14:03:17 #startmeeting 14:03:17 Meeting started Mon Jul 20 14:03:17 2015 UTC. The chair is tflink. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:03:17 Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:03:17 #meetingname fedoraqa-devel 14:03:17 The meeting name has been set to 'fedoraqa-devel' 14:03:17 #topic roll call 14:03:18 I was just about to ping you :) 14:03:21 * kparal is here 14:03:26 * lbrabec is here 14:03:30 * mkrizek is here 14:03:32 * jskladan is here 14:03:35 * garretraziel is here 14:03:38 #chair kparal jskladan lbrabec mkrizek garretraziel 14:03:38 Current chairs: garretraziel jskladan kparal lbrabec mkrizek tflink 14:04:34 who wants to go first with status updates? 14:04:49 #topic kparal's status report 14:04:49 #topic mkrizek status update 14:04:52 :) 14:04:55 you go 14:05:06 #info made final touches to latest stable build in the koji directive and pushed it 14:05:09 #link https://phab.qadevel.cloud.fedoraproject.org/T491 14:05:12 #link https://phab.qadevel.cloud.fedoraproject.org/D417 14:05:14 #info dev/stg machine artifacts deployment 14:05:17 #link https://phab.qadevel.cloud.fedoraproject.org/T505 14:05:19 #info various infra fixes 14:05:22 #info code reviews 14:05:24 #info preparing for Flock 14:05:27 #info set up a blog (mkrizek.wordpress.com) 14:05:36 yay 14:05:54 do you some blogpost draft already? 14:05:57 *have 14:06:14 I don't, it's on my todo list for tomorrow though 14:06:41 ok 14:07:01 any comments/questions? 14:07:18 nothing else from me 14:07:25 sounds like progress to me 14:08:07 kparal: next? 14:08:20 #topic kparal's status report 14:08:25 #info simplified development environment setup 14:08:26 #info many reviews 14:08:26 #info fixed artifact path url in resultsdb_directive 14:08:26 #link https://phab.qadevel.cloud.fedoraproject.org/D443 14:08:26 #info other small fixes (artifact path url, libsolv deps, etc) 14:08:56 now I'd like to have a look at the unit test integration with phab, as tested by lbrabec 14:09:54 comments? 14:10:14 none 14:10:15 nothing from me 14:10:24 but I also want to look at the unit test integration bits 14:10:41 tflink: https://phab.qadevel.cloud.fedoraproject.org/T364 14:10:44 and the linked diff 14:11:07 * roshi lurks 14:11:22 kparal: thanks 14:11:28 who's next? 14:11:37 #topic jskladan status updates 14:11:37 #info resultsdb now auto-creates unknown testcases for create_result() 14:11:37 #link https://phab.qadevel.cloud.fedoraproject.org/D431 14:11:37 #info posted testday app's playbook up for a review 14:11:37 #link https://phab.qadevel.cloud.fedoraproject.org/D441 14:11:37 #info changed the ugly font in Phab to something readable 14:11:38 #info reviews, Flock preparations, ... 14:12:05 ugly font? 14:12:49 it seems to default to some microsoft font, with a fallback to Lato 14:13:07 I also find it less readable than before. but nothing serious 14:13:31 I could not stand it (yes, there goes one afternoon :D) 14:13:55 how did you end up changing it? 14:14:08 via userscript, in the end 14:14:42 tried to hack it inside my firefox profile 14:14:58 cool 14:14:59 but this was more discoverable for future changes 14:14:59 userscript is a Firefox extension, I suppose 14:15:37 kparal: yes, greasemonkey, basically 14:16:02 other comments/questions? 14:16:15 no 14:17:06 looks like the font was rebuilt in upstream phab 14:17:13 * tflink will build and update new packages this afternoon 14:17:14 jskladan: has libtaskotron been patched already to stop trying to create the test case? I forgot 14:17:48 kparal: I don't think so, it needs to be done after the resultsdb with the proper changes is deployed 14:18:06 I need to create a ticket for myself, to provide the patch 14:18:12 well just a spec file bump is enough, then the deps will take care of it 14:18:19 ok 14:19:17 next volunteer? 14:19:33 ok, me 14:19:35 #topic jsedlak status update 14:19:35 #info OpenQA: learning about Docker version tags (it's a mess and it behaves weirdly) 14:19:35 #link https://medium.com/@mccode/the-misunderstood-docker-tag-latest-af3babfd6375 14:19:35 #info OpenQA tools: updated README and installguide 14:19:36 #link https://phab.qadevel.cloud.fedoraproject.org/D427 14:19:36 #info OpenQA tools: added code for creating prebuild Fedora Workstation disk image 14:19:36 #link https://phab.qadevel.cloud.fedoraproject.org/D446 14:19:37 #info OpenQA tests: moved logging-in logic into parent class to be able to use it from different places 14:19:37 #link https://phab.qadevel.cloud.fedoraproject.org/D440 14:19:38 #info OpenQA tests: working on using fedfind to provide url for downloading packages for fedup 14:19:38 #info OpenQA tests: created fedup desktop (workstation) tests 14:19:38 #link https://phab.qadevel.cloud.fedoraproject.org/D451 14:20:23 with the updated guides, can we now outsource it to amazon mechanical turk? 14:20:37 definitely 14:20:50 superb 14:20:57 but I've updated only install guide, not "how to write test" 14:21:02 hadn't done it yet 14:21:16 I am planning to sit with pschindl and teach him how to do it 14:22:01 garretraziel: have you ended up with using expect when adjusting the VM images, or have you found out something better? 14:22:20 yep, I am currently using expect 14:22:21 oh, I see the diff, it's expect 14:22:24 ok 14:22:40 if someone finds something better, I can change it 14:22:49 but expect is easy to use and just works 14:23:40 any other comments/questions? 14:23:48 just fyi, fedup functionality might change considerably in F23. but you'd have to talk to wwoods to learn the details 14:24:04 might? i thought that was pretty much decided 14:24:16 ok, I hope that it won't change much from user-side 14:24:19 yes, it sounded like decided 14:24:46 garretraziel: it sounded like there were significant changes in store 14:24:49 yeah, I would suppose that the fedup cli will stay very similar or all the same 14:25:00 one would hope :) 14:25:38 if they are only changes in usage, it will not be a problem 14:26:24 might be a good idea to see if there are more details available for the changeover, though 14:26:37 just in case the user-facing tool changes significantly 14:26:48 yup, I will look into it 14:26:57 lbrabec: you and I are left 14:27:02 I go 14:28:56 did something go wrong? 14:29:24 * tflink assumes that lbrabec is still preparing stuff or got disconnected, goes first for sake of time 14:29:31 #topic "tflink - status report" 14:29:31 #info attempting to improve packages for phabricator, need to get stuff redeployed first 14:29:31 #info preparing to migrate qadevel from cloud to actual infra machine 14:29:31 #info fixed strange issue on taskotron.stg after database outage 14:29:31 #link https://phab.qadevel.cloud.fedoraproject.org/T527 14:29:31 #info helped a little with artifacts deployment 14:29:49 * tflink is really looking forward to getting done with this migration so he can get back to being productive 14:30:48 will we end up with the qa.fp.o namespace for phab? 14:31:01 that's the current plan 14:31:12 struggling a bit with hsts and applicaiton routing at the moment, though 14:31:39 that's a way to avoid status report :) 14:31:54 tflink: thanks 14:32:20 * lbrabec is back in game 14:32:27 * tflink was planning to give a quick overview of the planned changes after status updates are done 14:32:35 if there's enough interest, anyways 14:32:53 any other questions/comments? 14:32:54 sure 14:33:03 no comments 14:33:10 none here 14:33:25 nope 14:33:41 lbrabec: go for it 14:33:45 #topic lbrabec status report 14:33:50 #info exitcode directive 14:33:50 #link https://phab.qadevel.cloud.fedoraproject.org/T450 14:33:51 #link https://phab.qadevel.cloud.fedoraproject.org/D447 14:33:51 #info search fix and improvement in resultsdb frontend 14:33:53 #info finished and pushed package installation from formulae 14:33:55 #info finished and pushed new argument `--local` 14:33:57 #info playing with pytest wrapper for arcanist 14:33:59 #link https://phab.qadevel.cloud.fedoraproject.org/T364 14:34:01 #info will continue to work on taskotron 14:34:47 cool, sounds good 14:35:06 any other comments/questions? 14:35:10 lbrabec: as for the pytest integration, even simple pass/fail indicator that will prevent people breaking tests unawares is a great progress from my pov 14:35:38 we don't need all the features right now 14:35:59 but of course the more the merrier 14:35:59 that shouldn't be hard to do, current implementation of pytest wrapper (from phab/arc team) has a little issue with virtualenv 14:36:37 there is no way how to specify that test from testing/ should be run 14:37:03 or I didn't find one, so I had to hardcode that 14:37:37 if that's something that is not currently in the ticket, let's update it please, so that we can also have a look at it 14:37:48 thanks 14:37:55 ok 14:38:28 * tflink wonders if there's a default that can be set in tox.ini or a pytest config file 14:38:53 good idea 14:39:21 any other comments/questions? 14:39:24 none here 14:39:43 are we forgetting anyone? i think we got everyone for status reports 14:40:23 * tflink takes that as a no 14:40:35 #topic qadevel migration plans 14:40:45 * tflink wanted to go over this quick 14:41:13 there will be emails once more details are figured out, though 14:41:58 basically, instead of just migrating from qadevel.cloud.fp.o, we figured that it'd be worth migrating to qa.fp.o 14:42:14 which would leave us with: phab.qa.fp.o, docs.qa.fp.o etc. 14:42:41 the main advantage here is that we can do subdomains of qa.fp.o and keep a valid ssl cert 14:43:01 since I already have a wildcard cert for qadevel that can be transferred to qa.fp.o 14:43:41 the biggest downside of this is that blockerbugs will need to move - the current plan is to move it from qa.fp.o/blockerbugs to apps.fp.o/blockerbugs 14:43:47 * kparal likes the new names 14:43:55 any questions/comments/concerns? 14:44:16 tflink: what will change with blockerbugs being moved? 14:44:30 mkrizek: hostname and url 14:44:44 * tflink is planning to have a redirect from qa.fp.o/blockerbugs to apps.fp.o/blockerbugs 14:44:58 ok, so the downside is just the work that needs to be done in order to move it 14:45:19 once production is migrated (looking like wednesday at this point), I'll need some help updating a bunch of links, though 14:45:37 yeah, i'm hitting a few more issues than I was anticipating 14:46:11 but it should mostly be a matter of testing stg and making sure everything still works 14:46:57 what's the reason to move blockerbugs from qa namespace to apps namespace? 14:47:02 yes, I used the s-word 14:47:18 kparal: routing 14:48:09 ok, /me doesn't need to know all the details, I'd make heads and tails of it either :) 14:48:15 with the way that it's currently set up and how we have the current qadevel set up, keeping blockerbugs on qa.fp.o would lose load balancing and proxying, adding a bucnch of works 14:48:20 *wouldn't 14:48:20 s/works/work 14:48:40 *either->anyway 14:48:56 in order for blockerbugs to work at it's current url, the main fp.o proxies need to own qa.fp.o 14:49:21 ok 14:49:25 but since we're doing git hosting, qadevel can't use those proxies and would need to own qa.fp.o in order to get phab.qa.fp.o etc. 14:50:06 * tflink just had an idea 14:50:13 if it works, there will be much facepalming 14:50:30 but this has already gone longer than I was hoping 14:50:37 so we go on to ... 14:50:41 #topic open floor 14:51:21 I'm wondering if we might be better served going to the style of meeting coordination that infra does now that we have a few more people 14:51:34 ie, creating a shared document ahead of time 14:51:47 but that discussion can wait for list or another day 14:52:52 if there's nothing else, we can end just barely before the hour :-/ 14:52:58 yay 14:53:01 * tflink sets fuse 14:53:41 boom! 14:53:47 thanks for coming folks 14:53:53 * tflink will send out minutes shortly 14:53:57 #endmeeting