#openstack-meeting-3 log

19:00:41 <briancurtin> #startmeeting python-openstacksdk
19:00:41 <openstack> Meeting started Tue Dec  8 19:00:41 2015 UTC and is due to finish in 60 minutes.  The chair is briancurtin. Information about MeetBot at http://wiki.debian.org/MeetBot.
19:00:43 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
19:00:45 <openstack> The meeting name has been set to 'python_openstacksdk'
19:00:57 <etoews> o/
19:00:57 <briancurtin> sorry about that, my client disconnected but appeared as if it was connected so i was in a meeting with no one
19:01:09 <etoews> the best kind of meeting
19:01:47 <briancurtin> does anyone have anything specific to bring up?
19:02:42 <terrylhowe> Only exciting news I have is OSC/SDK integration moving along thanks to help from Tang Chen and Richard Theis
19:02:44 <briancurtin> overall things are moving along reasonably well, and have had good contributions from a few others lately
19:03:04 <briancurtin> yeah, richard has been pitching in good stuff
19:03:22 <etoews> yep. good stuff from rtheis!
19:03:31 <etoews> can we have a quick look at https://review.openstack.org/#/c/254307/
19:03:46 <rtheis> etoews: thanks
19:04:40 <etoews> keystoneauth isn't too forthcoming about what's behind that 400 error
19:04:53 <etoews> http://logs.openstack.org/07/254307/2/check/gate-sdk-dsvm-functional/7f58e6c/console.html#_2015-12-07_23_49_46_783
19:04:58 <briancurtin> etoews: that’s good, and i agree with terry to fix the other stuff in another review. we’ve generally been pretty good about fixing one thing in one commit
19:05:37 <etoews> and i haven't been able to reproduce it with my own devstack :(
19:05:40 <terrylhowe> I just merged it. It is progress, fixining these tests is tricky but super important
19:05:56 <etoews> agreed. and thanks.
19:06:29 <etoews> i'll add some deletion examples next for key pair and servers. just so stuff gets cleaned up.
19:07:13 <etoews> terrylhowe: do you want to abandon this patch set then? https://review.openstack.org/#/c/250463/
19:07:41 <terrylhowe> yeh, well there is still the fix for the shell script there
19:08:34 <etoews> ah. i see it. well, whatever direction you want to go with it.
19:10:43 <etoews> does the 1.0 list warrant a review? https://launchpad.net/python-openstacksdk/+milestone/1.0
19:12:28 <briancurtin> etoews: i’ve been pruning here and there but haven’t made any changes in about 2 weeks. do you have any suggestions on things to promote/demote?
19:13:03 <briancurtin> or did you mean to just have people take a look at it for what’s next?
19:13:19 <etoews> why don't we step through the confirmed/in progress one by one and just see if anything is to be done on the issue?
19:14:19 <briancurtin> on https://bugs.launchpad.net/bugs/1466192 i just need to do it, so that’s an easy review
19:14:19 <openstack> Launchpad bug 1466192 in OpenStack SDK "Update contributor guide" [Critical,Confirmed] - Assigned to Brian Curtin (brian.curtin)
19:14:37 <briancurtin> for https://bugs.launchpad.net/bugs/1479977 that seems like it’s still stuck in a holding pattern
19:14:38 <openstack> Launchpad bug 1479977 in OpenStack SDK "Get sdk docs on developer.openstack.org/sdks/python/openstacksdk" [Critical,Confirmed] - Assigned to Everett Toews (everett-toews)
19:14:43 <etoews> one sec
19:14:44 <briancurtin> we now have the right docs theme, though
19:14:49 <etoews> yep
19:15:42 <etoews> https://review.openstack.org/#/c/246591/
19:16:10 <etoews> no new comments on my latest patch set. i think i just need to highlight it again to infra.
19:16:38 <briancurtin> im not confident in actually reviewing that, but i can say things look spelled correctly...
19:17:43 <etoews> there's a bunch of infra magic behind it. i have no idea what i did there.
19:19:14 <briancurtin> I’m assuming they’ll take a look and if it’s wrong, hopefully just
19:19:25 <briancurtin> help in figuring it out so you don’t have to guess at it?
19:20:01 <etoews> ya. i've already got a lot of help from andreas. i think it's in good shape now.
19:20:09 <briancurtin> cool
19:20:38 <briancurtin> https://bugs.launchpad.net/python-openstacksdk/+bug/1468086 — what should we do here?
19:20:38 <openstack> Launchpad bug 1468086 in OpenStack SDK "Mark services as Beta to indicate that the interfaces may not be stable" [Medium,Confirmed]
19:21:43 <etoews> is there anything we would mark as beta today?
19:22:17 <etoews> said another way, are we confident in supporting all of the interfaces as is in their current form?
19:23:09 <briancurtin> i kind of think we are, but i still need to do somewhat of an audit on that. i did a mini-audit around find a few weeks ago that just found us naming things differently, which was a trivial fix, but functionally was fine
19:23:46 <briancurtin> so maybe we do nothing and just have it in LP but untargeted?
19:24:02 <terrylhowe> the only thing that makes me nervous really is telemetry
19:24:40 <etoews> briancurtin: did you want to create a bug for an interface audit?
19:24:56 <briancurtin> yeah, that’s one. there’s also a review out that would potentially restructure its resources
19:24:56 <briancurtin> etoews: yeah i’ll file one
19:25:03 <terrylhowe> but looking at it the proxy has nothing right now, the resource classes seem a little messed up
19:25:07 <etoews> terrylhowe: do you want to mark the telemetry interface as beta?
19:25:31 <terrylhowe> maybe or quick clean up
19:25:38 <briancurtin> terrylhowe: yeah, that one is fairly disconnected, which probably feeds into why it often breaks functional tests
19:25:55 <etoews> terrylhowe: whatever works for you.
19:25:56 <terrylhowe> marking as beta would probably be easiest
19:26:41 <etoews> terrylhowe: do you want to create a bug for that or just go for it?
19:27:13 <etoews> and what mechanism do we use to mark it as beta?
19:27:40 <etoews> docstring and a warning?
19:27:59 <briancurtin> etoews: something in restructuredText markup would be ideal, and the warnings module is actually a really good idea
19:29:18 <briancurtin> etoews: i have been meaning to look at creating some rst shortcut to make identifying admin calls easier, but that’s actually a whole different conversation/organizational structure from this…but still, something like “.. beta:: <blah>”
19:29:44 <briancurtin> i forget how to make new admonitions, but it can be done
19:31:28 <etoews> terrylhowe: do you want to create a bug for that or just go for it?
19:32:05 <terrylhowe> the existing bug is fine, I just don’t know the steps etoews
19:32:39 <terrylhowe> the proxy only exposes capabilities list, so it isn’t much
19:33:01 <terrylhowe> unfortunately, that seems to fail as devstack changes their capabilities
19:33:45 <etoews> wait. are we talking about metric or telemetry?
19:33:47 <briancurtin> if the proxy only includes /capabilities we should just remove the proxy entirely, leave the resources, and wait for someone to come in and make a useful telemetry proxy
19:33:57 <etoews> it sounds like you're talking about metric
19:34:07 <briancurtin> (whichever this is about)
19:34:37 <briancurtin> i think somewhat related to any of this is https://review.openstack.org/#/c/226689/
19:35:56 <etoews> ya
19:36:29 <etoews> terrylhowe: is it metric or telemetry you want to mark as beta?
19:37:34 <terrylhowe> I think it was metric that was blowing up, but now I’m not sure
19:38:24 <briancurtin> that should be another general audit - the telemetry, metric, and alarm related things have been more fragile than the other services, both on devstack in the gate but also quite often on my own local devstack
19:38:45 <terrylhowe> no, it was telementry
19:41:09 <etoews> okay. well how about this.
19:41:10 <etoews> for every function in the _proxy of a beta service include at the top of the docstring "BETA: This function is subject to change" (or something like that)
19:41:52 <etoews> for every resource of a beta service, include at the top of the class docstring "BETA: This class is subject to change" (or something like that)
19:42:24 <etoews> then include a warning in the whatever_service.py file?
19:43:09 <terrylhowe> That definitely works for me and I’m fine with that, but I’d be okay being less through and just having a warning at the top of the proxy
19:43:22 <terrylhowe> I can’t seem to spell today
19:44:05 <briancurtin> i would think the beta notification would go into the top of something like doc/source/users/proxies/a_beta_proxy.rst
19:44:38 <etoews> sure. for the Proxy class of a beta service, include at the top of the class docstring "BETA: This class is subject to change" (or something like that)
19:45:24 <etoews> briancurtin: good idea to include beta info there too.
19:45:43 <briancurtin> including it in the class docstring means they’d have to read somewhat far into the doc, through the initial argument list, in order to get to it. i think we can head that off even earlier by just including it straight into the page where the docs are being held and the docstrings are pulled into
19:46:37 <etoews> but i'd want to include it in the code docstrings too for people who skip the docs and jump straight into the code
19:46:38 <briancurtin> i’m not 100% sure where to raise a warning from. now that i think of it, the blah_service.py file will always get loaded, but they might not be using that thing, so i’d rather not warn then.
19:46:59 <briancurtin> yeah, good point, i guess we’ll just have to write something up and see how it turns out
19:47:33 <etoews> do the proxies always get loaded?
19:48:16 <briancurtin> unless loading changed with ksa, everything gets loaded
19:48:35 <briancurtin> so you’ll always have a conn.telemetry or whatever
19:48:59 <etoews> hmmm...skip warnings for now?
19:49:05 <briancurtin> (further down the line we should probably get smarter and only load what comes back from the service catalog)
19:49:06 <terrylhowe> yes
19:49:25 <terrylhowe> I think there is a ticket for lazy loading
19:49:32 <briancurtin> maybe. they’re probably a little tricky to get right without over-warning people
19:49:37 <terrylhowe> definitely need that]
19:49:38 <etoews> that's my inclination. i truly hate noisy warnings.
19:49:53 <etoews> (that aren't relevant to me)
19:50:21 <etoews> okay. no warnings for now.
19:50:33 <etoews> i'll summarize the beta docstring stuff in that bug.
19:51:29 <etoews> briancurtin: is this still relevant? https://bugs.launchpad.net/python-openstacksdk/+bug/1494494
19:51:29 <openstack> Launchpad bug 1494494 in OpenStack SDK "Expose the full URL to an object in object_store" [Medium,Confirmed]
19:51:54 <briancurtin> yeah, still need to break a hole through something to expose that
19:52:23 <briancurtin> haven’t looked at it since the change though
19:52:29 <etoews> terrylhowe: is this still relevant? https://bugs.launchpad.net/python-openstacksdk/+bug/1461200
19:52:29 <openstack> Launchpad bug 1461200 in OpenStack SDK "Create resource.props type=Resource for attributes with _id names" [Critical,In progress] - Assigned to Terry Howe (thowe-g)
19:53:07 <terrylhowe> that gets into the path args world
19:53:21 <briancurtin> that’s very relevant and i really hope we can get that done so it’s equal across the board, but yeah, the path args thing
19:53:35 <terrylhowe> definitely need something there, I haven’t come up with a better solution than what is out there
19:53:52 <terrylhowe> something simplier would be nice
19:54:23 <briancurtin> i’ll take a look soon at that, would like to get those _id removals rolling
19:56:02 <etoews> briancurtin: is this done? https://bugs.launchpad.net/python-openstacksdk/+bug/1488631
19:56:02 <openstack> Launchpad bug 1488631 in OpenStack SDK "Converge on upload/download names instead of create/save/etc" [Critical,In progress]
19:56:06 <briancurtin> would rather not have to release with the blah_id names, then phase them out via deprecation in favor of blah (i also don’t know how to raise deprecation warnings on attribute access)
19:56:33 <briancurtin> etoews: i don’t think so, unless i pushed the branch and it didn’t have Closes-Bug - i’ll take a look
19:57:04 <terrylhowe> there aren’t a huge number of resources with path args atm
19:57:31 <etoews> terrylhowe: can this be considered done? https://bugs.launchpad.net/python-openstacksdk/+bug/1513908
19:57:31 <openstack> Launchpad bug 1513908 in OpenStack SDK "Move to keystoneauth" [Critical,In progress] - Assigned to Terry Howe (thowe-g)
19:57:33 <briancurtin> maybe in another one of my “audit” things i’ll round up where we have _id suffixed things and work them out
19:57:46 <terrylhowe> yes
19:57:49 * etoews desperately wants to kill at least one bug
19:57:52 <etoews> \o/
19:58:24 <terrylhowe> killed
19:58:37 <etoews> and there was much rejoicing
20:01:07 <briancurtin> #endmeeting