremy> do we have plans to have a box in phx for doing composes for f10?
< f13> what came out of that time was a more clear idea of what we use rawhide for, and identifying some tools needed and some enhancements needed to our current tools
< f13> jeremy: I'm going to ask mmcgrath for one of the blades of each arch for that.
-!- Irssi: #fedora-meeting: Total of 106 nicks [1 ops, 0 halfops, 0 voices, 105 normal]
< f13> jeremy: we have releng2, which is going to be RHEL5 based, for doing other releng tasks, but since it's RHEL5, I don't think we'll be able to compose on it
< jeremy> f13: *nod*
< f13> at least last time I tried to use RHEL5 with selinux permissive to do an F9 pungi compose things went *boom*
< f13> and well livecds.
< f13> with Mike I'll work on a kickstart set we can use to provision the blade (or a guest on the blade) to do our composes, then return the blade back into the builder pool
< jeremy> might be worth getting that infrastructure ticket filed and moving
< f13> yep, I'll do that right after the meeting
< jeremy> sounds good
< f13> so back to the 'what is rawhide' thing, some of the things will be improvements to pungi, and to the scripts that generate rawhide each night
< f13> those changes will probably wait for the releng2 cut over, which right now is hinging on luke's bodhi rollout
< notting> he said that's this week. we have witnesses.
< poelcat> f13: skvidal wrote http://skvidal.fedorapeople.org/misc/verifytree.py
< lmacken> which I am working on now
< lmacken> I'm working on making our bodhi app instances talk to an instance on releng2, which will do the mashes, and scheduled jobs
< f13> poelcat: right, seth already cranked out one tool we need
< wwoods> I actually had a pretty good idea about how to test the simple booting case
< wwoods> working on it now
< wwoods> it's a subset of the problem I was working on already, so that helps
< f13> cool
< f13> As for pungi, I'm deep in API breaking land, leading up to a 2.0 release
< f13> earlier I split the notion of what's /in/ the compose vs what's used to make the compose, since buildinstall can now handle multiple yum repos
< f13> today I'm working on the ability to gather debuginfos for the things in the compose. Not totally necessary for Fedora, but useful nonetheless
< f13> I've also alerted the Fedora Unity folks that now would be the time to make any feature requests
< f13> same really goes to all of you. Now is a good time to throw ideas my way
< f13> I don't really want to be working on pungi too much longer, since I have lots of other fun things I could be working on.
< wwoods> (also it's really not a good idea to do active devel on the compose tools later in the cycle)
< f13> oh yeah, pungi also as of today inherited splittree.py and pkgorder
< f13> In other news, we'll be using our proof of concept Testopia instance for F10 testing
< f13> which is pretty damn cool.
< wwoods> yeah we're still figuring out how to work with the UI and best practices and stuff
< f13> I'm going to talk with bpeck and co later today about getting me an xmlrpc script I can attach to the tail end of rawhide creation that will create test runs for each day's rawhide, so that if anybody feels like testing rawhide, they have some structure
< f13> *cough* andy *cough*
< wwoods> I'm cooking up info on Testopia targeted towards various audiences (testers, packagers, spinners/secondary arch folks, etc)
< f13> and finally, for the signing server I've ordered a couple of OpenPGP Smartcards, with the idea that these will provide the gpg key storage for the server
< f13> I've got to play around with them a bunch and ensure that they'll offer the protection we need first.
< notting> does current testopia use a) bz accounts b) fedora accounts c) its own account store?
< f13> c
< poelcat> f13: was there any tie-in with the spacewalk folks on that?
* poelcat sent them your way
< poelcat> that == signing
< f13> poelcat: sortof.
< f13> notting: before leaving proof of concept, it'll either be a or b
-!- f13 changed the topic of #fedora-meeting to: Fedora releng - open floor
< f13> on the releng spam front, the spam-filter plugin for Trac doesn't seem to have made an impact, although I'm not sure we're using it right. I think I just need to keep feeding SA more spam to learn from.
< jeremy> f13: I stopped using spamassassin with evo due to lack of effectiveness
< jwb> i'm here now
* jeremy has more luck with bogofilter these days
< f13> jeremy: got a way to use that on the hosted1 server?
< f13> jwb: anything you can tell us about the spins meeting?
< jwb> sure
< jwb> the spins sig seemed to be very in-tune with the idea of having spins as Features, and follow that process
< jwb> e.g. tracked and composed from Gold bits, etc
< jeremy> f13: I've only used it via the evo plugins adly
< jwb> they are still trying to figure out how to deal with localized spins. not sure where to go there, but one option is to have the main torrent tracker point to another server that is hosting the torrent
< jwb> that might also be an option for other spins that aren't done through the Feature process (for one reason or another)
< jwb> they wanted rel-eng's thoughts on doing something like that
< jwb> for the F-9 spins specifically, i think i have some todos there. it's been 2 weeks, so i need to go back and read logs to remember that
< notting> i thought we were going to do some f9 spins from gold at some point. maybe i'm forgetting
< jwb> notting, i think that was one of my todos, yes
< jwb> there were some questions on how to do "update" spins, but i don't think we came to any sort of resolution there
< f13> k
< jwb> so, i think our biggest "item" is to think about how we want to deal with non spins/torrent.fp.o spins
< jwb> hm... that came out wrong
< jwb> spins not hosted on spins/torrent.fp.o
< f13> well, our biggest 'item' right now is to get the F9 ones knocked out, then we can talk about the future
< jwb> oh, good point
< jwb> i have cable as of 2 hours ago, so starting tomorrow/wed-ish i'll be back in my normal "ALWAYS ON THE INTERNET" status
< jwb> hopefully we can get the f9 spins done soon
< jwb> FEL (revived now), Games, and XFCE
< jeremy> hoorays!
< f13> yay
< f13> Anything else before we wrap it up?
< drago01> jeremy: have you already "solved" this https://bugzilla.redhat.com/show_bug.cgi?id=450852 ? I don't really know how to fix it (well without adding huge non upstream patches)
< buggbot> Bug 450852: low, low, ---, Kristian Høgsberg, ASSIGNED , compiz-gnome requiring compiz-fusion?
< jeremy> drago01: nope. the answer probably needs to involve a cluestick for upstream...
< rdieter> jeremy: +1
< jeremy> that or making it so that our defaults don't use plugins that are in -fusion
< drago01> jeremy: we have a "disable cube option" ... and the only non cube plugin that let you have multiple workspaces in compiz was plane ... which got dropped in favor of wall which is in -fusion
< jeremy> drago01: hence, cluestick...
< drago01> jeremy: ;)
< drago01> jeremy: splitting wall into a subpackage should work but its ...ugly
< drago01> jeremy: are we _that_ space limited? compiz-fusion + compiz-fusion-gnome ~ 2.1 MB on x86_64
< f13> alright looks like we're done with releng stuff.
< f13> Thanks all!

Generated by irclog2html.py 2.6 by Marius Gedminas - find it at mg.pov.lt!