Wednesday, 2014-11-12

*** etoews has joined #openstack-sdks00:12
*** etoews has quit IRC00:17
*** Viswanath has joined #openstack-sdks00:25
*** Viswanath has quit IRC00:29
*** etoews has joined #openstack-sdks01:12
*** rmcall has quit IRC01:14
*** etoews has quit IRC01:17
*** sigmavirus24_awa is now known as sigmavirus2401:38
*** britthouser has joined #openstack-sdks01:47
*** Viswanath has joined #openstack-sdks01:47
*** Viswanath has quit IRC01:51
*** briancurtin has quit IRC01:56
*** briancurtin has joined #openstack-sdks01:58
*** etoews has joined #openstack-sdks02:13
*** etoews has quit IRC02:18
openstackgerritA change was merged to stackforge/python-openstacksdk: base class for proxy tests  https://review.openstack.org/13244302:19
*** ayoung has joined #openstack-sdks02:29
*** openstackgerrit has quit IRC02:34
*** tellesnobrega_ has joined #openstack-sdks02:46
*** rmcall has joined #openstack-sdks02:47
*** mattfarina has joined #openstack-sdks02:56
*** mattfarina has quit IRC02:57
*** etoews has joined #openstack-sdks03:05
*** etoews has quit IRC03:10
*** rmcall has quit IRC03:16
*** rmcall has joined #openstack-sdks03:17
*** Viswanath has joined #openstack-sdks03:22
briancurtinsigmavirus24: most of the current proxies are going to need a rebase since i changed that base test case. should just be a button click, i think, and not any actual code change03:22
sigmavirus24ah cool.03:22
sigmavirus24I was just going through reviews and noticed that03:22
*** etoews has joined #openstack-sdks03:24
sigmavirus24I dont' see the button on my end so I didn't realize that03:24
*** Viswanath has quit IRC03:25
briancurtinsigmavirus24: yeah it'll only appear for terry, so at first it just looks bad that there's a bunch of red text and the red X, but in this case it's relatively easy to work from (i think/hope)03:27
sigmavirus24I'm happy to help if it's not trivial03:27
*** etoews has quit IRC03:28
briancurtinsigmavirus24: i dont remember the mechanics there, but i think in the end you would be reproducing what terry would have to do anyway, at best. (i think you can manually rebase and push, and he can push a button, or something)03:34
sigmavirus24Yeah, I just mean if Terry doesn't have time to rebase them (and the button doesn't work) I'd be happy to pick up some slack /cc terrylhowe03:35
*** testing has joined #openstack-sdks03:48
*** etoews has joined #openstack-sdks03:56
*** etoews has quit IRC04:00
*** rmcall_ has joined #openstack-sdks04:05
*** rmcall has quit IRC04:05
*** rmcall_ is now known as rmcall04:05
*** sigmavirus24 is now known as sigmavirus24_awa04:19
*** etoews has joined #openstack-sdks04:20
*** etoews has quit IRC04:25
*** subscope has joined #openstack-sdks04:36
*** etoews has joined #openstack-sdks04:39
*** etoews has quit IRC04:43
*** briancurtin has quit IRC04:52
*** stevemar has quit IRC05:14
*** tellesnobrega_ has quit IRC05:23
*** etoews has joined #openstack-sdks05:39
*** etoews has quit IRC05:44
*** rmcall has quit IRC06:11
*** stevemar has joined #openstack-sdks06:27
*** stevemar has quit IRC06:28
*** stevemar has joined #openstack-sdks06:28
*** k4n0 has joined #openstack-sdks06:37
*** etoews has joined #openstack-sdks06:40
*** etoews has quit IRC06:44
*** rmcall has joined #openstack-sdks06:46
*** subscope has quit IRC06:55
*** etoews has joined #openstack-sdks06:56
*** etoews has quit IRC07:00
*** etoews has joined #openstack-sdks07:13
*** etoews has quit IRC07:18
*** rmcall has quit IRC08:14
*** etoews has joined #openstack-sdks08:14
*** etoews has quit IRC08:19
*** rmcall has joined #openstack-sdks08:20
*** rmcall has quit IRC08:30
*** stevemar has quit IRC08:34
*** etoews has joined #openstack-sdks09:15
*** etoews has quit IRC09:19
*** tellesnobrega_ has joined #openstack-sdks10:11
*** tellesnobrega_ has quit IRC10:23
*** tellesnobrega_ has joined #openstack-sdks10:29
*** tellesnobrega_ has quit IRC10:50
*** tellesnobrega_ has joined #openstack-sdks11:20
*** terrylhowe has quit IRC12:18
*** tellesnobrega_ has quit IRC12:56
*** tellesnobrega_ has joined #openstack-sdks13:03
*** briancurtin has joined #openstack-sdks13:21
*** briancurtin has joined #openstack-sdks13:21
*** k4n0 has quit IRC13:46
*** tellesnobrega_ has quit IRC13:47
*** sigmavirus24_awa is now known as sigmavirus2413:54
*** mattfarina has joined #openstack-sdks14:02
*** stevemar has joined #openstack-sdks14:35
*** etoews has joined #openstack-sdks14:37
*** openstackgerrit has joined #openstack-sdks14:40
*** rmcall has joined #openstack-sdks14:41
*** etoews has quit IRC14:41
*** rmcall has quit IRC14:46
*** rmcall has joined #openstack-sdks14:50
*** ycombinator_ has joined #openstack-sdks14:51
*** ycombinator_ has quit IRC14:52
*** rmcall has quit IRC14:55
*** etoews has joined #openstack-sdks15:15
*** etoews has quit IRC15:28
*** ayoung is now known as ayoung-afk15:53
*** etoews has joined #openstack-sdks16:00
*** rmcall has joined #openstack-sdks16:05
*** terrylhowe has joined #openstack-sdks16:19
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: Redo auth plugins  https://review.openstack.org/13180416:32
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: Make some v3 auth data default to the keystone default domain  https://review.openstack.org/12586516:32
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: Add additional support for --or-show  https://review.openstack.org/13399616:32
*** sigmavirus24 is now known as sigmavirus24_awa16:32
openstackgerritBrian Curtin proposed stackforge/python-openstacksdk: Initial "Getting Started" guide  https://review.openstack.org/13388316:34
*** sigmavirus24_awa is now known as sigmavirus2416:34
*** Viswanath has joined #openstack-sdks16:36
*** Viswanath has quit IRC16:39
*** ayoung-afk is now known as ayoung16:57
*** mattfarina has quit IRC17:03
*** mattfarina has joined #openstack-sdks17:04
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: cleanup files that are created for swift functional tests  https://review.openstack.org/13400917:09
*** rmcall has quit IRC18:00
*** rmcall has joined #openstack-sdks18:01
*** subscope has joined #openstack-sdks18:36
briancurtinterrylhowe: if i could steal a few minutes of your time, would you be able to look at https://review.openstack.org/#/c/133883/ so i can push a doc build to readthedocs?18:59
*** rmcall has quit IRC19:10
*** rmcall has joined #openstack-sdks19:12
britthouserAfter seeing the SDK session at summit, I thought I'd take the SDK for a whirl.19:33
britthouserHaving issues on CentOS - anybody got this working on 6.6 or 7.0? or is everyone pretty much using ubuntu?19:34
britthouserProblem I'm having is when I do 'from openstack import connection' I get, "Exception: Versioning for this project requires either an sdist tarball, or access to an upstream git repository. Are you sure that git is installed?"19:34
britthouserand I do have git installed.19:34
Alex_Gaynorbritthouser: how'd you get a copy of the project?19:35
britthouserI did a 'pip install openstack-sdk'19:35
britthousersorry19:35
britthouserpip install python-openstacksdk19:35
Alex_Gaynorbritthouser: hmm, I think there's a bug in how the package was generated19:36
Alex_Gaynorbriancurtin, terrylhowe: It looks like the current package was uploaded manually, can we move to uploading packages through the openstack system? Can stackforge projects do that?19:37
briancurtinAlex_Gaynor: looking at this right now, and the package is looking weird to me, and im not sure how to do them through any openstack system. fwiw, the package that's up there now was somewhat cobbled together between master plus one or two in-review changes, basically what it took to run the examples from our taelk19:38
briancurtini had been running that stuff out of a checkout, but i believe terrylhowe ran it out of the actual installed package19:39
Alex_Gaynorbriancurtin: I don't know what needs to happen to use the openstack system; probabyl mordred could tell us19:41
briancurtinbritthouser: fwiw, yeah i got that versioning/git installed exception now running from the package. weird...19:44
britthouserOk. =)  At least its not me.19:44
britthousernot just me.19:44
sigmavirus24stevemar: for the record, I opened a bug against testtools to see if we could get that feature in19:44
sigmavirus24https://bugs.launchpad.net/testtools/+bug/1392007 is the bug link if you wish to comment. It'd probably be worthwhile to have for more than just the SDK19:45
briancurtinbritthouser: once we figure out the package i'll let you know. i would suggest you try a source checkout in the meantime, but then you'd have to pull a bunch of branches that are currently under review, so it's probably a hassle for right now19:47
Alex_Gaynorbriancurtin: fwiw ``pip install git+https://github.com/stackforge/python-openstacksdk`` gives the same error19:47
Alex_GaynorI don't really understand pbr super well unfortunately19:48
britthouserbriancurtin: Ok I'm not in a huge rush.  If I can save myself some headache by waiting a day or two, I'll probably do that.19:48
stevemarsigmavirus24, lookin'19:48
britthouserbriancurtin: I did want to contribute a one line change to the Readme (git should be a requirement), so maybe I'll walk through that "Getting started guide" and see if I can submit a patch for that in the meantime.19:50
Alex_Gaynorbritthouser: sure, you won't actually need a fully functioning install to send a patch :-)19:51
britthouserOk cool.  Thanks!19:51
Alex_Gaynorlet us know if you've got any questions getting started contributing!19:51
Alex_Gaynor(Though I'm about to get lunch :-))19:52
britthouserI will.  I saw https://review.openstack.org/#/c/133883/ earlier, so I thought I'd walk through that.19:52
Alex_Gaynorsounds good; please leave comments on that if anything is unclear or could be improved!19:53
britthouserWill do. =)19:53
*** subscope has quit IRC20:01
sigmavirus24it's a problem with the line in __init__.py that determines the version, something to do with pbr.version.VersionInfo('openstack').version_string()20:01
sigmavirus24other projects have that in a separate version.py file typically20:02
sigmavirus24Seems like a general PBR issue https://github.com/rackspace/pyrax/issues/450 maybe20:04
mordredAlex_Gaynor: on my phone ... I'll getcha hooked up when I'm on my laptop20:08
*** etoews has quit IRC20:09
mordredalso, versioninfo('openstack') is gonna just flat be wrong, I believe20:10
sigmavirus24Yeah I'm thinking that's the problem20:15
openstackgerritIan Cordasco proposed stackforge/python-openstacksdk: Use project name to retrieve version info  https://review.openstack.org/13404620:16
sigmavirus24^^ fixes it20:17
britthousersigmavirus24: if I manually make that change to my install, the import succeeds.20:20
sigmavirus24Sounds good britthouser20:20
*** etoews has joined #openstack-sdks20:25
stevemarsigmavirus24, re: https://review.openstack.org/#/c/134009/ i was wondering if i should use setupClass and teardownClass, since we only test against py27 and py3320:26
sigmavirus24I thought I saw that we were testing against 2620:26
sigmavirus242nd to last test in the jenkins check table20:26
stevemaroh jeez...20:27
stevemarwhy is that not in order :P20:27
sigmavirus24If we drop that, then we can use setUpClass, tearDownClass20:27
sigmavirus24lol *shrug*20:27
stevemarnah20:27
sigmavirus24stevemar: either way, it's the betterâ„¢ solution for testtools to add this since I can't fathom that no other project needs this20:27
sigmavirus24britthouser: even though that getting started guide was just +W'd you can still submit patchset(s) for any fixes you find20:30
openstackgerritMerged stackforge/python-openstacksdk: Initial "Getting Started" guide  https://review.openstack.org/13388320:31
britthousersigmavirus24: Sure20:33
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: cleanup files that are created for swift functional tests  https://review.openstack.org/13400920:37
Alex_Gaynorsigmavirus24: nice catch, thanks20:49
sigmavirus24HTH20:49
sigmavirus24It seemed odd at first that packaging wouldn't use the package name and when mordred mentioned it I tried it out and it just worked so... pure luck really20:49
openstackgerritMerged stackforge/python-openstacksdk: Use project name to retrieve version info  https://review.openstack.org/13404620:50
mordredsigmavirus24: the other way to accomplish the desire there is via testresources21:03
mordredinstead of setupclass21:03
sigmavirus24I'm not familiar with testresources honestly. I'll have to look into it21:03
mordredI'll be on laptop in about 15 and will be more verbose :)21:04
sigmavirus24take your time21:06
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: Create functional tests for the example code  https://review.openstack.org/13405321:12
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: Create functional tests for the example code  https://review.openstack.org/13405321:15
stevemarsigmavirus24, you caught me not running pep821:16
* stevemar hangs head in shame21:16
sigmavirus24stevemar: I am pep8721:16
sigmavirus24*pep821:16
* sigmavirus24 maintains flake821:16
sigmavirus24Feel free to curse at me liberally =P21:17
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: Create functional tests for the example code  https://review.openstack.org/13405321:17
stevemarsigmavirus24, bah humbug!21:17
openstackgerritTerry Howe proposed stackforge/python-openstacksdk: Minor docs updates to index, installation and usage  https://review.openstack.org/13242321:17
stevemarsigmavirus24, i was too eager anyway, i wanted a few more cosmetic changes to the comments and such21:18
sigmavirus24Fair enough21:19
openstackgerritTerry Howe proposed stackforge/python-openstacksdk: Minor docs updates to index, installation and usage  https://review.openstack.org/13242321:22
openstackgerritTerry Howe proposed stackforge/python-openstacksdk: Minor docs updates to index, installation and usage  https://review.openstack.org/13242321:24
mordredsigmavirus24: so, the general reason people usually use setupClass is as a performance optimization so that you don't have to run setUp and tearDown for expensive fixtures, right?21:24
sigmavirus24mordred: yes21:24
mordredsigmavirus24: the way that works with testresources is that you essentially define a testresource class that holds the expensive resource, then add a fixture to the test class that is an instance of that testresource21:26
mordredthen in the testresource class, if it's expensive, you can override the reset method to not teardown/setup21:26
mordredand to optionally do scrubbing if that makes sense21:27
mordredit's then possible to inform the test scheduler about resources so that in parallel runs things that need an expensive resource can clump onto one of the parallel threads21:27
mordredso that you don't have to make one in each parallel path21:27
sigmavirus24Yeah I was looking at testresources' pypi page (because there don't seem to be any docs anywhere else) and it looks fairly simple to use.21:28
mordredit's a bit more work up front, but gives a bit more flexibility on how you manage those things ... which most of the time are fairly rare anyway, and should be considered21:28
mordredyah21:28
mordredI need to get someone to write some better docs on that21:29
sigmavirus24it just seems like overkill for one small set of tests that need a file21:29
sigmavirus24file creation *could* slow down the tests but it really shouldn't block that patchset21:29
sigmavirus24More so, this resource isn't used in any of the tests. We just create the file and delete it. We're not holding on to a reference in it21:31
sigmavirus24mordred: https://review.openstack.org/#/c/134009/2/functional/tests/test_object.py21:31
sigmavirus24So this feels even more like overkil than not21:32
mordredsigmavirus24: yes. I totally agree21:34
mordredsigmavirus24: also - there isa  tempdir fixture that knows how to cleanup after itself21:34
mordredjust while we're on the subject21:34
sigmavirus24yeah but I think that still creates the file for each test21:34
mordredyah21:34
* mordred just normally tries to avoid tearDown21:35
mordredaddCleanupi is gauranteed to run, tearDown is not21:35
mordredbut I'm nitpicking :)21:35
mordredI agree with the substance of what you are saying21:35
*** Viswanath has joined #openstack-sdks21:36
sigmavirus24yeah, this *should* be doable with testresources though21:38
mordredhttp://paste.openstack.org/show/13255121:38
*** Viswanath has quit IRC21:40
sigmavirus24stevemar: ^^21:40
stevemarsigmavirus24, ah neat21:41
*** Viswanath has joined #openstack-sdks21:41
*** Viswanath has quit IRC21:44
openstackgerritSteve Martinelli proposed openstack/python-openstackclient: cleanup files that are created for swift functional tests  https://review.openstack.org/13400921:45
openstackgerritMerged openstack/python-openstackclient: Use fixtures from keystoneclient for static data  https://review.openstack.org/13225921:54
openstackgerritMerged stackforge/python-openstacksdk: Minor docs updates to index, installation and usage  https://review.openstack.org/13242321:56
*** mattfarina has quit IRC22:01
stevemarsigmavirus24, whats the point of: https://review.openstack.org/#/c/133957/22:15
sigmavirus24stevemar: if I understand the related bug well enough, usage of PYTHONHASHSEED should be removed so that (on the python's that pick it automatically) our tests take advantage of hash randomization22:16
stevemarfair enough i suppose22:17
stevemarsigmavirus24, if you have time, https://review.openstack.org/#/c/131804/ i think dtroyer is still in paris mode :)22:18
sigmavirus24It also appears to be an artifact of prior dependency on hash ordering (which I didn't think would affect us)22:18
sigmavirus24stevemar: I don't have the magical +W powers ;P22:19
stevemarsigmavirus24, i know, but i consider your +1 a very strong  +1 :)22:19
sigmavirus24Ah, thanks. I'll look over it22:19
stevemari really want to cut a 1.0 release, if we can get a few of these last lingering patches, i'll quickly make some release notes and bug the heck out of dea22:20
stevemardean22:20
sigmavirus24heh22:24
jamielennoxstevemar: i love the 1.0 idea, i think you/we should look at the client manager object a bit more first - this is the object that is passed to all external managers and the interface is fairly unstable22:42
stevemarjamielennox, i'm down with that, what else are you thinking? also, you're up too early22:43
jamielennoxstevemar: no, i just think i get a daylight savings advantage22:43
jamielennoxstill 8.40 here22:43
jamielennoxstevemar: the other major one that i've always thought that OSC should do is define an abstract base class and load plugins based on that rather than the current weird module import thing you currently do22:44
*** tellesnobrega_ has joined #openstack-sdks22:44
jamielennoxhowever it might already be a little late to define that one22:44
stevemarjamielennox, you're thinking that can't wait til 1.0?22:48
stevemarerr, til after 1.022:49
sigmavirus24jamielennox: are you in australia?22:49
stevemarsigmavirus24, he sure is22:50
sigmavirus24Yeah. We just learned this with our team. +2 hour overlap in our work schedules22:50
sigmavirus24It's awesome22:50
sigmavirus24also prepare for a bunch of comments on the plugins patchset22:53
jamielennoxsigmavirus24: yea22:55
jamielennoxsigmavirus24: however i'm not in Brisbane and we don't do DST (idiots) so it's only 1 hour for us22:56
jamielennoxstevemar: i think the abc thing can wait22:56
jamielennoxyou have too many people implementing the current interface already to change that last minute - and it should be a fairly simple case of defining a new setuptools entrypoint for the new class and also loading them22:57
sigmavirus24jamielennox: ah, that stinks22:57
jamielennoxsigmavirus24: ah - i'm not in Sydney - am in brisbane, obviously editted that sentence mid-thought22:57
sigmavirus24that said, there was a state like that here that recently decided to actually follow DST now so it's terrible22:57
sigmavirus24Yeah I understood that to mean you were in Brisbane because i read it as a growl notification and it parsed correctly anyway for me22:58
jamielennoxit's mostly just annoying to be different to everyone else22:59
jamielennoxanyway,22:59
sigmavirus24must be convenient to not bother with dst though22:59
* sigmavirus24 missed a bunch of meetings this week because they were all scheduled in local time instead of utc23:00
jamielennoxsigmavirus24: lol, yea - i guess i've only worked on international things in non-dst23:00
sigmavirus24stevemar: just realized (from the fact that you followed me) that we met in SAT during the midcycle meetup at dinner23:02
stevemarsigmavirus24, i thought it was you, there were too many ian's that day23:03
sigmavirus24Fair enough. I'm just terrible at being introduced to people in restaurants because i have kind of terrible hearing23:03
stevemari could see how that doesn't help23:04
openstackgerritMerged openstack/python-openstackclient: Tests work fine with random PYTHONHASHSEED  https://review.openstack.org/13395723:19
*** tellesnobrega_ has quit IRC23:20
*** rmcall has quit IRC23:29
jamielennoxdhellmann: ping re oslo.messaging23:35
*** tellesnobrega_ has joined #openstack-sdks23:35
*** etoews has quit IRC23:40
*** tellesnobrega_ has quit IRC23:45

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