| openstackgerrit | OpenStack Proposal Bot proposed openstack/swauth: Updated from global requirements https://review.openstack.org/257032 | 10:36 |
|---|---|---|
| openstackgerrit | Merged openstack/swauth: Updated from global requirements https://review.openstack.org/257032 | 11:08 |
| onovy | peterlisak, fyi ^ | 11:08 |
| openstackgerrit | Ondřej Nový proposed openstack/swauth: Use pbr for versioning https://review.openstack.org/257042 | 13:40 |
| openstackgerrit | Ondřej Nový proposed openstack/swauth: Use pbr for versioning https://review.openstack.org/257042 | 13:58 |
| openstackgerrit | Ondřej Nový proposed openstack/swauth: Rename of "test_swauth" to "test" fix https://review.openstack.org/257047 | 14:41 |
| openstackgerrit | Ondřej Nový proposed openstack/swauth: WIP: Release 1.1 https://review.openstack.org/257048 | 14:44 |
| openstackgerrit | Ondřej Nový proposed openstack/swauth: WIP: .mailmap https://review.openstack.org/257049 | 14:45 |
| *** nadeem has joined #openstack-swauth | 17:18 | |
| openstackgerrit | Ondřej Nový proposed openstack/swauth: Unit tests for swift_version https://review.openstack.org/257055 | 17:20 |
| notmyname | onovy: careful on pbr integration. use swift as an example and if you must use pbr, then make it only a build-time dependency | 17:21 |
| notmyname | instead of a runtime dependency | 17:21 |
| onovy | https://review.openstack.org/#/c/257042/2/swauth/__init__.py | 17:25 |
| onovy | so this is bad idea? | 17:25 |
| notmyname | https://github.com/openstack/swift/blob/master/swift/__init__.py | 17:26 |
| onovy | https://github.com/openstack-dev/cookiecutter/blob/master/%7B%7Bcookiecutter.repo_name%7D%7D/%7B%7Bcookiecutter.module_name%7D%7D/__init__.py :) | 17:26 |
| onovy | i'm using cookiecutter as reference | 17:26 |
| notmyname | from https://github.com/openstack/swift/commit/0717133197d8929a03ec267d4e3c5b1696f02fb3#diff-225f9353128e1c217cb079c110ff2b72 | 17:26 |
| onovy | hmmm | 17:29 |
| onovy | simple question: why? :) | 17:29 |
| notmyname | there are some historical reasons (eg older things we don't support any more, like lucid), but also just to keep dependencies down on the production machines | 17:30 |
| notmyname | as a build time dependency, it means I only have to install pbr on my build box | 17:30 |
| notmyname | as a runtime dependency, it means I have to build a package for pbr and deploy that with swift | 17:31 |
| notmyname | of course, if you're using other openstack projects, you probably have already done pbr packaging. but that's where we are | 17:31 |
| onovy | but you added dependency to pkg_resources, which is not in requirements.txt right? | 17:33 |
| onovy | it seems to be "standard" to have run time deps to pbr | 17:36 |
| onovy | openstack/python-openstackclient/requirements.txt:pbr>=1.6 | 17:36 |
| onovy | openstack/python-keystoneclient/requirements.txt:pbr>=1.6 | 17:36 |
| onovy | openstack/neutron/requirements.txt:pbr>=1.6 | 17:36 |
| onovy | openstack/keystone/requirements.txt:pbr>=1.6 | 17:36 |
| onovy | openstack/horizon/requirements.txt:pbr>=1.6 | 17:37 |
| notmyname | yeah. and I'm sad about that ;-) | 17:37 |
| onovy | peterlisak, what do you think? | 17:39 |
| openstackgerrit | Ondřej Nový proposed openstack/swauth: Use pbr for versioning https://review.openstack.org/257042 | 17:39 |
| onovy | i think if it's "standard" depenedency and if it's mainstream :) we can go with pbr runtime dep. you don't like mainstream opinions right? :) for example not integrating global-requirements | 17:41 |
| notmyname | I don't like to update global-requirements because I don't want to force deployers to rebuild a whole set of packages when it isn't needed. eg if we use FooLib and so does Horizon, I don't want to force swift deployers to update FooLib when the older version still works just fine with swift | 17:53 |
| notmyname | especially since many swift deployers aren't using other openstack components | 17:54 |
| onovy | yep, but i think this should work better. for example if swift needs: foolib>1.0 and horizon>2.0, then swift is fine with horizon. glob-req should update only if swift have foolib>1.0 && <2.0 | 17:55 |
| onovy | reason for glob-req is to be able to install all openstack on one machine | 17:56 |
| onovy | and foolib 2.1 is fine for swift and for horizon | 17:56 |
| notmyname | I agree! but the current way things work is looking for an exact string match | 17:56 |
| onovy | hmmm, thats new info for me | 17:56 |
| onovy | i thinked it's more smart | 17:57 |
| openstackgerrit | Merged openstack/swauth: Rename of "test_swauth" to "test" fix https://review.openstack.org/257047 | 17:58 |
| *** nadeem has quit IRC | 17:59 | |
| openstackgerrit | Ondřej Nový proposed openstack/swauth: WIP: Release 1.1.0 https://review.openstack.org/257048 | 18:01 |
| *** onovy_ has joined #openstack-swauth | 18:13 | |
| *** onovy has quit IRC | 18:13 | |
| *** openstackgerrit has quit IRC | 18:13 | |
| *** onovy_ is now known as onovy | 18:13 | |
| *** nadeem has joined #openstack-swauth | 18:49 | |
| *** nadeem has quit IRC | 18:56 | |
| *** nadeem has joined #openstack-swauth | 19:22 | |
| *** nadeem has quit IRC | 19:49 | |
| *** nadeem has joined #openstack-swauth | 21:58 | |
| *** nadeem has quit IRC | 21:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!