=== gary_poster is now known as gary_poster|away === gary_poster|away is now known as gary_poster [13:13] hi benji, gary_poster: i'm trying to use pdb to debug charmworldlib but it is terribly confused about where to find source files. it tries to find them in /home/bac/projects/charm-tools/build/bdist.linux-x86_64/egg/charmworldlib/bundle.py which does not exist. (the build directory is not present.) any ideas on how to get it sane? === BradCrittenden is now known as bac [13:14] on call [13:14] say hello for me [13:14] bac: the problem might be that the .pyc files were built in one location and then moved, if so, removing all the .pyc files will get them regenerated with correct paths [13:15] benji: tried that first. 'make clean' removes all pyc files from the tree and then i rebuilt. same same. [13:15] benji: if i manually import charmworldlib and then look at __file__ it is in a reasonable spot [13:16] bac: I don't know for sure, but if make is getting them moved, then it will just recreate the problem; I'd to an 'rm **/*.pyc' (or equiv in your shell) [13:16] I'm curious why it is doing such a strange thing. [13:17] benji: i'm not sure what you mean. 'make clean' removes them. i then verify they are gone. are you saying manually remove them after building so the debugger has to recompile them as they are encountered? [13:17] bac: exactly [13:17] ok, will try that [13:23] benji: doesn't work so well when accidentally removing *.py [13:23] heh :( [13:26] Makyo, hatch, could you please apply local pressure to hazmat to have him merge https://code.launchpad.net/~hazmat/juju-deployer/refactor-placement-and-validate-feedback/+merge/195903 ? veiled threats welcome ;-) thank you [13:26] :) can do [13:26] Hah. otp w/ sabdfl, after [13:27] ok thanks :-) [13:29] rick_h_: AIUI you are not around and we should resched. Lemme know if I'm wrong [13:54] gary_poster: I'm around now [13:54] gary_poster: so any time you're free [14:00] ack thanks rick_h_. Probably will be 11:30-ish [14:00] gary_poster: rgr [14:01] frankban: I figured out that wierd selenium error and the fix was just to upgrade, will you review the trivial branch for me? https://codereview.appspot.com/33230043 [14:02] benji: sure, on it in a minute [14:06] gary_poster I never got the email in question so Makyo is forwarding it to me [14:06] benji: the branch looks good, it's weird that the I see the MP linked to an old one (Use a random service name for tests.) [14:06] hrm, that is odd [14:07] benji: but the changes are pertinent to the selenium problem, so that's just an lbox oddity I guess [14:08] benji: or maybe you used the same "test-failure" branch name [14:08] frankban: oh, I did use the same branch name [14:10] benji: are you going to create another MP? [14:11] frankban: I guess so [14:11] benji: ok, I'll LGTM there then [14:35] benji: can we chat real quick? [14:36] bac: sure [14:36] bac: how about the team meeting hangout? [14:36] sure [14:37] bac: https://plus.google.com/hangouts/_/calendar/Z2FyeS5wb3N0ZXJAY2Fub25pY2FsLmNvbQ.j0rk5d371ph8331ijtf48t2uj0 [14:43] gary_poster: friend of mine found a talk to get me interested in clojure hah https://www.youtube.com/watch?v=ShEez0JkOFw [14:43] :-) cool [14:52] cool I'll check it out as well [14:53] I've been looking for a lispy language to spend some time on === kirkland` is now known as kirkland [15:35] rick_h_ testing some CI stuff? [15:36] hatch: yea [15:36] ok just fyi there will be test failures but that's because the tests are bad [15:36] hatch: supposedly the github integration will allow someone on the team (juju org) to trigger tests for pull requests from someone not in the org [15:36] hatch: that's fine, just curious. Was looking at the repo and saw the pull request as a chance to see it it'd work [15:36] hi marcoceppi i'm working on bug 1263112 for charmworld regarding bundle proofing and inheritance. i see you made some changes in that area to charm-tools 1.2 which may solve the bug. are you going to be releasing 1.2 soon? [15:36] <_mup_> Bug #1263112: charmworld proof and ingest does not properly support inheriting bundles [15:38] rick_h_ oh yeah cool np [15:39] gary_poster: did you see the guy behind lighttable is from chapel hill, now seattle? [15:50] jujugui call in 10 [15:53] bac: I'll be cutting a release this week [15:53] bac: but I think that already landed [15:56] did we ever change charm lint/best practice rules to accomodate gui not supporting '.' in config keys? [15:57] bac ah, it didn't make the 1.2.5 release, that will be in 1.2.6 on Friday [15:58] hazmat: I don't see it in the config.yaml lint rules, is that a "critical" thing? [15:58] marcoceppi: i don't see any charm-tools from 1.2 on pypi [15:59] jujugui call in 2 [15:59] bac: oh, I haven't been pushing to pypi for the 1.2.X releases. I'll rectify that. Either way the code hasn't landed yet [15:59] bac: I forgot you guys don't use the package, I'll make sure that finds it's way back in to my release process [16:00] marcoceppi: ok. would like to get us on 1.2 any way. thanks. [16:00] bac: uploading the old releases atm [16:00] or, trying to [16:00] hey marcoceppi i also noticed the charmworldlib package on pypi has .pyc files in it, which really mess up debugging. i think it is Not Good to include pyc in the package. could you take a look at that? [16:01] bac: can do [16:01] cool, thanks [16:01] marcoceppi, dunno, last i checked the gui had issues with that , but perhaps its been fixed. [16:02] yeah.. still has issues with it afaics [16:03] hazmat: I'll add it to the proof then, we should update the docs with this too - does juju have a problem with it? [16:03] its fairly common for some of the java charms [16:03] marcoceppi, no.. just gui [16:03] hazmat: I'm not sure if we should consider the GUI rendering an ERROR or a WARN - I mean it's pretty much integral at this point, right? [16:04] gary_poster: ping? [16:04] yeah sorry [16:04] trying to finish up [16:04] np [16:04] marcoceppi, prolly err, gui is an essential part of the ecosystem [16:04] hazmat: ack, I'll update proof and let jcastro know [16:04] marcoceppi, thanks [16:06] sorry sprry [16:14] bac: 1.2.5 (latest) uploaded to pypi, 1.2.6 to land friday [16:15] marcoceppi: thanks! [16:15] gary_poster: i replied to the email as requested and will remove my face from the card [16:15] :-) ty [16:20] bac: I'm trying to re-package charmworldlib for you, but it still includes .pyc for __init__ and version which are parsed at setup.py time. Is there something I can do to explicitly not allow .pyc files? [16:21] marcoceppi: unsure. benji, thoughts? [16:24] marcoceppi: i've got to step out for a bit but can help in an hour or so. [16:25] bac: np, thanks for everything so far [16:25] bac and marcoceppi: I think there is a way to do that with some sort of inventory file that includes/excludes files but I don't know the details off the top of my head. [16:26] it's somewhat bad form to import from the package itself in the setup.py because it can easily become impossible to do so (i.e., a compliled component isn't available at setup time) [16:26] the version.pyc shouldn't be an issue [16:27] benji: thanks, i've added a make clean rule to the project and will press a 0.3.1.1 that has a cleaner sdist [16:30] marcoceppi: http://stackoverflow.com/questions/8880661/getting-package-version-using-pkg-resources for versioning [16:30] if it reads something it should just be an empty text file [16:31] marcoceppi: and the version only exists in the setup.py file [16:31] * rick_h_ hasn't looked at the file in question and just going off irc conversation [16:31] rick_h_: so, can I then, using this, read the version from within the module code? [16:31] rick_h_: https://plus.google.com/hangouts/_/calendar/Z2FyeS5wb3N0ZXJAY2Fub25pY2FsLmNvbQ.j0rk5d371ph8331ijtf48t2uj0 when you are ready. no rush [16:31] rick_h_: since help and -v both spit out the software version for the python project [16:32] marcoceppi: yes [16:33] rick_h_: awesome, thank you. I'll update the projects to use that instead! [16:34] marcoceppi: https://github.com/bookieio/breadability/blob/master/breadability/__init__.py for example [16:34] marcoceppi: that's in the package __init__.py and reads from setup.py [16:36] rick_h_: <3 thanks [16:44] bac: neither project will have pyc files in the sdist going forward [16:44] bac: 0.1.4 fo charmworldlib uploaded to pypi with no pyc, charm-tools 1.2.6 will be clean [16:55] https://github.com/juju/juju-gui/pull/63 jujugui quick review/qa [16:57] Makyo: looks good [16:57] benji: cheers, thanks. [17:56] marcoceppi: hey that all sounds great. thanks! [18:33] Did someone put a bitcoin miner in the charm functional tests? These things take forever. [19:00] hey gary_poster [19:00] what's the status of quickstart landing in the proper juju ppa? [19:01] jcastro: isn't it there now? [19:02] gary_poster, oh I see, it just needs to be built for trusty [19:03] jcastro: ack. juju itself doesn't really work for trusty yet, does it? [19:03] it does, you just need to sudo for juju status [19:03] oh ok [19:03] but thumper was going to bring it up at their meeting to start testing trusty [19:04] so I figure, next time I see frankban just ask him to push a trusty build into the ppa? [19:04] for quickstart I mean [19:04] jcastro: yeah +1. Do you have any urgency? [19:04] not really [19:04] just pointing it out [19:05] :-) cool [19:48] gary_poster, hey so, the changelog formatting still looks weird, but it's readable now [19:48] so I guess good enough? [19:48] heh [20:02] hey marcoceppi it looks like you made changes to charmworldlib but didn't push them to launchpad. could you do that please? [20:02] bac: sorry about that, committed and pushed [20:03] cool, thanks. [20:13] gary_poster: I sent my MSDN email [20:16] gary_poster: also, I entered a swap day (20th -> 16th) and a vacation day (17th) [20:34] benji, did you have an days held over from 2013? [20:34] any [20:35] gary_poster: yep, 4; this is the first of them I'm using [20:36] benji, cool. did you see my email about needing to schedule them very soon? [20:36] I'm heading out to my appoitment now. Later. [20:36] ok bye [20:36] gary_poster: I understood "very soon" to be in the first quarter. Is there a different definition I should be aware of? [20:37] benji, Jan [20:37] benji, should be used before March 31 [20:37] scheduled before end of Jan [20:37] make sense? [20:47] hey gary_poster [20:47] so quickstart needs the bundle to be "bundles.yaml" [20:47] but we export export.yaml [20:47] and it also errors out because we include the -gui in the bundle, and asks me to remove it from the bundle [20:47] should I file these as bugs? [20:48] I was thinking maybe have the gui export "bundles.yaml", and not include itself in the export? [20:48] jcastro: does it do that on comingsoon? [20:49] jcastro: I thought we had a bug and fixed it for including gui itself [20:49] though maybe that's not in a release yet [20:50] checking [20:50] jcastro: hmm, https://bugs.launchpad.net/juju-gui/+bug/1249039 is the bug [20:50] <_mup_> Bug #1249039: Exporting from real environment exports juju-gui as well

[20:51] so guess it's still an issue, jujst low on the -gui and high on quickstart [20:51] rick_h_, nice! It comingsoon does the right thing [20:51] jcastro: ah, ok cool then [20:51] ok [20:52] so ... is making it so it export bundles.yaml a gui bug or should the bug be "quickstart/deployer should be able to deploy a yaml file, no matter what I name it" [20:52] the 2nd one seems more correct to me [20:52] jcastro: yea, same to me. frankban is EOD but would ask him about it tomorrow. If there's any history for the mandate to make it so [20:53] * jcastro nods [20:53] yeah so ... I just quickstarted in AWS and dragged over the brutal mongodb thing I am working on [20:53] it's pretty much awesome [21:47] yay charm tests passed [21:47] yay [21:58] Morning [23:24] morning huwshimi. dinner not served here yet. mind if we start talking at the top of the hour? [23:25] gary_poster: Sure! [23:25] cool thanks. talk to you then