13:02:18 <toabctl> #startmeeting rpm_packaging
13:02:19 <openstack> Meeting started Thu Mar  9 13:02:18 2017 UTC and is due to finish in 60 minutes.  The chair is toabctl. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:02:20 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:02:22 <openstack> The meeting name has been set to 'rpm_packaging'
13:02:34 <toabctl> ping toabctl, dirk, apevec, aplanas, IgorYozhikov, jpena, jruzicka, number80, kaslcrof
13:02:47 <number80> o/
13:02:56 <jpena> o/
13:02:58 <toabctl> #chair IgorYozhikov jpena number80 dirk
13:02:59 <openstack> Current chairs: IgorYozhikov dirk jpena number80 toabctl
13:03:01 <IgorYozhikov> o/
13:03:02 <witek> o/
13:03:33 <toabctl> please add you points to the agenda: https://etherpad.openstack.org/p/openstack-rpm-packaging
13:03:41 <IgorYozhikov> let's adjust ^^^
13:03:48 <IgorYozhikov> :)
13:03:49 <toabctl> might disapear so just go ahead without me...
13:03:58 <IgorYozhikov> ok
13:06:34 <IgorYozhikov> ... and let's start then
13:06:44 <IgorYozhikov> #topic https://review.openstack.org/#/q/status:open+project:openstack/rpm-packaging+branch:master+topic:monasca-rpm - usage of additional variables instead of sourceX
13:06:56 <witek> hello
13:07:14 <IgorYozhikov> witek, please continue
13:07:26 <witek> I wanted to discuss the usage of variables for SourceX
13:07:38 <witek> it has two advantages:
13:07:47 <witek> * better readability
13:08:04 <witek> %{buildroot}%{api_conf_dir}/%{api_conf}
13:08:09 <witek> instead of
13:08:23 <witek> %{buildroot}%{SourceX}%{SourceY}
13:08:52 <witek> * the variable values are sed'ed in templates
13:10:16 <witek> what do you think about this approach?
13:10:27 * IgorYozhikov suggested to use common approach like  %{_sysconfdir}/monasca/%{SourceX} as used for rest projects just to do things in similar way.
13:11:38 <witek> OK, I'm fine with it, common coding style is important
13:11:47 <number80> I agree w/ IgorYozhikov
13:11:51 <jpena> Personally, I think using sed to subsitute variable names around is cumbersome, and leads to a more complex spec
13:11:56 <IgorYozhikov> and it is required to clarify - are we good with such approach and what is written in packaging policy/rules
13:12:33 <toabctl> +1 for trying to be consistent
13:14:08 <witek> can you point me to 'packaging policy/rules' please?
13:14:13 <IgorYozhikov> number80, is there any documented stuff for such cases in fedora packaging rules?
13:17:07 <number80> Yes
13:17:21 <number80> fedoraproject.org/wiki/Packaging/Guidelines
13:17:38 <number80> SUSE: https://en.opensuse.org/openSUSE:Packaging_guidelines
13:18:08 <IgorYozhikov> I found https://fedoraproject.org/wiki/Packaging:RPM_Source_Dir
13:18:24 <IgorYozhikov> where all sources set as files not vars
13:18:36 <toabctl> witek, you can also just look into the other spec.j2 templates and follow what we do there.
13:19:30 <witek> thanks for the references
13:20:18 <witek> are you OK with using sed to substitute the variable values?
13:20:20 <IgorYozhikov> witek, as I saw, you use sed for configuration files update right?
13:20:28 <witek> correct
13:20:43 <IgorYozhikov> might be crudini could help here?
13:21:29 <IgorYozhikov> not sure that my suggestion is the best but for ini like files it should work
13:21:34 <toabctl> witek, we are not doing that in other spec.j2 templates
13:21:53 <toabctl> I would like to be consistent
13:23:29 <witek> I haven't found crudini usage in openstack/rpm-packaging
13:24:19 <IgorYozhikov> also you are modifying files you are already uploaded within commit, so might be better to upload already prepared files in such case ans do not update them during build step
13:24:27 <IgorYozhikov> just suggesting
13:24:41 <IgorYozhikov> s/ans/and/
13:25:05 <toabctl> let's try to avoid crudini
13:25:31 <IgorYozhikov> witek, I suggesting crudini as one of tools which could be used for modification of ini like files
13:25:36 <IgorYozhikov> ah, ok
13:25:51 <toabctl> I would just do what we already do in keystone
13:26:34 <toabctl> and if we decide that we need to support eg a username switch in the .spec we can fix it in all other places. let's try to keep it simple and consistent for now.
13:26:36 <IgorYozhikov> toabctl, are you suggesting to upload already modified files and do not update them during build if those files are not part of source code?
13:26:38 <toabctl> my 2 cents :)
13:27:04 <toabctl> IgorYozhikov I'm suggesting to follow what we already have.
13:27:11 <IgorYozhikov> ok
13:28:09 <IgorYozhikov> so, the rest changes should be provided by cm system, right?
13:28:23 <IgorYozhikov> I'm about config files update
13:29:36 <witek> ok, thank  you for the comments, we'll update the changes
13:29:49 <witek> and we'll be happy about reviews :)
13:30:17 <IgorYozhikov> witek, u r welcome :)
13:30:30 <IgorYozhikov> anything else related to this topic?
13:31:20 <witek> I'm not sure we understand Haikels comment here:
13:31:22 <witek> https://review.openstack.org/#/c/433030/37/openstack/monasca-notification/monasca-notification.spec.j2@28
13:32:50 <IgorYozhikov> number80, ^^^
13:33:43 <IgorYozhikov> I do not see any provides/obsoletes?
13:34:18 <number80> witek: if you rename a package, it has to provides/obsoletes old package name
13:34:25 <number80> if you don't, you'll break updates
13:35:08 <number80> python-monasca-notifications won't be replaced by openstack-xxxx and it will conflicts since they ship the same stuff
13:35:09 <witek> we're not renaming really, just moving to the subpackage
13:35:58 <number80> okie
13:36:04 <number80> lemme look
13:36:17 <witek> ok, thanks
13:37:21 <IgorYozhikov> #topic python3 - feedback to singlespec macros
13:39:02 * IgorYozhikov took a look privately -I'm  more || less fine, but unfortunately still didn't have enough time for hands on experience.
13:39:12 <IgorYozhikov> It's on my TODO list
13:41:13 <IgorYozhikov> any feedback about single spec from RDO side?
13:45:34 <IgorYozhikov> #topic packages reviews (https://review.openstack.org/#/q/project:openstack/rpm-packaging+status:open )
13:45:48 <IgorYozhikov> I want to discuss my update for nova
13:46:02 <IgorYozhikov> how to be with pacement-api
13:46:18 <IgorYozhikov> https://review.openstack.org/#/c/425673/
13:46:33 <IgorYozhikov> and suse tests failures
13:46:57 <IgorYozhikov> 1. about placement-api
13:47:17 <IgorYozhikov> I created wsgi config and put it in package
13:48:22 <IgorYozhikov> question here - do I need to create a separate systemd unit for it or leave as is under wsgi
13:48:44 <jpena> I'd leave it under wsgi, it seems to be the recommended setup
13:49:05 <IgorYozhikov> yes, I did that according to setup.cfg
13:50:53 <IgorYozhikov> ok, and what is about failed tests in SUSE CI - https://build.opensuse.org/package/live_build_log/home:suse-cloud-ci:rpm-packaging-sles12-Master-d6540599133dd567765834b65c02ee2915c63983/openstack-nova/standard/x86_64
13:51:04 <IgorYozhikov> toabctl, could you help with ^^^^^
13:51:29 <toabctl> IgorYozhikov, fix is on it's way. newer pbr needed
13:51:36 <toabctl> aplanas is working on it
13:51:44 <IgorYozhikov> ah, cool thanx
13:52:06 <IgorYozhikov> so, moving forward then
13:52:43 <IgorYozhikov> jpena, according to neutron deps, OVS and Co is on the way.
13:52:56 <jpena> nice
13:53:00 <IgorYozhikov> I found issues with job configs
13:53:33 <IgorYozhikov> some of jobs were using 7.2 snapshots and this is fixing right now by our infra team
13:54:28 <IgorYozhikov> anything else we need to discus here?
13:54:40 <IgorYozhikov> we have 5 min
13:58:50 <IgorYozhikov> open floor || finishing meeting?
13:59:11 <jpena> yep
13:59:49 <IgorYozhikov> #endmeeting