| openstackgerrit | Merged openstack/ara-plugins master: Fixed namespacing to use python3 style namespace packages. https://review.openstack.org/605817 | 03:26 |
|---|---|---|
| openstackgerrit | Merged openstack/ara-clients master: Fixed namespacing to use python3 style namespace packages. https://review.openstack.org/605819 | 03:27 |
| dmsimard | apollo13: oh wow I didn't know how much I wanted isort in my life until today | 03:28 |
| openstackgerrit | David Moreau Simard proposed openstack/ara-server master: Only ara is a namespace package, subpackages are not. https://review.openstack.org/605503 | 03:33 |
| openstackgerrit | Merged openstack/ara-web master: Connect ara-web with ara-server https://review.openstack.org/603105 | 03:34 |
| openstackgerrit | Merged openstack/ara-web master: Read versions in Navbar from config https://review.openstack.org/603110 | 03:34 |
| openstackgerrit | Merged openstack/ara-web master: Introduce ListItem in playbooks container https://review.openstack.org/603144 | 03:35 |
| openstackgerrit | Merged openstack/ara-web master: Add playbook parameters in playbook item https://review.openstack.org/603166 | 03:35 |
| dmsimard | sshnaidm|off, tristanC, apollo13: I started brain dumping what's up and todo with 1.0 https://etherpad.openstack.org/p/ara-todo-1.0 | 04:30 |
| dmsimard | I'll continue tomorrow, going to sleep | 04:30 |
| tristanC | dmsimard: about static log generation, it seems like that this is still needed for zuul swift log upload and it may be an issue for zuul to use the version 1.0 | 04:33 |
| dmsimard | tristanC: I am aware | 04:34 |
| dmsimard | html generation (or an equivalent feature fulfilling the use case) would come in a 1.1 | 04:35 |
| dmsimard | I need to ship something, I can't do everything or it won't ship. | 04:35 |
| dmsimard | I've officially been talking about 1.0 for more than a year now: https://dmsimard.com/2017/08/16/whats-coming-in-ara-1.0/ | 04:36 |
| dmsimard | (Interesting how a lot of those ideas I wrote a while back stood the test of time despite changing so much) | 04:37 |
| dmsimard | good night :D | 04:40 |
| tristanC | dmsimard: i meant, the first web interface may fulfill that needs, if only we could make it fetch a static ara.sqlite from zuul log server | 05:01 |
| apollo13 | dmsimard: haha, yes isort & black are basically go fmt for me :) | 06:24 |
| openstackgerrit | Florian Apolloner proposed openstack/ara-server master: Only ara is a namespace package, subpackages are not. https://review.openstack.org/605503 | 06:31 |
| openstackgerrit | Florian Apolloner proposed openstack/ara-server master: Added CI for black/isort. https://review.openstack.org/605439 | 06:33 |
| apollo13 | dmsimard: hrmpf, my fix still breaks ansible integration: https://dpaste.de/PbOv -- the problem is that pbr seems to convert that git requirement into a package name which does not exist on pypi & it's mirrors | 06:46 |
| apollo13 | so much for build reasonably | 06:50 |
| apollo13 | anyways, that is something we have to tackle together; will wait till your are back and then talk options :) | 06:51 |
| openstackgerrit | Florian Apolloner proposed openstack/ara-server master: Small cleanups for django best practices. https://review.openstack.org/605421 | 06:53 |
| *** gvincent has joined #ara | 07:52 | |
| dmsimard | apollo13: the namespace stuff is unfortunately not a priority for me right now, if it's going to break stuff I'd revert the patches on clients and plugins and we can revisit later | 12:45 |
| apollo13 | dmsimard: well it is not the namespacing per se imo; python setup.py install is broken imo since it installs ara 0.1.6 due to how pbr resolves packages | 12:46 |
| dmsimard | yeah, it's not ideal, I agree | 12:46 |
| dmsimard | It's something I want to fix but it's not a blocker is what I mean | 12:46 |
| apollo13 | dmsimard: I have a fix idea :) | 12:53 |
| apollo13 | dmsimard: could you upload ara-server and ara-clients right now onto pypi? | 12:53 |
| apollo13 | with version 0.0.0.something or so | 12:54 |
| apollo13 | the locally installed version from the integration tests should be higher, this way we might get around the issue | 12:54 |
| apollo13 | + it is probably a good idea to reserve the names on pypi | 12:54 |
| apollo13 | also ara-plugins is already on pypi, so that would seem fitting to | 12:55 |
| apollo13 | and would allow us to test if our strategy works (after all we will have that situation in the future too; ie released versions do exist on pypi, but we want to test against the local git checkouts) | 12:55 |
| apollo13 | I understand your concerns about low priority, but I'd argue that I'd want to have a testing strategy now and in the future that works; not that it turns out one repository would be better | 12:58 |
| apollo13 | which it probably would be aside from the fact that it most likely makes packaging harder *sigh* (can something be easy once?) | 12:58 |
| dmsimard | apollo13: if you're willing to hunt that down I'm not going to stop you :p | 13:13 |
| dmsimard | apollo13: before uploading ara-server to pypi we need to land https://review.openstack.org/#/c/605503/ | 13:13 |
| dmsimard | I had to pull it before it'd let me upload it to testpypi | 13:13 |
| dmsimard | otherwise it'd complain about namespacing | 13:13 |
| *** tbielawa has joined #ara | 13:15 | |
| dmsimard | are we good to merge that ? | 13:16 |
| apollo13 | Imo we are good to merge it but zuul won't let it pass | 13:17 |
| apollo13 | we kinda have a chicken egg problem :) | 13:17 |
| apollo13 | also ara-clients should probably not even have a dependency on ara-server (after all it is just optional if you use the offline client) | 13:19 |
| apollo13 | maybe that would already fix the test issue for now, would you be good with that? | 13:19 |
| apollo13 | on the other hand ara-plugins requires ara-clients, so that at least would need to be on pypi | 13:20 |
| dmsimard | zuul will let it pass | 13:29 |
| dmsimard | Because ansible-integration is non-voting | 13:29 |
| dmsimard | apollo13: yes, +1 on dropping dependency from client to server | 13:30 |
| apollo13 | oh clients does not have any tests yet, so that should be "easy" :D | 13:32 |
| openstackgerrit | Florian Apolloner proposed openstack/ara-clients master: Dropped hard dependency to ara-server. https://review.openstack.org/606082 | 13:33 |
| *** apollo13 has quit IRC | 13:48 | |
| *** apollo13 has joined #ara | 13:49 | |
| dmsimard | apollo13: thanks a lot for your help, it's much appreciated | 13:56 |
| openstackgerrit | Merged openstack/ara-server master: Only ara is a namespace package, subpackages are not. https://review.openstack.org/605503 | 14:04 |
| openstackgerrit | Merged openstack/ara-server master: Small cleanups for django best practices. https://review.openstack.org/605421 | 14:15 |
| *** tbielawa is now known as tbielawa|errand | 14:35 | |
| apollo13 | dmsimard: np, way more fun than my current fight with an oracle db | 14:37 |
| *** themurph has joined #ara | 14:46 | |
| *** tbielawa|errand is now known as tbielawa | 14:57 | |
| dmsimard | https://pypi.org/project/ara-server/ | 15:16 |
| dmsimard | https://pypi.org/project/ara-plugins/ | 15:17 |
| dmsimard | https://pypi.org/project/ara-clients/ | 15:17 |
| dmsimard | apollo13: ^ | 15:17 |
| apollo13 | oh nice, I failed against oracle anyways, let me test that :) | 15:17 |
| dmsimard | apollo13: We can also set up Zuul to automatically release each commit to pypi until we hit 1.0 or something | 15:18 |
| apollo13 | I rather not to be honest | 15:18 |
| apollo13 | wasted storage on pypi | 15:18 |
| apollo13 | let me see what ansible-integration now produces locally | 15:19 |
| dmsimard | Well, the main reason why I set up git in the reqs.txt was because I was interested in testing the master branch of everything together | 15:19 |
| dmsimard | since nothing had yet been released | 15:19 |
| dmsimard | Granted, we can probably handle that business logic in the jobs instead of reqs.txt | 15:19 |
| apollo13 | jupp | 15:20 |
| apollo13 | wow, the integration tests seem to work now | 15:20 |
| apollo13 | gotta check which packages it actually used | 15:20 |
| dmsimard | with the packages from pypi ? | 15:20 |
| apollo13 | ara-clients==0.0.1.dev13,ara-plugins==0.0.1.dev12 | 15:20 |
| apollo13 | seems like it | 15:20 |
| apollo13 | that was not what I expected :/ | 15:21 |
| apollo13 | I thought the git urls would win | 15:21 |
| apollo13 | actually I do not know | 15:22 |
| apollo13 | which version would it use from the git checkout? | 15:22 |
| apollo13 | Running setup.py (path:/tmp/pip-install-7wpxukfq/ara-plugins-dummy/setup.py) egg_info for package ara-plugins-dummy produced metadata for project name ara-plugins. Fix your #egg=ara-plugins-dummy fragments. | 15:23 |
| apollo13 | I am testing, but I think pip no longer likes a version there | 15:24 |
| apollo13 | pbr should probably drop that behavior XD | 15:24 |
| apollo13 | dmsimard: ha, win; verbose output confirms that it is __not__ downloading from pypi | 15:25 |
| apollo13 | https://dpaste.de/ergx/raw or do you see anything there? if not, win on the whole line :D | 15:26 |
| dmsimard | apollo13: looks ok to me | 15:43 |
| openstackgerrit | Florian Apolloner proposed openstack/ara-server master: Ensure that add files to a playbook succeeds completly or not at all. https://review.openstack.org/606121 | 15:44 |
| apollo13 | dmsimard: so, thoughts on black and isort? do you want me to add a patch that runs isort & black over the codebase? | 15:46 |
| dmsimard | apollo13: did you see my last comment on that patch ? | 15:48 |
| dmsimard | I legitimately don't know if we want to enforce (gate) very strict standards | 15:49 |
| dmsimard | pep8 is one thing, black is one step higher | 15:49 |
| apollo13 | dmsimard: so to make this easy for you; black does not support a set of parameters | 15:50 |
| apollo13 | it ensures a coding style valid in the eyes of pep008 and tries to format the code in a way that change requests are minimal | 15:50 |
| dmsimard | does django enforce those, for example ? | 15:50 |
| apollo13 | no | 15:50 |
| dmsimard | do you know of a project that does ? | 15:50 |
| apollo13 | but django would loose a bunch of history | 15:50 |
| dmsimard | I'm just curious what it looks like in practice | 15:50 |
| apollo13 | ie git blame would break | 15:50 |
| apollo13 | which is not really an issue for ara in it's current form | 15:50 |
| apollo13 | well black on it's own uses black; and that style be what __every__ project using black gets | 15:51 |
| apollo13 | (since there are no options) | 15:51 |
| apollo13 | as for who is using it, no idea; I know people using it but not sure how much of that is public and not just enforced inside company projects | 15:52 |
| dmsimard | I mean, looking at the diff, it's mostly good things: http://logs.openstack.org/39/605439/3/check/tox-format/2f8258f/ara-report/result/ab792ee3-3cb3-4a54-af97-f178d0524821/ | 15:52 |
| dmsimard | Pretend we fix that, I just don't know if we want to make it voting | 15:53 |
| dmsimard | We could have the job non-voting but then the burden is on the reviewers to actually go look when it fails | 15:53 |
| dmsimard | I have to drop, I'll be back in ~2h | 15:54 |
| apollo13 | have fun | 15:54 |
| openstackgerrit | Florian Apolloner proposed openstack/ara-server master: Added CI for black/isort. https://review.openstack.org/605439 | 15:54 |
| openstackgerrit | Florian Apolloner proposed openstack/ara-server master: Fixed tox warning. https://review.openstack.org/605443 | 15:55 |
| apollo13 | rebased my remaining patches | 15:55 |
| apollo13 | dmsimard: to be honest I see black rather helpful; I've set my editor to use black on save. This is imo easy to tell contributors and ensures a consistent style | 15:57 |
| apollo13 | no back and forth because project X wants style Y and project Z doesn't know what it wants :D | 15:58 |
| *** tbielawa is now known as tbielawa|lunch | 17:10 | |
| dmsimard | apollo13: yeah, don't get me wrong, black is nice | 17:26 |
| dmsimard | the question is only about whether or not we want to gate on it, that's all | 17:26 |
| *** tbielawa|lunch is now known as tbielawa | 18:13 | |
| *** openstackgerrit has quit IRC | 20:07 | |
| *** tbielawa is now known as tbielawa|g0n3 | 21:14 | |
| *** tbielawa|g0n3 has quit IRC | 21:14 | |
| *** themurph has quit IRC | 21:32 | |
| *** openstackgerrit has joined #ara | 23:05 | |
| openstackgerrit | Merged openstack/ara-server master: Ensure that add files to a playbook succeeds completly or not at all. https://review.openstack.org/606121 | 23:05 |
| openstackgerrit | David Moreau Simard proposed openstack/ara-clients master: Add requests to requirements.txt https://review.openstack.org/606218 | 23:20 |
| openstackgerrit | Merged openstack/ara-clients master: Dropped hard dependency to ara-server. https://review.openstack.org/606082 | 23:20 |
| openstackgerrit | David Moreau Simard proposed openstack/ara-clients master: Add requests to requirements.txt https://review.openstack.org/606218 | 23:22 |
| dmsimard | apollo13: do you have an opinion on how we should load the configuration for the backend ? I'm not a fan of settings.py in terms of UX | 23:36 |
| dmsimard | I guess we could have settings.py parse an ini file ? a yaml file ? something else ? | 23:36 |
| dmsimard | I don't know enough about django to be able to tell what are the best practices out there | 23:37 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!