[09:45] * danilos -> breaklunch [12:28] bac benji danilos gmb, hiya and call in 2 or 3 [12:28] ok [12:32] https://answers.launchpad.net/launchpad/+question/172089 [12:44] https://launchpad.net/bugs/847485 [12:44] <_mup_> Bug #847485: process-mail.py crashing with Unicode logging errors < https://launchpad.net/bugs/847485 > [13:00] * gary_poster needs to take kids to school [13:19] ...and we're back. [14:09] gary_poster: let me know when you want to talk about 175545 [14:11] ack benji, hopefully I'll ping in 5 or 10 [14:11] sounds good [14:11] benji, if I don't ping in 30, ping me back please :-) [14:11] k :) [14:48] hi gary_poster, this is your reminder [14:49] benji, hi, thank yo [14:49] will call in a min [14:49] k, need to prepare [14:50] ready when you are [14:56] bug 175545 [14:56] <_mup_> Bug #175545: BugLinkTargetMixin.linkBug() and unlinkBug() should take a user parameter rather than using LaunchBag < https://launchpad.net/bugs/175545 > [15:02] gary_poster: lib/lp/bugs/tests/buglinktarget.txt [15:02] lib/lp/bugs/model/buglinktarget.py [15:05] lib/lp/code/interfaces/branch.py [15:14] hi gmb, it turns out i was wrong wrt attribute syncing on conjoined bug tasks...so i don't think i need to pick your brain after all [15:15] bac: Glad to hear it. As it happens, I hate conjoined bugtasks with a fiery passion :). [15:15] gmb: i'm nearing that camp [15:15] :) [20:37] gary_poster: got a sec for a buildout question? [20:37] sure bac [20:37] i'm trying to review julian's txlongpollfixture branch [20:37] right [20:38] to do that i have to manually build txlongpoll [20:38] he and I talked about that for awhile. Oh? that's not the intent, is it? [20:38] he includes this bit of advice [20:38] You also need to get lp:rabbitfixture, build it and drop the egg in [20:38] ./download-cache/dist in order for txlongpoll to build. You will also possibly [20:38] need to do the same for testresources-0.2.4_r58 until it reaches PyPI. [20:38] so the rabbitfixture part is fine [20:38] i got revno 58 of lp:testresources and built it [20:39] moved it into download-cache/dist/testresource-0.2.4_r58.tar.gz [20:39] (maybe I misunderstand, but we should be using sdists in the download-cache; that doesn't sound like your question though, so I won't worry about it for now) [20:39] so far so good [20:40] yeah but then i get: [20:40] Installing testresources 0.2.4-r58 [20:40] Caused installation of a distribution: [20:40] testresources 0.2.4 [20:40] with a different version. [20:40] Got None. [20:40] While: [20:40] Installing test. [20:40] Error: There is a version conflict. [20:40] We already have: testresources 0.2.4 [20:41] if i don't provide testresource buildout complains that it cannot be found [20:41] if i do it complains as above [20:41] perhaps your comment about sdists is relevant? did i make it incorrectly? [20:43] IMO this should be supplied by Julian in download-cache or for downloads. bac, how did you build the testresources? I suspect you just did "./bin/buildout setup sdist .", or perhaps just "python setup.py sdist", and then you munged the built sdist to have the filename you wanted, right? [20:43] If so, that's not quite right [20:43] you need to use egg_info [20:43] something like this [20:43] yes [20:44] that's exactly what i did [20:44] * gary_poster goes to look at the docs to remind himself [20:44] these packages should have F&#$& makefile targets [20:44] the tx ones do [20:45] bac, try "./bin/buildout setup . egg_info -r sdist" [20:46] bac, oh wait [20:46] that only works for svn [20:46] try this one: [20:46] ./bin/buildout setup . egg_info -b-r58 sdist [20:47] (or python setup.py egg_info -b-58 sdist) [20:48] if it works, it should give you a tar.gz with the name you expect [20:48] without munging [20:48] (the second example was missing an "r": it should have been "python setup.py egg_info -b-r58 sdist" [20:49] ) [20:49] bac ^^^ [20:51] maybe -b_r58? [20:51] the fact that setuptools only supports svn by default is annoying [20:52] setup.py doesn't recognize 'egg_info' as a command [20:53] bac, use the bin/buildout version then [20:53] setup.py must not be setup to support setuptools [20:54] It's too bad Python packaging is such a wasteland. :-/ [20:56] it looks like testresources doesn't use buildout? [20:56] no buildout.cfg [20:57] bac, bah [20:57] bac, you may have to manually muck with setup.py then [20:57] this makes no sense [20:57] testresources uses neither buildout nor setuptools nor distribute apparently [20:57] mortals have built this stuff before [20:58] bac, muck with setup.py locally. [20:58] change the version there [20:58] then run sdist [20:58] and move on [20:58] the mortals who did this before apparently like to muck with their files when they make temporary releases [21:00] that worked [21:00] thank you [21:02] cool bac, sorry for the goose chase. When no one agrees on how to do something, we're left with guessing. :-/ [21:02] yep. [21:10] gary_poster: bug you again? [21:10] http://pastebin.ubuntu.com/695840/ [21:10] what does the extras_require do? [21:11] julian claims that it causes setup.py to pull txlongpoll into the bin/ but it doesn't seem to [21:12] i'm not sure how it would find it... [21:12] ok, enough [21:12] bac sure. that is supposed to mean that when you build the test target of txlongpollfixture, you need txlongpoll. Otherwise, if you are just using it, you shouldn't have a dependency [21:12] where does it expect to find txlongpoll? [21:12] bac, pypi, or similar [21:13] i'm confused by the two seemingly contradictory statements about PATH [21:13] in that paste [21:14] bac, he is trying to say the same thing twice. Let me try: [21:14] (and then I need to go) [21:14] Actually I need to go now bac, sorry [21:14] np [21:14] Ask me Monday :-) [21:14] Have a great weekend [21:14] looks like a NEEDSINFO to me