[11:04] okay, ephemeral disk setup is pretty broken on Azure. stuff from fstab is mounted way before cloud-init-config.service, which is needed to re-create the filesystem === dmsimard is now known as dmsimard|afk === Toger_ is now known as Toger === contrapumpkin is now known as copumpkin [18:06] harlowja, https://code.launchpad.net/~harlowja/cloud-init/cloud-init-net-refactor/+merge/293957 [18:07] thats what i owe you right ? [18:13] harlowja, why do you randomly change things some times. tox.ini at that mp [18:33] let me see , i think that smoser was also fix the 26 stuff, there isn't much of a need to have a requirements.txt file and a seperate list of requirements for 26 [18:33] seeing as thats the point of a requirements file ? [18:33] k [18:34] we had listed the envs specifically as trusty (i think) does not support the newer tox [18:34] that gleans the python version from the testenv name [18:34] testenv name still there [18:34] just not a special list of deps [18:36] you yanked [18:36] [testenv:py3] [18:36] basepython = python3 [18:36] and you can't do that i dont hink. [18:37] ok, that doesn't seem like it will hurt to put it back in, though that it was inferred [18:38] *thought [18:38] only in newer tox [18:38] k [18:46] ok dokie smoser cleaned that up [19:10] harlowja, how did you drop if PY26 in tests/unittests/helpers.py ? [19:11] the TestCase define [19:11] oh. i see . use unittest2 [19:11] ya, unittest2 should handle all we need here [19:11] unless u want me to undo that? [19:11] but damn was that testcase nasty [19:11] lol [19:11] (custom testcase) [19:11] lol [19:12] wonder who wrote that garbage [19:12] :-/ [19:12] not it [19:12] lol [19:12] me of 2 years ago? [19:12] it was probably 2014 harlow. man... *that* guy. [19:12] lol [19:12] ya, what a jerk [19:12] that old me [19:12] lol [19:12] the new jerk is much cooler [19:12] lol [19:13] so to run test, we'l need python-unittest2 [19:13] right ? [19:13] right [19:13] should be everywhere, its not a new thing [19:14] ./packages/bddeb is busted then. i'll fix that. [19:14] i suspect that bdrpm is probably busted too [19:16] do you actually need unittest2 on python3 ? [19:16] i guess so. [19:18] no, prob not, but then we need a pip that can understand python version constraints :-P [19:19] ? [19:21] there is a way to do unittest2 ; python_version < 2.6 [19:21] but needs newer pip to understand that [19:23] oh. that wasnt what i was askin g i dont htink [19:24] is unittest2 what you get in python3's unittest ? [19:24] and http://paste.ubuntu.com/17182223/ is what i have in my diff on your tree right now. [19:24] the last 2 hunks just cause you'd done that in some tests too [19:25] smoser yes, unittest2 is pretty much a backport lib [19:25] https://pypi.python.org/pypi/unittest2 [19:25] but it should work on ' 2.6, 2.7, 3.2, 3.3, 3.4 and pypy.' [19:25] alot of these backport libs also internally just use the newer stuff on versions where they aren't needed [19:25] so it should be ok [19:26] thats fine [19:29] http://paste.ubuntu.com/17182223/ looks ok, want me to add, or u? [19:30] you can please [19:30] yes sir [19:30] generally speaking i think i want to avoid 'six' if i can at all do that. [19:30] in the net/ module [19:31] as i want it to be esxportable to curtin where we do not have a ddependency on six [19:31] next change, next change :-P [19:31] six though should be ummm, pretty normal [19:31] i'd wonder why six is that big of a deal :-P [19:32] otherwise u just make mini-six [19:32] which i've seen alot, lol [19:32] (in other libs/...) [19:41] well, heres the thing. [19:42] curtin runs in an ubuntu cloud iamge back to precise (12.04) [19:42] six in 12.04 was universe. [19:42] https://launchpad.net/ubuntu/+source/six [19:42] this i think would be the first usage of universe package by curtin [19:43] let me see for sure on that.. but even adding any dependency notin the image is less than ideal [19:48] rharper, so the reason i bothered you in #curtin was usage of six in cloud-init/net [19:48] https://server-team-jenkins.canonical.com/job/curtin-vmtest-venonat/266/artifact/output/PreciseHWETBcacheBasic/logs/install-serial.log [19:48] that is a precise install [19:49] y [19:49] http://paste.ubuntu.com/17183262/ [19:49] that is the interesting piece for this conversation [19:49] right,. what's not in precise by default [19:49] currently we are getting some things from universe [19:50] gdisk and bcache-tools [19:50] adding a dependency on six would be one more [19:50] which is not really the end of the world. [19:50] but precise is the only image supported that does not have python-six (or python3-six). [19:51] curtin is six free, do we have a lot more use of six in cloud-init already ? I mean what's the impact ? [19:55] well, if cloudinit/net adds six and cloudinit/net is used in curtin [19:55] then curtin gains six dependency [20:15] harlowja, ok. i think its sane [20:16] i'm still not sold on six. as we're using it for a *very* minisix [20:16] but sure [20:20] ya, its either that or for py2/py3 u make something minisix [20:21] either can be done, i've seen both, its not anything imposssible/that hard [20:21] just less crap to be written [20:21] lol [20:21] and i like less crap [20:22] right [20:29] harlowja, 'skip_first_boot' [20:29] ya [20:29] that's for testing [20:30] in that during testing and getting data from say a local file, don't mess with my networking, lol [20:31] although i can just change it to mock out 'on_first_boot' there [20:32] i think i'd prefer that. it just looks odd as that function signature is different [20:32] kk [20:32] i do understand the desire to not break your networking :) [20:32] :-P [20:33] ya, test starts messing with my networking == bad [20:33] lol [20:33] and grab http://paste.ubuntu.com/17185178/ [20:33] and i'm happy i think [20:35] i hit approve, and commented. go ahead and do those things an then you can pull to trunk [20:35] thank you josh. [20:36] no thank you [20:36] lol [20:36] smoser is the coolest person ever [20:36] lol [20:39] e-ver [20:39] e-vah! [20:42] amen [20:42] ? [20:42] lol [20:43] oh evar [20:43] got ya [20:46] =) [20:49] also smoser another one u might be interested in (when running on a mac) [20:49] https://gist.github.com/harlowja/b7bbf4f18057b3668f16bbf49ccffe26 [20:49] probably something small off there [20:50] did i just regress taht ? [20:50] unsure [20:50] what is this /Users dir [20:50] i dont have one of those [20:50] mac? [20:50] nor a /var/folders [20:50] i think you computer is broken [20:50] maybe install Ubuntu [20:50] mac [20:50] lol [20:50] mac mac [20:50] mac [20:51] not that big of a deal, mac [20:51] lol [20:51] you'll be cool, mac [20:56] i would be pretty cool if i had a mac and a handlebar mustache [20:56] lol [20:56] Odd_Bloke, ^ would be nice if you could fix that test [20:57] not high priority (obviously) [20:57] damn macs [20:57] lol [20:57] i'm not really sure what test_generate_certificate_uses_tmpdir is tryin gto prove [21:16] man i keep on doing `git diff` in the cloudinit dir [21:16] smoser that's gonna be fixed over the weekend right? [21:16] lol [21:20] other ones that happen on centos i think are already known https://gist.github.com/harlowja/aa2b506c069fe874a51a774cd65a745c [21:21] smoser ok, merging that refactor in [22:24] ok, https://code.launchpad.net/~harlowja/cloud-init/cloud-init-net-sysconfig/+merge/297115 [22:25] eyes on that would be cool [22:25] or new tests [22:53] rharper ^