ReleaseEngineering/Meetings/2008-jul-07

From FedoraProject

Jump to: navigation, search

Contents

Fedora Release Engineering Meeting :: Monday 2008-07-07

Ticket Review

Dropping Orphaned Packages

Adjust Alpha Schedule

IRC Transcript

-!- f13 changed the topic of #fedora-meeting to: Fedora Release Engineering - Roll Call 13:01
f13 ping: notting jeremy rdieter wwoods lmacken warren spot jeremy poelcat jwb 13:02
* spot yawns 13:02
* wwoods smashy smashy 13:02
* notting is here 13:02
rdieter yo 13:02
notting at least until i break my kernel 13:02
* poelcat here 13:03
* lmacken is here 13:03
* jeremy is here-ish 13:04
* jwb is sort of here 13:04
* rdieter goes to stand between jeremy's in here-ish-ville and jwb's sort-of-here-land 13:05
jeremy heh 13:05
f13 haha 13:05
f13 ok, lets get started. 13:05
-!- f13 changed the topic of #fedora-meeting to: Fedora releng - F9 tickets 13:05
f13 .rel 24 13:05
zodbot f13: #24 (spins for Fedora 9) - Fedora Release Engineering - Trac - https://fedorahosted.org/projects/rel-eng/ticket/24 13:05
f13 I worked some on this, test3/4 are setup, one i686 one x86_64 13:06
f13 turns out we don't really need i686, the x86_64 can produce i686 images just fine with setarch 13:06
f13 (unless jeremy has a hidden gotcha here) 13:06
jeremy nope 13:06
* jeremy always does the i686 images with setarch 13:06
f13 however I think the spin-kickstarts git repo is dire. 13:06
f13 the games spin isn't even in the F-9 branch, and none of XFCE or FEL actually worked from the F-9 branch 13:07
* jeremy will brb 13:07
f13 FEL fails to get a package (knetworkmanager), and XFCE can't manage to login before it boots people out 13:07
f13 I need to get in touch with mether or kanarip and verify that the configs in the F-9 branch are the ones I'm supposed to be using, and that they're actually really honestly tested 13:08
notting quality 13:08
notting kanarip is theoretically here. kanarip? 13:08
jeremy he's on vacation I think 13:08
rdieter f13: trying to pull in knetworkmanager is definitely wrong, broken. 13:08
f13 rdieter: yeah, I have a feeling that the fixed configs are on the master branch, but I don't want to make that assumption as those could be F10ish configs 13:09
rdieter you're prob right 13:09
f13 so I need to respond to rahul's comment in the ticket and wait for more information. 13:09
f13 this is definitely an area that needs better process next time around 13:10
f13 alright, anything else with this ticket? 13:11
f13 moving on 13:11
-!- f13 changed the topic of #fedora-meeting to: Fedora releng - F10 tickets 13:12
f13 .rel 23 13:12
zodbot f13: #23 (Fedora 10 Naming) - Fedora Release Engineering - Trac - https://fedorahosted.org/projects/rel-eng/ticket/23 13:12
notting dooom 13:12
wwoods blerg. 13:12
f13 there has been some more back/forth between the board and legal trying to get a reasonable amount of names to vote on 13:12
f13 sadly almost all the names that were suggested don't pass legal muster 13:12
f13 but as I understand it, there is still more being vetted by legal right now 13:13
f13 stickster: ping are you here? 13:13
wwoods can we go back to *not* voting on names 13:13
wwoods and pre-approve like the next 8 13:13
* stickster here 13:13
wwoods and never ever deal with this again? 13:13
f13 stickster: is my current status accurate? 13:13
stickster f13's summary is correct 13:13
f13 k 13:13
f13 wwoods: but that would be taking away FREEEEEEEEeeeeeeedooooooooooom! 13:14
notting f13: no, freedoom is still in the repo 13:14
f13 hahah 13:14
* stickster thinking about names being suggested via wiki for F11. 13:14
f13 +e 13:14
f13 stickster: have we yet exhausted our supply of user suggested names to throw at legal? 13:15
notting wwoods: preapproving the next 8 names and keeping links would make a nasty tree 13:15
stickster f13: The latest batch exhausts the list, but Josh and I went through them and combed google for collisions first, this time. 13:15
jeremy and just because they're approved today doesn't mean they would be in six months 13:15
stickster Lesson learned for next time. 13:15
notting jeremy: zod objects to you sullying his reputation in such a way 13:16
stickster The list that most recently went to Legal includes only those names crossing that hurdle. 13:16
f13 alright, what happens if legal comes back and says nyet? 13:16
stickster We have at least 4 names already that do pass. 13:16
f13 ah ok 13:16
stickster We are looking for a roster of 8-10 total. 13:16
f13 (does that include Red Hat Linux?) 13:16
stickster f13: There was a collision with a major North American -- 13:17
stickster oops. 13:17
f13 heh 13:17
stickster f13: I mean, yes, the dream is still alive ;-) 13:17
* f13 would rather it not be that... but oh well. 13:17
quaid <delurk> (btw, jwb++ thanks for all this extra lifting to get the naming ahead this time) </delurk> 13:17
* stickster prepares slapping hand for notting 13:17
f13 alright, so next time around we're going to pre-prune the suggestions for the obvious legal nonos right? 13:18
stickster quaid++, jwb++ 13:18
jwb f13, yes 13:18
jwb stickster, quaid: np 13:18
stickster f13: Right. And I think we're going to proactively encourage way better names. 13:18
f13 k 13:18
f13 worksforme 13:18
stickster "LOVE"? "LOVE"?!? 13:18
f13 anything else on this topic? 13:18
notting i think we might have to veto 'red hat linux' just because of the 'will cause mass confusion' reason 13:19
notting but... heee. 13:19
wwoods I think the Pirate Captain sums up my feelings pretty well: "Yarr. I hate the sea and everything in it." 13:19
f13 ok, moving on. 13:20
-!- f13 changed the topic of #fedora-meeting to: Fedora releng - Alpha 13:20
f13 Alpha is the end of this month, time to get gears forward on what we need to do for it 13:21
f13 with an added wrinkle that both jwb and I will be at OLS for the freeze area. 13:21
f13 crap, why didn't we notice this when we did the schedule :/ 13:21
f13 I wonder if it's not too late to push alpha out a week to avoid OLS 13:21
jeremy likely not 13:22
jwb i say we should 13:22
jwb i have selfish reasons though 13:22
jeremy (not too late -- that is, it is fine to do so) 13:22
f13 ok, I'm going to ask poelcat to adjust out a week then. 13:23
notting so freeze moves from next tuesday? 13:23
f13 that means we'd freeze on the 22nd, the day before OLS starts, but there will be others not at OLS who can handle tagging 13:23
jwb "freeze" 13:23
f13 and rawhide can be used for the testing grounds 13:23
jwb wait, confused 13:23
f13 once I get back we can spin up the release candidates 13:23
poelcat f13: yeah, i've been meeting propose a more detailed schedule too 13:23
f13 oh crap 13:23
f13 that's right 13:23
f13 alpha is non-blocking freeze 13:24
poelcat s/meeting/meaning 13:24
jwb f13, right 13:24
f13 boy, we gave ourselves a ton of time for a non-blocking freeze release. 13:24
notting it's like we're expecting brokenness 13:24
f13 well and making it harder to verify fixes 13:24
* f13 wonders if nonblocking really helped anything here 13:24
notting if we push it out a week, i won't be here for GA 13:25
* poelcat thought we added more freeze time on purpose 13:25
f13 poelcat: we did, I just was thinking of the mechanics of getting the fixed packages into trees for alpha, and getting any testing on those trees 13:25
f13 well, since we'll be doing the composes in PHX we can likely make the trees available somehow 13:25
jwb are we finally doing that? 13:26
f13 notting: may need some mash work to mash out trees from the freeze tag. 13:26
f13 jwb: theoretically 13:26
jwb that would be nice 13:26
f13 the RCs at least. 13:26
notting f13: shouldn't matter, you can point it at whatever tag 13:26
wwoods link to the schedule? 13:27
f13 notting: we could just push alpha forward but not any others 13:27
f13 wwoods: http://poelstra.fedorapeople.org/schedules/f-10/f-10-all-tasks.html 13:27
notting f13: i meant GA of alpha. *shrug* no big deal 13:28
f13 oh GA of alpha yeah.. 13:28
f13 probably not a big deal if you miss that 13:28
poelcat f13: one thing we had on F9 schedule was "drop orphaned packages at EOL of FC6"... did we do that this time when F7 EOLd? 13:28
f13 unfortunately since alpha is non-blocking, it requires the most releng work to make fixes testable for other people. 13:28
f13 poelcat: hrm, I don't think so. 13:28
f13 warren: ping; didn't you deal with that last time, dropping of the old orphaned packages? 13:28
warren I can do it 13:29
f13 warren: create yourself a ticket would you, toss it in the F10 rawhide tracker. 13:29
f13 (so that we remember to make a ticket for it for F11) 13:29
warren where is the tracker? 13:29
warren f13: and what are the upcoming deadlines for this? 13:30
f13 warren: https://fedorahosted.org/rel-eng/milestone/Fedora 10 Rawhide 13:30
f13 ugh, silly trac and spaces 13:30
f13 warren: well, according to poelcat we did it at EOL of FC6 last time, IE 1 month after GA. So we're behind schedule this time 13:30
f13 warren: https://fedorahosted.org/rel-eng/newticket use that, and just select the Fedora 10 Rawhide milestone 13:31
jeremy f13: so we are planning to have our composes being done in phx for the alpha? 13:32
f13 jeremy: if mmcgrath makes good with the resource request, yes I plan on making the RCs in PHX 13:32
warren meaning I should send out a warning to the list NOW? 13:32
f13 jeremy: I may also do composes locally just to verify tooling works and for quicker local testing, but the RC's and GA I plan to compose in PHX 13:32
warren f13: we want to remove all orphans before alpha? 13:33
f13 warren: that would be helpful. I'd target things that have been orhpaned since before F9 GAd. Anything that was orphaned after that is less targetted for removal. 13:33
jeremy f13: it'd be good to be spinning the final releases with the same environment we use for beta, if not alpha. even if it means that we have to work through the issues of lag time for testing 13:33
f13 warren: yes, we'd like to remove orphans before alpha freeze. 13:33
f13 jeremy: sorry, I meant the release candidates of alpha 13:33
f13 jeremy: not release candidates of F10 13:34
jeremy f13: aha, then we're all good then :-) 13:34
f13 violent agreement 13:34
warren https://fedorahosted.org/rel-eng/ticket/251 Euthanize orphans filed 13:34
jeremy fwiw -- I've been building near daily live images and we are currently remaining cd-sized. I expect that to change the day before freeze ;) 13:34
drago01 jeremy: what did you do about the compiz stuff? 13:35
f13 Proposal: push alpha dates back 1 week to make room for OLS, don't adjust any other dates at this time. 13:35
f13 +1 13:35
jeremy f13: +1 13:35
warren +1 13:35
notting wait, where does OLS fall in the current schedule? 13:35
jeremy drago01: for the moment, other stuff has shrunk to allow us to fit again. but I really don't expect that to last 13:35
drago01 jeremy: ok... 13:36
notting does it make sense to even have a 'freeze' if you aren't going to spin trees for testing for part of it? 13:36
warren f13: should I announce by say July 18th if packages in rawhide are orphaned they will be removed? 13:37
f13 notting: there are other releng people around and we can use that day's rawhide to be the basis of finding things that need fixing. 13:37
f13 notting: OLS is the 23rd through the 26th 13:37
f13 notting: but you bring up a good point 13:38
f13 if I'm not doing much those days, will anybody else? 13:38
notting well, i can run pungi and see what happens ;) 13:38
f13 right 13:38
warren f13: I'm at the OSCON and LTSP hackfest that week 13:38
f13 mash + pungi are the interesting parts 13:38
* jeremy can build and smoke test livecds 13:39
f13 well and repoclosure 13:39
notting are we aiming for full repoclosure, or 'Fedora' repoclosure? 13:39
wwoods I don't think full repoclosure is possible, is it? 13:40
wwoods aren't there known, intentional conflicts 13:40
notting and if we're doing a non-blocking freeze, is https://fedorahosted.org/rel-eng/ticket/65 even valid? 13:40
dgilmore wwoods: conflicts are differnet to broken deps 13:40
f13 conflicts are not allowed 13:41
f13 what do we have in Fedora that is "known intentional" conflict? 13:41
notting f13: fedora-logos and generic-logos 13:41
f13 we couldn't figure out a way to make them not conflict? 13:41
f13 did FESCo approve a waiver to the no-conflicts policy? 13:41
notting as it's based on raw file paths, not really 13:41
rdieter alternatives! (duck) 13:42
f13 and what happens to users who accidentally add both to their install? 13:42
f13 notting: it's validish, we still ahve to create the freeze tag 13:42
f13 but no, we don't actually freeze rawhide 13:42
f13 I'll fix that 13:43
notting f13: then they've kickstarted themselves into a pickle 13:43
notting generic-logos isn't selectable 13:43
f13 I'm less happy about this. 13:43
f13 we've always had a strong no conflicts rule 13:43
f13 for the reason that there is no recovering from it, you just plain lose 13:44
notting syslog-ng also Conflicts: with rsyslog, and that's selectable 13:44
notting f13: we can push it to fesco, i suppose. sort of off-topic here 13:44
f13 that needs to be fixed. I consider that a bug. 13:44
warren sendmail and postfix conflict with each other, and alternatives gets around that? 13:46
f13 yes 13:48
notting not explicitly 13:48
rdieter alternatives seems to be an option here, but that's a path to the dark side, if you ask me. Unfortunately, I don't see any other alternative (to Conflicts). 13:48
f13 anyway, that's somewhat offtopic 13:49
f13 are we all happy with moving alpha, even though jwb and I will be out the first 4 days of the "freeze" ? 13:49
notting wfm 13:49
* spot nods 13:50
wwoods so we're moving the freeze to July 22 and GA to Aug 5? 13:50
f13 yes 13:50
notting i get to miss two weeks of alpha bugs. perfect! 13:51
f13 anything else on alpha? 13:52
notting same arch set? 13:52
notting are secondary arches on their own? 13:52
f13 secondary are on their own. 13:54
* notting wonders now, since we have ia64, at what point do we have the ppc discussion 13:54
f13 notting: we don't have ia64 in any sort of automated fashion 13:55
notting true 13:55
f13 I don't think we're ready for secondary ppc until F11~ 13:55
-!- f13 changed the topic of #fedora-meeting to: Fedora Release Engineering - Open Fllor 13:57
-!- f13 changed the topic of #fedora-meeting to: Fedora Release Engineering - Open Floor 13:57
notting there was discussion on another channel about getting bodhi support for removing packages from karma-based autopush 13:58
notting lmacken: status? 13:58
lmacken yes 13:58
lmacken that feature exists 13:58
lmacken the karma automatism is configurable now 13:59
notting is it wired in the UI? 13:59
lmacken yes 13:59
lmacken there is a checkbox in the new update form, and also you can configure the high/low karma thresholds for pushing/unpushing 13:59
lmacken as for the bodhi deployment 13:59
lmacken it's coming along.. a bit slower than expected. The majority of the python-fedora issues have been ironed out, and I'll be pushing out a new release today 13:59
lmacken I polished up bodhi's RPM last night, and started working on it's mod_wsgi deployment 14:00
lmacken shouldn't be too long now.. I'll send an outage notice when I'm ready to take the releng1 bodhi instance down 14:00
notting f13: what's the ticket for the resource deployment for composes-in-PHX? 14:01
f13 lmacken: don't suppose thats frobbable from 'make update' ? 14:01
warren As soon as I figure out how to query for orphans again, I plan on doing an announcement warning people that June 18th is the deadline. 14:02
lmacken f13: it will be 14:02
f13 .ticket 652 14:02
zodbot f13: #652 (Use PHX machines to compose Fedora) - Fedora Infrastructure - Trac - https://fedorahosted.org/projects/fedora-infrastructure/ticket/652 14:02
lmacken i've got a bunch of Makefile.common changes to make 14:02
lmacken (can I have privs to the common repo, btw?) 14:02
f13 don't you already? 14:02
lmacken nope 14:02
f13 hrm, I thought you were a cvsadmin 14:02
lmacken I'll apply 14:03
f13 k 14:03
f13 I wonder if that's something I have to run through FESCo to get you approved for that 14:03
warren I dunno 14:03
warren I think we did vote on adding tibbs as cvsadmin? 14:03
tibbs I am a CVS admin. 14:04
tibbs But I don't know if anyone voted on it. 14:04
warren tibbs: we're wondering if we need fesco to approve lmacken as cvsadmin 14:04
tibbs It's been long enough now that I've forgotten how it was done. 14:04
notting doubt fesco. maybe an informal poll of current cvsadmin 14:05
notting i asked ville to apply b/c he keeps sending patches 14:05
f13 heh 14:05
tibbs I would certainly vote for it were it put before any group or committee I'm on. 14:05
warren Either take cvsadmin or stop being helpful. 14:05
f13 well I'll bring it up to FESCo at the next meeting 14:05
f13 to decide if FESCo itself should vote, or if its happy with some other body voting on it 14:05
warren I'm prepared to just approve lmacken now 14:05
warren but if we should ask fesco, i'll wait 14:06
f13 but since cvsadmin gets pretty darn wide access, I think it should be more than just two people agreeing on it 14:06
warren How many people would be better served with a Makefile.common only access? 14:07
warren (worthwhile to create it?) 14:07
f13 warren: I don't thikn it's necessary 14:07
f13 we can take the one-off patches we get from time to time, but for people like luke who are doing more serious and more frequent work it's worth granting them access 14:07
f13 ok, I've got a visitor, so need to cut the meeting short. 14:09
f13 thanks all! 14:09

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!