16:00:41 #startmeeting Fedora QA Meeting 16:00:41 Meeting started Mon Jan 18 16:00:41 2010 UTC. The chair is jlaska. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:41 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:00:47 #meetingname qa 16:00:47 The meeting name has been set to 'qa' 16:00:52 meeting powers go! 16:01:07 queue the robot sounds 16:01:10 #topic gathering 16:01:12 * kparal is here 16:01:20 wwoods: kparal: howdy 16:01:35 anyone else primed and ready for another edition of the QA meeting? 16:01:47 * tk009 16:01:56 morning 16:02:01 tk009: adamw hi there 16:02:15 sorry, /me is tweaking with stuff again 16:02:45 not sure if Viking-Ice or maxamillion are around too 16:02:52 * Viking-Ice here 16:02:59 Viking-Ice: oh good, welcome! 16:03:57 okay, I put a ping out to Adam Miller, but I think he might be busy at the moment 16:04:01 so let's get started ... 16:04:13 Working off the agenda here http://lists.fedoraproject.org/pipermail/test/2010-January/088008.html 16:04:20 #topic Previous meeting follow-up 16:04:40 #info maxamillion bring back conversation about Xfce 4.8 updato on xfce@lists.fp.o and inform about conclusion 16:05:06 I don't have any updates on this, anyone else? I'll leave it on the list until we get a chance to sync up with Adam M. 16:05:43 moving on 16:05:46 #info Viking-Ice write lxde tests and find a date for the test day 16:06:05 Viking-Ice: any updates on definingsome LXDE tests for a possible test day? 16:06:19 Not really looking at it.. 16:06:27 Created ticket in QA 16:06:46 [Fedora QA] #46: Write Test cases for LXDE components. 16:06:58 To keep track will update there as thing progress.. 16:07:18 okay thanks, lemme add that to the 'draft' test day schedule too (https://fedoraproject.org/wiki/Talk:QA/Fedora_13_test_days) 16:07:26 Ok 16:07:53 Viking-Ice: thanks for the update 16:08:09 next up .... 16:08:13 #info adamw and rhe to discuss ways to add install testing as a QA activity 16:08:32 hopefully rhe is sleeping, adamw ... sounds like you two made some progress? 16:08:32 well, rhe's not around, so i'll do it 16:08:49 rui did, i was too busy playing around with shiny shiny toys. =) 16:09:06 rui drafted an installation testing page: 16:09:13 #link lxde fedora list 16:09:15 grr! 16:09:18 #link https://fedoraproject.org/wiki/QA/Installation_Test 16:09:23 stupid paste buffers 16:09:38 the idea is to have that big explanatory page, and a bit in the 'joining' page which would link to it 16:10:04 I wondered if the big page should merge with https://fedoraproject.org/wiki/QA:Fedora_12_Install_Test_Plan in some way; rui thinks not at present 16:10:08 #info the idea is to have that big explanatory page, and a bit in the 'joining' page which would link to it 16:10:39 i'm going to take some time today to have a look at both pages and see whether I should disagree some more =), do a bit of a proofread, and come up with a draft for the bit to be added to the Joining page 16:11:08 adamw: I can see the drive to avoid duplicate content. Hopefully we can do that while preserving the concept of a 'test plan' 16:11:25 ayup 16:11:28 rather ... while addressing the same issues that the test plan addresses 16:11:38 whether it takes a different form or not 16:11:48 cool, nice update 16:12:04 next up ... 16:12:09 #info jlaska to reach out to beland for guidance/ideas on how to document the process (or point to existing documentation) for how bugs are noted (common_bugs, release notes, install guide etc...) How to determine which bugs land in which place? 16:12:51 another week and I've not kicked off this small task :( I'll prioritize this for today and see if beland and awilliam have some thoughts as to whether there are any ideas for improvement here 16:13:16 #action jlaska - reach out to beland for guidance/ideas on how to document the process for how bugs bubble through different release documents 16:13:36 cwickert: perfect timeing 16:13:37 timing 16:13:53 cwickert: there's a previous meeting note to follow-up on ... 16:13:57 #info cwickert to file an infrastruture ticket for the lxde mailing list 16:14:09 did you have any updates you wanted to share? 16:15:27 okay, we can come back to that later 16:15:28 jlaska: can be closed 16:15:36 cwickert: ah okay, thanks 16:15:39 nirik fixed it 16:15:44 what's the new list? 16:15:59 lxde@lists.fpo 16:16:03 #link https://admin.fedoraproject.org/mailman/listinfo/lxde 16:16:08 nice, thanks 16:16:21 alright, moving along ... 16:16:23 #topic Security Policy Update 16:16:32 On Friday's FESCO Meeting, Adam and Will responded to questions concerning building project-wide consensus around a security policy for Fedora (see FESCO ticket#297) 16:16:56 .fesco 297 16:16:57 cwickert: #297 (Please consider the idea of a security (privilege escalation) policy) - FESCo - Trac - https://fedorahosted.org/fesco/ticket/297 16:17:02 adamw: can you talk to how things went, and what's next? 16:17:09 yeppers 16:17:14 short version: they kicked it back to us 16:17:32 with a promise that if we come up with some kind of draft policy, people who have a clue about security will be gracious enough to review it 16:17:53 so as far as FESCo is concerned the next thing that happens is I come back with a draft privilege escalation policy in hand 16:18:13 fesco needs at least some kind of draft 16:18:41 #info as far as FESCo is concerned the next thing that happens is I come back with a draft privilege escalation policy in hand 16:19:16 adamw: is that something you can absorb for this release? 16:19:37 i can write up something quick 'n' dirty based on the famous blog post 16:19:43 i'll run it by you lot for review before taking it to fesco 16:20:24 #action adamw to build on the famous spot security blog post and draft something quick'n'dirty for QA review 16:20:47 adamw: what's this look like when it's all done to your liking? 16:21:23 translation ... what does 'finished' look like for the security stuff? 16:21:45 dont we need to have a strong base security policy in place that spinners need to document ( including desktop/ all DE ) on what they differ from that policy ? 16:21:57 we have some kind of fesco-approved policy and some test cases for it 16:22:07 Viking-Ice: you'd've thought so. :) 16:22:35 adamw: okay 16:22:35 It's the only thing that makes sense to do 16:24:09 adamw: Viking-Ice: anything else to discuss, otherwise, let's move on 16:24:27 nope, i'm okay 16:24:47 and kudos for the double contraction :) 16:25:02 #topic fedora-release-rawhide 16:25:10 #info wwoods noted that a new fedora-release-rawhide package may impact how we document opting into (and out of) testing rawhide via yum (see Releases/Rawhide#Testing_Rawhide) 16:25:24 wwoods, got anything you'd like to add on this topic? 16:27:10 take that as a no 16:27:12 no, just that we'll need to talk to nirik et. al. to make sure we understand how to get on (and off!) rawhide 16:27:24 and make sure that all the places in the wiki where we talk about that 16:27:28 get updated to reflect the new state of the art 16:27:37 ah, good 16:27:40 yeah, I was going to update the Rawhide page when it lands. 16:27:43 any help welcome. 16:27:55 nirik: when you say rawhide wiki page ... is that https://fedoraproject.org/wiki/Releases/Rawhide ? 16:28:03 yeah, 16:28:24 I'm sure we can ask for help on test@l.fp.org following any instructions stubbed on the wiki 16:28:26 there really doesn't seem to be any/much other in the way of docs that talks to people about rawhide or how to move to it. 16:28:33 esp since rawhide involvement will be heating up again 16:29:28 #action nirik intends to update the [[Releases/Rawhide]] wiki page to reflect the changes (help appreciated) 16:29:54 that's another area I think beland has some experience in as well, he's helped tweak that page in the past iirc 16:30:02 basically it will just be 'yum install fedora-rawhide-release' then edit /etc/yum.repos.d/fedora-rawhide.repo and set 'enabled=1' 16:30:22 cool 16:30:38 alrighty ... wwoods, nirik: thanks for the updates 16:30:45 moving on ... 16:30:57 #topic AutoQA Updates - rpmguard (kparal) 16:31:11 right 16:31:33 so last week me and james enabled sending the rpmguard results to autoqa-results ML 16:31:40 so you can have a look 16:31:45 together with rpmlint results 16:32:13 in the beginning we experienced some networking issues and many tests failed, but it seems to be working well now 16:32:13 #info last week, enabled rpmguard test results to autoqa-results@l.fh.org 16:32:27 kparal: I don't think we ever really figured those errors out, did we/ 16:32:28 ? 16:32:33 nope 16:32:52 well, it works now :D 16:33:09 well, nothing much more about rpmguard, even won't be next week, since I'm doing RHCT 16:33:21 Good luck on the RHCT front 16:33:29 but I also started some discussion in autoqa-devel about possible autoqa architecture 16:33:49 about creating server to receive all results and provide an API to access them 16:33:55 so all comments welcome 16:34:06 there are even pictures there! ;) 16:34:15 heh! 16:34:18 #info initiated discussion on autoqa-devel@ about refactoring capturing test results 16:34:21 #link https://fedorahosted.org/pipermail/autoqa-devel/2010-January/000120.html 16:34:27 definitely agree we need *some* better way to store results 16:34:52 I miss comments "no, you're completely wrong, it should be this way: ...." 16:35:01 heh 16:35:02 anyone provide them, please :) 16:35:24 ok, that's about it from me 16:36:08 kparal: does a larger discussion on the types of package update tests influence the discussion around how to store/present their test results? 16:36:19 or is that something that can happen in parallel? 16:36:59 well, the package update tests will probably have very similar set of output information that should be stored in the results database 16:37:31 so we can do it simmultaneously I believe, we don't have to wait for it to finish 16:37:46 okay, gotcha. thanks 16:38:11 #info next steps ... continue discussion around improving the mechanism for storing test results 16:38:21 alright, moving on to ... 16:38:32 #topic AutoQA Updates - deps/conflicts prevention (wwoods) 16:38:59 okay, so the notes I have from last week are that work is underway w/ lmacken to figure out how to gather information for a post-bodhi-update trigger 16:39:16 yeah, it turns out there's some API calls that alllmost do what we need 16:39:57 lmacken said he was would work on adding the call we need in the next bodhi code update 16:40:20 so in the meantime I started work on the depcheck test itself 16:40:34 which is.. *close* to working, but not quite there. it's tricky business. 16:41:05 the code is in git: http://git.fedorahosted.org/git/?p=autoqa.git;a=blob;f=tests/depcheck/depcheck 16:41:21 hah whoops I forgot to add the GPLv2 boilerplate text 16:41:28 if it involves prco sets, I think my brain will melt! :D 16:42:01 yeah it's kind of complicated - involves PRCO data and package sacks and processing obsoletes and whatnot 16:42:07 but not insurmountable 16:42:15 just gonna take some time to get it right 16:42:33 I like how it's another test designed to be run by hand first 16:42:52 you and kparal are trailblazers there :) 16:43:13 right, that's the design - you give it the name of a repo to check against and a set of new packages to test 16:43:23 #info lmacken is working on adding a missing API call into the next bodhi code update 16:43:51 it checks all the dropped provides, added requires, new conflicts, etc. 16:43:53 #info wwoods started design on an initial depcheck test ... not 100% complete, but close 16:43:59 #link http://git.fedorahosted.org/git/?p=autoqa.git;a=blob;f=tests/depcheck/depcheck 16:44:22 and returns failure if there's any problems that would break repoclosure 16:44:35 it typically takes only a few seconds to run, even when downloading the metadata 16:44:44 that's not too bad at all 16:44:56 which means we can very likely run it for every new rawhide package build / bodhi update 16:45:10 without too much trouble 16:45:49 nim-nim: ping 16:46:16 so that's depcheck. 16:46:41 sweet, anything else on the depcheck/conflicts front you want to note? 16:47:14 I think you said so already, not sure if there were other non-depcheck items to discuss 16:47:57 I'm going to move on just to keep things going ... but we can come back if needed 16:48:05 nothing comes to mind 16:48:13 wwoods: okay, thanks for the updates! 16:48:24 #topic AutoQA Updates - install automation (lili/rhe) 16:49:04 no big updates to share just yet ... I discussed with lili last week and he is refining a sample python script that automates a virt DVD install 16:49:47 the plan is to get that working, refine it per the requirements discussed in previous meeting around what inputs an automated virt install test would need to accept 16:50:30 wwoods: kparal: while lili is refining his test, do you two have any thoughts on a good way to share his test development progress? 16:50:56 meaning, should I ask lili to contribute the script into git ... or should we continue working it outside of git until it's at a stable point? 16:51:35 I think it can surely be in git right now, that's what we have VCS for 16:51:52 that was my thought, but wasn't sure how you two felt about it 16:51:55 it can be a separate branch 16:51:59 and merged later 16:52:11 yeah - does he already have commit access? 16:52:30 wwoods: I don't believe he does yet. He's still new to git and python 16:52:37 branches are cheap and (fairly) easy in git 16:53:05 but if he's more comfortable just working on a local git repo until he's satisfied 16:53:09 and sending patches to the list for review 16:53:19 that works too 16:53:36 wwoods: kparal okay thanks gents ... I'll be sure to pass that along 16:53:45 and lili can choose which route he is more comfortable with 16:54:02 #info next steps - continue refining virt dvd install test 16:54:21 #info next steps - look for ways to contribute tests back into git (private checkout, or public branch) 16:54:28 okay, next up ... 16:54:35 #topic AutoQA Updates - packaging/deployment (jlaska) 16:54:49 my two tasks last week were 16:55:21 #info respond to package review feedback for autotest-client (bug#548522) 16:55:22 Bug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=548522 medium, medium, ---, nobody, NEW, Review Request: autotest-client - Autotest is a framework for fully automated testing 16:55:43 #info scope out packaging work required for building the autotest BuildRequires gwt (and all of it's bundled deps) 16:56:04 I've got some re-work I'll be attempting with permissions on the autotest-client package 16:56:11 should have that in this week 16:56:34 I'm getting a better handle on the effort required to package gwt (and deps) 16:56:37 #link https://fedoraproject.org/wiki/User:Jlaska/gwt 16:57:02 #info I'll be spending time this week to identify action plans for each of the items listed under 'status uncertain' 16:57:44 once autotest-client is blessed ... I'll file a review request for autoqa 16:57:58 after that ... we should have all we need to deploy the results front-ends in Fedora infrastructure 16:58:06 That's all I have on the packaging/deploy front 16:58:07 yaaay 16:58:40 okay, so let's open things up ... 16:58:43 #topic Open discussion - 16:58:54 anything folks would like to discuss before closing out the meeting? 16:59:40 when are we scheduled to get rawhide install images? 16:59:46 wwoods: aha, thanks for bringing that up! 17:00:05 so this Thursday (1/21) is a scheduled install image drop to run the rawhide acceptance test plan 17:00:19 #topic open discussion - When is next rawhide install image drop? 17:00:26 #info so this Thursday (1/21) is a scheduled install image drop to run the rawhide acceptance test plan 17:00:42 #link http://poelstra.fedorapeople.org/schedules/f-12/f-12-quality-tasks.html 17:01:04 I sent an email out to dcantrell and jkeating for guidance on indentifing who is doing what for this drop 17:01:12 and filed a rel-eng ticket just to track creating the images 17:01:13 * jlaska finds link 17:01:26 #info https://fedorahosted.org/rel-eng/ticket/3277 17:02:05 so we ran through rats_install (finished up last monday) ... and this time should include some bug fixes for issues identified 17:03:52 Rats test results against anaconda-13.16 (using http://clumens.fedorapeople.org/updates.img) - http://jlaska.fedorapeople.org/rats.png 17:03:59 #topic open discussion - 17:04:12 wwoods thanks for reminding me! :D 17:04:19 all nicely green 17:04:35 an i386 compose wasn't created ... so there were no results 17:04:56 hi 17:05:23 any other issues to discuss? Otherwise, let's close things out and get back to work 17:05:34 RodrigoPadula: hi 17:05:53 famsco meeting ? 17:06:05 RodrigoPadula: one moment, wrapping up fedora-qa meeting 17:06:11 okay gang ... thanks for your attention 17:06:12 ops.. soryy 17:06:21 as usual, I'll follow-up to the list with minutes 17:06:27 #endmeeting