Wednesday, 2016-03-09

*** bdemers has joined #openstack-rpm-packaging01:19
*** paragan has joined #openstack-rpm-packaging02:50
*** chandankumar has joined #openstack-rpm-packaging07:50
*** dguitarbite has joined #openstack-rpm-packaging08:10
*** mpavone has joined #openstack-rpm-packaging08:13
*** apevec has joined #openstack-rpm-packaging08:16
*** apevec has left #openstack-rpm-packaging08:17
*** openstackgerrit has quit IRC08:47
*** ihrachys has joined #openstack-rpm-packaging08:48
*** openstackgerrit has joined #openstack-rpm-packaging08:48
toabctlIgorYozhikov, dirk, number80: can you have a look at https://review.openstack.org/#/c/290330/ ?08:48
toabctlthis is to simplify the version handling for the spec files.08:49
number80ack08:49
openstackgerritIgor Yozhikov proposed openstack/rpm-packaging: Add initial spec file for python-glanceclient  https://review.openstack.org/28024608:51
IgorYozhikovtoabctl, will look, test & comment08:53
IgorYozhikovbtw, glanceclient was updated with new epoch syntax08:53
openstackgerritThomas Bechtold proposed openstack/rpm-packaging-tools: Add rpm-packaging-status.py for generating reports  https://review.openstack.org/28190609:01
number80toabctl: do you have a test failures due to oslosphinx?09:28
number80(w/ renderspec)09:28
toabctlnumber80: ? where? what?09:29
number80http://fpaste.org/336089/45751579/09:30
number80I had to explicitly avoid 4.3 as it freezes my laptop09:30
number80looks like a pbr failure09:31
toabctlnumber80: so how does this happen?09:31
toabctlnumber80: when installing renderspec in a venv?09:31
number80Same failure w/ tox or python setup.py test (either in venv or not)09:31
toabctlnumber80: tox -epy27 (and removing before .tox) works here09:33
toabctlwith oslosphinx==4.3.009:34
number80I need to investigate this09:34
number80(had another freeze)09:34
toabctlnumber80: maybe http://stackoverflow.com/questions/21136266/typeerror-dist-must-be-a-distribution-instance-when-resolving-beautifulsoup-d ?09:35
toabctlI'm using setuptools 19.6.209:36
*** paragan has quit IRC09:37
number80we have a slightly older one09:37
number80ok, i managed to retrieve a decent backtrace before freeze, it seems that venv calls system distutils????09:40
number80that's the same issue, but for some reason, I get a much older setuptools09:44
toabctlnumber80: I'm getting setuptools 18.2 when I create a new virtualenv09:55
toabctlnumber80: maybe you need a newer virtualenv package?09:57
*** chthon has joined #openstack-rpm-packaging09:58
*** chandankumar has quit IRC10:00
number80toabctl: maybe, I'm also finding issue with setuptools 20.something but less critical10:02
number80funnily cleaning tox worked but with old setuptools 12.0.5 ...10:03
number80well, I'll continue the review10:04
number80gives me a good excuse to switch to F24 pre-alpha on laptop10:05
toabctl:)10:05
*** bdemers has quit IRC10:17
*** bdemers has joined #openstack-rpm-packaging10:22
IgorYozhikovhttps://review.openstack.org/#/c/284906/ - merge?10:32
toabctlIgorYozhikov: no. it doesn't build.10:32
toabctlIgorYozhikov: I#m looking into that currently10:32
IgorYozhikovhow to recheck?10:32
toabctlIgorYozhikov: I'm testing locally currently10:32
IgorYozhikovok, please let me know, also keystoneclient is on the way, mivanov is going to publish it today10:33
toabctlIgorYozhikov: commented for ostestr with a possible fix10:35
*** paragan has joined #openstack-rpm-packaging10:36
*** paragan has joined #openstack-rpm-packaging10:36
*** chandankumar has joined #openstack-rpm-packaging11:13
openstackgerritIgor Yozhikov proposed openstack/rpm-packaging: Add initial spec for os-testr  https://review.openstack.org/28490611:21
IgorYozhikovtoabctl, os-testr ^^^11:21
toabctlIgorYozhikov: already commented :)11:22
openstackgerritIgor Yozhikov proposed openstack/rpm-packaging: Add initial spec for os-testr  https://review.openstack.org/28490611:23
IgorYozhikovupdated :11:23
toabctlIgorYozhikov: can you also mark the comments as done when they are done? makes reviewing easier (not in this case, but in general )11:23
IgorYozhikovtoabctl, sure, will do11:24
*** mpavone has quit IRC11:50
*** mpavone has joined #openstack-rpm-packaging11:51
*** dguitarbite has quit IRC12:27
toabctlIgorYozhikov: would be cool to get feedback for https://review.openstack.org/#/c/290330/12:28
toabctlnumber80: still fine withj https://review.openstack.org/#/c/284906/ ?12:29
IgorYozhikovtoabctl, going to check in 10 min12:32
*** dguitarbite has joined #openstack-rpm-packaging12:41
*** jpena is now known as jpena|lunch13:39
*** IgorYozhikov has quit IRC13:41
*** dteselkin has quit IRC13:41
*** IgorYozhikov has joined #openstack-rpm-packaging14:16
*** dteselkin has joined #openstack-rpm-packaging14:18
*** chandankumar has quit IRC14:20
openstackgerritMikhail Ivanov proposed openstack/rpm-packaging: Add initial spec file for python-keystoneclient  https://review.openstack.org/29047214:30
openstackgerritMikhail Ivanov proposed openstack/rpm-packaging: Add initial spec file for python-keystoneclient  https://review.openstack.org/29047214:31
IgorYozhikovtoabctl, http://paste.openstack.org/show/489851/14:37
toabctlIgorYozhikov: does it work for you? or is anything missing? I tested it with another spec and looks like it worked14:39
IgorYozhikovtoabctl, doubling bounds14:40
toabctlIgorYozhikov: can you post your source spec.j2 file?14:40
IgorYozhikovtoabctl, http://paste.openstack.org/show/489853/ - in old syntax, so, I believe that I need to retest it with something new, going to check14:44
toabctlyeah. but anyway that's broken. I'll add a test and fix it. thanks!14:45
*** jpena|lunch is now known as jpena14:47
*** chandankumar has joined #openstack-rpm-packaging14:51
IgorYozhikovtoabctl, http://paste.openstack.org/show/489854/14:53
IgorYozhikovwith os-testr14:53
toabctlIgorYozhikov: can you post the spec.j2 as well? I guess the version is still hardcoded there. so if there's a version in py2pkg() then this version is used and the version from global-requirements is *not* used14:55
IgorYozhikovunderstood, os-testr spec.j2 is the same as on review - https://review.openstack.org/#/c/284906/15:01
*** pradk has joined #openstack-rpm-packaging15:09
toabctlhm. so old-style filtering needs a special case.15:19
toabctlI may just go ahead and remove all currently used old-style stuff before we start to build workarounds for it.15:19
IgorYozhikovtoabctl, right - unmanaged versions handled by renderspec via requirements - BuildRequires:  python-pbr >= 1.915:21
IgorYozhikovso it works as was described in commmit message15:21
IgorYozhikovplease, let men know when you will finish with small fixes 4 tests :)15:22
openstackgerritThomas Bechtold proposed openstack/renderspec: Support package versions from global requirements file  https://review.openstack.org/29033015:22
toabctlIgorYozhikov: updated the renderspec PR. should now still work with old-style filters15:22
IgorYozhikovI'm just unset version in spec.j2 for pbr and relaunch rendering - works15:22
IgorYozhikovgoing to check with old oslo15:24
IgorYozhikovstill doesn't work with old syntax. renderspec --spec-style fedora --epoch epo.yaml --requirements /opt/oslo.i18n/requirements.txt rpm-packaging/openstack/oslo.i18n/oslo.i18n.spec.j2 -o - | grep pbr15:31
IgorYozhikovBuildRequires:  python-pbr15:31
IgorYozhikovRequires:python-pbr15:31
toabctlyeah. with the old syntax, the version is not automatically filled.15:31
IgorYozhikovif it is as designed - I'm OK15:31
toabctlbut imo that's not a problem15:31
IgorYozhikovsure, so +215:31
IgorYozhikovand going to prepare local update of renderspec.rpm15:32
openstackgerritThomas Bechtold proposed openstack/rpm-packaging: Use py2pkg() contextfunction everywhere  https://review.openstack.org/29061916:03
toabctlIgorYozhikov: ^^ this removes all the filter usage16:03
toabctlfor py2pkg16:03
toabctlnumber80, dirk: ^^16:04
*** mpavone has quit IRC16:30
IgorYozhikovtoabctl, please check BuildRequires: lower bounds (minimal versions)16:47
toabctlIgorYozhikov: where?16:51
IgorYozhikovtoabctl, I'm about your last commit with sed16:54
toabctlah. something wrong there?16:54
toabctlcan you add a comment?16:54
*** paragan has quit IRC16:54
toabctlah. you did :-)16:55
toabctlIgorYozhikov: well - I can't do that with sed. the goal of that commit is to get rid of the py2pkg filter usage16:55
IgorYozhikovI love sed magic, when use it, not every1 can even read the code :)16:55
toabctlso I wouldn't fix the missing versions here tbh16:56
toabctllet's do that cleanup in another commit16:56
IgorYozhikovtoabctl, ok - sounds reasonable16:56
toabctlIgorYozhikov: if we agree on the global-requirements usage, we have to delete all versions anyway.16:57
IgorYozhikovtoabctl, - done16:57
toabctlthx!16:57
IgorYozhikovtoabctl, not sure that we need to do that. because of some projects have version requirements higher that set in requirements.txt16:58
IgorYozhikovpbr >= 1.8 for example from setup.py | cfg16:58
IgorYozhikovI suggest to update versions which are lower | older that set in requirements.txt16:59
toabctlhm.17:00
IgorYozhikovie. flexibility with manual versions setup with automations17:00
toabctlusually all openstack projects follow the versions from  https://github.com/openstack/requirements/blob/master/global-requirements.txt17:01
IgorYozhikovyes, I know that17:01
toabctlmaybe pbr is a special case17:01
IgorYozhikovhttps://github.com/openstack/oslo.db/blob/4.5.0/setup.py#L28, https://github.com/openstack/oslo.i18n/blob/3.4.0/setup.py#L28 VS https://github.com/openstack/requirements/blob/master/global-requirements.txt#L13517:02
toabctlIgorYozhikov: vs https://github.com/openstack/oslo.db/blob/4.5.0/requirements.txt17:04
toabctlit's somehow strange that in setup.py and requirements.txt are 2 different versions. imo that's a bug17:04
IgorYozhikovtoabctl, I think that we should not cut ability to set up versions manually. :)17:05
IgorYozhikovkeystoneclient has the same situation17:05
toabctlIgorYozhikov: sure. I don't want to do that. but I would try to minimize the manual handling17:05
IgorYozhikovhttps://github.com/openstack/python-keystoneclient/blob/master/setup.py17:05
IgorYozhikovtoabctl, it will be cool if requirements file usage will override only outdated | old versions17:06
IgorYozhikovand of course absent version values17:06
IgorYozhikovand of course absent version values substitutions with values from requirements files17:07
toabctlbut maybe you *want* to have a outdated version (because you know this version also works)17:07
toabctlhm. but if you *want* it, then you can just set it in the requirements.txt file for yourself.17:08
toabctljust thinking loud....17:08
IgorYozhikovadditional cli flag|param ? --keep-old-versions17:08
IgorYozhikovadditional cli flag|param ? --update-old-versions17:08
IgorYozhikovone of them could be used as default behavior17:09
IgorYozhikovand if one of mentioned flags set - will be updated only entries without versions17:10
toabctlhm. still not sure why we really need this. if global-requirements.txt has pbr>=1.6 and we know we need >=1.8 , then we can just set >=1.8 in the global requirements17:10
IgorYozhikovand if you want to use automated things like Jenkins Jobs - how to handle such cases?17:11
toabctland it's anyway not possible to have 2 different rpm's with different versions installed. so having the requirements globally make sense to me.17:11
toabctlyou mean in the pbr example case?17:11
IgorYozhikovagree that within one OS release should be used same versions for all components17:12
IgorYozhikovwe use pbr 1.8 and  minimum requirements set as >= 1.6 - fits us17:14
toabctlyeah17:14
toabctlso my idea is to use the global-requirements.txt from upstream and if downstream (suse miratin redhat ...) needs different requirements they can adjust the file and use that when rendering the specs17:15
IgorYozhikovjust want to be more flexible in setting up an automated environments17:15
toabctlwe can discuss it tomorrow in the meeting again. maybe number80, dirk, ... have other input17:16
IgorYozhikovyour point taken, will think on addition for special key for updates of old versions17:16
IgorYozhikovyep, good idea - discuss17:16
*** chthon has quit IRC17:34
*** ihrachys has quit IRC18:10
*** jpena is now known as jpena|zzz18:14
*** chandankumar has quit IRC18:35
*** ihrachys has joined #openstack-rpm-packaging20:25
*** ihrachys has quit IRC20:51

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!