16:01:25 #startmeeting FAS3 discussion 16:01:25 Meeting started Thu May 26 16:01:25 2016 UTC. The chair is stickster. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:01:25 Useful Commands: #action #agreed #halp #info #idea #link #topic. 16:01:25 The meeting name has been set to 'fas3_discussion' 16:01:29 #meetingname fas3 16:01:29 The meeting name has been set to 'fas3' 16:01:35 .hello pfrields 16:01:36 stickster: pfrields 'Paul W. Frields' 16:01:49 .hello pingou 16:01:50 pingou: pingou 'Pierre-YvesChibon' 16:01:50 #chair SmootherFrOgZ skrzepto pingou puiterwijk 16:01:50 Current chairs: SmootherFrOgZ pingou puiterwijk skrzepto stickster 16:01:55 .hello skrzepto 16:01:56 skrzepto: skrzepto 'Szymon Mucha' 16:02:03 .hello puiterwijk 16:02:04 ^ this guy catches on fast ;-) 16:02:04 puiterwijk: puiterwijk 'Patrick "マルタインアンドレアス" Uiterwijk' 16:02:17 skrzepto++ 16:02:17 pingou: Karma for skrzepto changed to 2 (for the f23 release cycle): https://badges.fedoraproject.org/tags/cookie/any 16:02:22 skrzepto++ 16:02:22 puiterwijk: Karma for skrzepto changed to 3 (for the f23 release cycle): https://badges.fedoraproject.org/tags/cookie/any 16:02:24 skrzepto++ 16:02:25 stickster: Karma for skrzepto changed to 4 (for the f23 release cycle): https://badges.fedoraproject.org/tags/cookie/any 16:02:30 COOKIES!!! 16:02:43 \ó/ 16:02:44 yum yum! 16:02:47 cookie avalanche! 16:03:03 SmootherFrOgZ++ ? 16:03:03 pingou: Karma for laxathom changed to 1 (for the f23 release cycle): https://badges.fedoraproject.org/tags/cookie/any 16:03:11 moar cookies 16:03:14 SmootherFrOgZ++ 16:03:15 puiterwijk: Karma for laxathom changed to 2 (for the f23 release cycle): https://badges.fedoraproject.org/tags/cookie/any 16:04:02 SmootherFrOgZ++ 16:04:02 stickster: Karma for laxathom changed to 3 (for the f23 release cycle): https://badges.fedoraproject.org/tags/cookie/any 16:04:03 \o/ 16:04:13 OK, cookies are out on the table now... let's talk about FAS3 16:04:17 ok now let's take 5 minutes to digest all these 16:04:24 oh wait :D 16:04:31 While you guys do that, I'm on a diet so I leave them to you :-) SmootherFrOgZ: So to bring you up to speed... 16:04:53 pingou, hey daily update I have been trying to setup pagure on my friends aws machine, and I kind of got poor man ci to work on my local machine!! I have a lab exam tomorrow so today I didn't do much progress but I will be working tomorrow properly 16:05:09 skrzepto is a summer intern at Red Hat and he is on our Fedora Engineering team full-time. 16:05:12 thanks fhackdroid :) 16:05:40 SmootherFrOgZ: one of the things pingou, puiterwijk, and I have discussed is the idea of having him help with FAS3 as part of his duties this summer 16:06:03 This might range from building out additional unit testing, to fixing bugs/issues, with the goal of getting to staging 16:06:47 skrzepto: FYI, SmootherFrOgZ started FAS3 as a principal author -- he is one of our long-time volunteer contributors in Fedora :-) 16:06:56 https://fedoraproject.org/wiki/User:Laxathom 16:07:03 * SmootherFrOgZ waves at skrzepto 16:07:21 i dont know how the * thing works by * waves back :) 16:07:30 skrzepto: "/me " 16:07:35 pingou, ^^ 16:07:44 Is there any way I can resolve datagrepper topics into categories? Is there an inbuilt function that does that? 16:07:47 * skrzepto thanks everyone 16:08:28 OK, hopefully that was a good introduction. I was hoping during this hour, we could identify the critical tasks needed to get FAS3 to staging by (suggestion) end of July 16:08:30 Oops. Sorry. Didn't see a meeting was going on. 16:09:15 SmootherFrOgZ: so from the todo list we made last time, are there remaining items? 16:09:23 * pingou has one 16:09:46 we have the last one which is to drop itsdangerous deps 16:10:01 also, include the great artwork made by Maria 16:10:27 the critical part I thin, is the binding with Ipsilon 16:10:46 think* 16:10:51 Ipsilon is? 16:10:52 * stickster notes that http://fas3-dev.fedorainfracloud.org/ doesn't seem to be working again 16:11:09 #info goal is to get FAS3 to staging no later than end of July 16:11:25 #link https://fedoraproject.org/wiki/User:Laxathom/Drafts:FAS3.0 16:11:36 ha, didn't get notification about this one :( 16:13:06 there is something that'll be some work for stg: packaging 16:13:23 we need to look at the list of dependencies and see which are in epel7 16:13:28 and start packaging the missing ones 16:13:36 skrzepto: Ipsilon is the authentication system 16:13:50 runs at https://id.fedoraproject.org 16:13:51 skrzepto: https://id.fedoraproject.org/ -- 16:13:54 *jinx 16:14:15 ah okay :) 16:14:19 Also https://pagure.io/ipsilon 16:14:22 for the code ^^ 16:14:29 github.issue.comment -- sergiomb2 commented on issue #122 on fedora-infra/the-new-hotness https://github.com/fedora-infra/the-new-hotness/issues/122#issuecomment-221919207 16:14:29 Also, https://ipsilon-project.org/ for general info 16:14:36 puiterwijk++ 16:14:43 github.issue.edited -- sergiomb2 edited issue #303 on fedora-infra/anitya https://github.com/fedora-infra/anitya/issues/303 16:15:31 OK, who is going to take the task of pulling out the itsdangerous deps? 16:15:54 should we list all the items first? 16:15:59 and is it better to #action those here, or is it better to look for the issues and assign? 16:16:02 maybe an etherpad could be useful :) 16:16:31 http://etherpad.osuosl.org/fas3-staging-plan 16:16:44 thanks :) 16:16:55 * SmootherFrOgZ will take it 16:17:59 SmootherFrOgZ: do you have a ref about the artwork? 16:18:09 is it like including it in the css? or plain html? 16:18:10 btw, http://fas3-dev.fedorainfracloud.org/ is back up 16:18:16 \o/ 16:18:18 SmootherFrOgZ++ 16:18:49 pingou: no, it should be in the tff and in the icons dir of fas's assets 16:19:18 SmootherFrOgZ: if you have the link to the ticket handy, could we add it to the etherpad? 16:19:32 lemme find the email 16:19:43 skrzepto: ever played with RPMs? :) 16:20:05 https://fedorahosted.org/design-team/ticket/403 16:20:06 pingou, not much was reading up on it last night about how to package them 16:20:56 pingou, I am actually really interested in learning more though :) 16:21:04 If you want experience, I know a piece of software you can package up? :-) 16:21:16 Let's make it something not too scary for a first try :-) 16:21:18 * pingou knows a sponsor who can help 16:21:28 stickster: it's a trivial python package.. 16:21:32 disco 16:21:43 puiterwijk, interested :) 16:21:51 skrzepto: github.com/puiterwijk/flask-oidc 16:21:55 * puiterwijk hides 16:21:59 ^^ 16:22:39 github.star -- Devyani-Divs starred pypingou/pagure https://github.com/pypingou/pagure/stargazers 16:22:59 * skrzepto up for the challenge 16:23:17 Cool. Should be an easy one to start with 16:23:48 skrzepto: we can discuss more about later tonight or tomorrow 16:23:59 SmootherFrOgZ: what do you have in mind regarding the ipsilon binding? 16:24:00 pingou, sounds good 16:24:45 pingou: I think he might mean the Ipsilon login plugin 16:25:04 puiterwijk: so is this a FAS or an Ipsilon issue? 16:25:20 I guess both? 16:25:53 SmootherFrOgZ: puiterwijk: any idea how hard/easy it will be? 16:25:55 The other part of the binding is where FAS needs to accept Ipsilon's tokens, but I think that's not needed for 1.0 16:26:51 pingou: not entirely sure yet. Depends on whether the FAS endpoint is there yet, and how complicated it is 16:26:57 SmootherFrOgZ: ? 16:27:17 pingou: well, should be easy. We have all the FAS scope already 16:27:26 cool 16:27:36 * SmootherFrOgZ looks for his sample 16:27:37 puiterwijk: I guess you'll take this part? 16:27:45 pingou: sure 16:27:51 puiterwijk++ 16:27:56 thanks :) 16:28:07 https://gist.github.com/laxathom/56eed584ec87c913eee4 16:28:08 some thoughts about stg: 16:28:24 * make fas3 be: accounts.stg.fp.o 16:28:39 * let fas2 and fas3 co-exist for a little while 16:28:40 this is a sample on how to request an authorization from an id service 16:28:51 * use this time to port python-fedora to fas3 16:29:04 * drop fas2 when python-fedora is fas3 compatible 16:29:05 SmootherFrOgZ: awesome. I think I might have an Ipsilon plugin by the weekend, looking at the ease of that example 16:29:21 :) 16:29:29 pingou: well, "python-fedora is fas3 compatible" - not everything 16:29:30 this means: deploy fas3 earlier, w/o breaking the existing system but still allowing to test it 16:29:48 The workflow will stay the same even once itsdangerous is dropped 16:29:49 puiterwijk: I'm speaking about stg he' :) 16:29:51 especially, the login function should not work anymore. 16:30:09 pingou: I know. But I'm not planning to support all features that are currently in python-fedora for fas3 16:30:25 for python-fedora, I added support a while ago in a dedicated branch 16:30:27 also, the API will most likely change significantly because of API tokens instead of username/password to FAS itself 16:30:34 puiterwijk: sure thing 16:30:35 I will need to do a repass on that one 16:30:47 I also did it for python-supybot 16:30:54 SmootherFrOgZ: ah. we should work on that. As said, I don't want full backwards compat :) 16:30:57 my point was: do not wait for python-fedora to deploy FAS3 in stg 16:31:09 Right. Absolutely 16:31:19 but do drop FAS2 when we have python-fedora updated, so that while stg is broken, devs have something to test against 16:31:29 pingou: sure thing but python-fedora should be ready-to-deploy as well based on my previous work 16:31:37 and then it's the: Let's fix stg game :) 16:32:10 SmootherFrOgZ: hopefully it'll be that smooth :) 16:32:21 * stickster crosses fingers :-D 16:33:04 :) 16:33:12 SmootherFrOgZ: regarding the propagation of users/groups from FAS to systems, working fine? 16:33:36 You mean system-side as in fasClient? 16:33:51 yes 16:34:24 Yes, you have a new git repo for this 16:34:46 https://github.com/fedora-infra/fas-client 16:35:15 needs to be packaged as well then no? 16:35:21 yup 16:35:34 requires ffi? 16:35:43 github.star -- skrzepto starred fedora-infra/fas-client https://github.com/fedora-infra/fas-client/stargazers 16:36:03 One of the goal is to have it in fp repo 16:36:55 note that using ffi is very dangerous. It is very likely that that will make it not work under apache due to selinux requirements 16:37:02 looks to me like the README needs adjustments :) 16:37:28 If possible, please try to avoid using anything that depends on ffi in its codepath 16:38:12 SmootherFrOgZ: so fas-client depends on fas itself? 16:38:17 The code has not been touched since. So this def need a pass and check if it still required 16:38:33 * pingou wonders about the deps on cliff 16:38:58 Yeah cliff, we had a few talk with Toshio about this one 16:39:17 the list of dependencies of cliff is... interesting 16:39:18 regarding the namespace at build-time 16:39:20 https://pypi.python.org/pypi/cliff 16:40:28 It's openstack.. SmootherFrOgZ: if you want a library to make CLI tooling easier, maybe take a look at http://click.pocoo.org/5/ ? 16:40:55 openstack's lib are... interesting 16:41:01 click sounds nicer 16:41:13 puiterwijk: but click os py3 only 16:41:23 pingou: oh, fun.. hadn;t seen that 16:41:38 so back to good old argparse :) 16:41:39 but then, I'm not even sure fas-client is going to be that complicated of a CLI that it needs an extra tool... 16:41:47 +1 16:41:53 yeah, just argparse should work for what fas-client needs... 16:41:59 puiterwijk: heard of it. didn't have time to dig into it. I will 16:42:07 the less deps the better since it'll be installed on all machines 16:42:11 Yes 16:42:38 wfm 16:42:40 agreed 16:42:50 SmootherFrOgZ: what's the submodules for? 16:43:42 this is to sync-up the release version working against fas-client 16:43:57 and to eady hack on the software 16:44:19 without having to git clone back and forth 16:44:31 the bootstrap handle everything 16:44:54 and you ready to write code with all the required deps we develop 16:45:05 SmootherFrOgZ: so libffi is in fact used by fas3 not fas-client? 16:47:48 * stickster is out of his depth here. 16:47:53 * stickster just listening... 16:47:59 * pingou waits for SmootherFrOgZ :) 16:49:09 I think too. I will confirm this. Sorry I did copy-past the readme from FAS 16:49:37 hm so looking at the setup.py all deps for fas-client are already in epel7 16:49:37 so I might left useless deps 16:49:56 the only one I wonder (beside cliff) is fedmsg 16:50:01 fas-client depends on fedmsg? 16:50:11 Yes, I did tested it on a redhat/centos7 beta release 16:50:15 So what was the decision/consensus on ffi? is that a TODO? 16:50:32 stickster: yes. check ffi for removal 16:51:03 SmootherFrOgZ: ok I'm seeing an issue in fas-client here 16:51:11 pingou: yup. Unless we want to keep the current mechanism we have as in no dynamic sync of accounts 16:51:36 our old set-up: cron pull data from FAS at a regular rate -- issue: all machines would contact FAS at about the same time 16:51:57 current set-up: a fedmsg consumer listens to fedmsg and trigger an ansible playbook updating (sequencially) all machines 16:52:05 * stickster is starting to get a little shaky and may go to get food shortly 16:52:14 basically the new fas-client as it is, brings back the old behavior 16:52:38 all the machine will be listening to fedmsg, all of them will see the change in FAS and all of them will query FAS for info/update 16:53:06 SmootherFrOgZ: I wonder if we shouldn't keep the ansible approach 16:53:27 We could have both I would say 16:53:35 and make fedmsg an optional dep 16:53:43 was going to say that :) 16:54:20 pingou: Since I'm losing the ability to follow the details here... just want to restate that we should end up with a few tasks for skrzepto to be involved in fixes 16:54:33 stickster: agreed :) 16:54:48 now back to your details in progress ;-) 16:55:04 * stickster yields baton 16:55:09 stickster: I do think we'll make good use of skrzepto around the packaging item :) 16:55:13 (the more the merrier) 16:55:34 And perhaps we can get him involved in pulling out some of the unwanted libs, and rewriting those pieces 16:55:35 then, there is the unit-tests question 16:55:42 pingou: +1, that too 16:55:51 * skrzepto taking notes 16:56:22 I think the unit-tests would be a good start so skrzepto can learn the software at the same time 16:56:34 I agree 16:56:41 lmacken: Here's a ticket for the TaskServer instance from yesterday's discussions. https://fedorahosted.org/fedora-infrastructure/ticket/5325 16:56:42 How thing are put together 16:56:55 jflory7: excellent 16:56:59 the only thing about it is that there are currently none, which mean skrzepto has to set-up the entire structure around them 16:57:34 pingou, structure? 16:58:14 skrzepto: how to organize the tests, how to get some data in the db & so on 16:58:30 pingou, ahh i see thanks 16:58:56 no pb :) 16:59:27 having tests, especially for the API will be most useful when working w/ python-fedora 17:00:30 ok, we're 1h in, should we wrap up? 17:00:43 pingou, is python-fedora the way we package python apps into fedora? 17:00:48 skrzepto: of all the points we discussed, are there any that interest you more than others? 17:01:11 skrzepto: no it's a python library providing a number of module allowing to interact with apps in the infra 17:01:34 github.star -- skrzepto starred fedora-infra/python-fedora https://github.com/fedora-infra/python-fedora/stargazers 17:01:35 skrzepto: like the module to login to any of our apps using openid, or the module to query FAS for info about an user, a group 17:01:58 pingou, ahh okay found it on github as well 17:01:59 (the module to interact w/ pkgdb used to be there but got pulled out w/ pkgdb2 and didn't make it back) 17:02:34 pingou, so we need to put the orignal back in? 17:02:51 skrzepto: the original? 17:03:13 pingou, nvm read your last message wrong 17:03:17 ah ok :) 17:04:34 skrzepto: so any preferred task? 17:05:11 just to make sure i didn't miss any points. 1. I should package the python project that puiterwijk posted and 2.) unit tests on python-fedora? 17:05:21 for FAS3 17:05:35 and 3. work on FAS3 17:05:41 forgot that pt. 17:05:42 python-fedora doesn't have unit-tests, may be cool to add some but it'll be hard 17:05:53 so 2. is for FAS3 :) 17:06:32 okay, so write some unittests and do bug fixes for fas3? 17:06:49 if that's interesting to you :) 17:06:56 pingou, it is :) 17:07:23 just want to make sure im working on the correct projects 17:07:45 oh fear not, we'll tell you otherwise :D 17:08:11 SmootherFrOgZ: the ffi deps, do you want to look at it or is that something skrzepto could investigate as well? 17:08:13 skrzepto: I think you also need an introduction to FAS3 thus, would you have a time slot so we can schedule a video call to describe fas? 17:08:37 skrzepto: can investigate it, sure 17:08:41 cool :) 17:08:46 skrzepto: assigning it to you? 17:09:15 pingou, thats the removal of ffi right? 17:09:18 yes 17:09:29 pingou, ill take it :) 17:09:37 assigned :) 17:09:44 excellent! 17:09:47 so we all have something to do that is a blocker for stg :) 17:10:01 I've got two remaining questions: 17:10:18 a) is everyone ok w/ the proposed plan (regarding stg) (cf etherpad) 17:10:35 SmootherFrOgZ: puiterwijk skrzepto ? 17:10:46 if so I'll send it to the list for comments/review 17:11:06 I am 17:11:09 +1 17:11:16 same 17:11:31 quick question I can't seem to find the fas3 repo 17:11:46 skrzepto: check the FAS_3 branch in the fas project 17:12:02 b) when do we schedule the next progress meeting? 17:12:09 next week same time? 17:12:16 github.star -- skrzepto starred fedora-infra/fas https://github.com/fedora-infra/fas/stargazers 17:12:25 pingou: maybe 30 minutes later? 17:12:34 so 16:30 - 17:30 UTC? 17:12:34 16:30 UTC? 17:12:35 pingou, thanks, and anytime works for me 17:12:40 SmootherFrOgZ: ? 17:13:20 not sure, I shold be in FPC meeting at this time. hopefully, today we didn't have too much to discuss 17:13:28 how about friday same time? 17:13:50 puiterwijk: ? 17:13:54 * pingou would prefer 16:00 then 17:14:25 one sec 17:14:35 SmootherFrOgZ: 16:00 UTC would work for you? 17:14:44 next week 16:00 UTC should probably work. 17:14:57 let's do that 17:15:08 #info next meeting Friday June 3rd 16:00 UTC 17:15:17 anything else? 17:15:22 * pingou being called away 17:15:46 pingou, my priority first is to remove ffi then package the python app? 17:15:55 since ffi is a blocker i think? 17:15:58 skrzepto: yes 17:16:24 skrzepto: it's remove/investigate where ffi is used and if it is replacable 17:16:43 Oh wait, I might be in the US so I need to check the time 17:17:06 pingou, okay 17:17:21 ok I'm going to close the meeting here and we can discuss more the next meeting date/time 17:17:23 #endmeeting