[12:27] <oSoMoN> hi all
[12:28] <oSoMoN> I had a libreoffice build going in a PPA of mine, which I saw failing towards the end of the build due to a unit test failure, and when I went back to it to get the full log I saw that the build had just been restarted under my feet
[12:28] <oSoMoN> how is that possible?
[16:14] <cjwatson> oSoMoN: Can happen if your build crashes the builder.
[16:15] <oSoMoN> cjwatson, I guess in that case there's no way to recover the build log until the builder crashed, right?
[16:15] <cjwatson> Correct.
[16:15] <oSoMoN> right, so I guess it makes sense to trigger a new build
[16:16] <oSoMoN> it's kind of a pain for libreoffice and similar packages that take forever to build though
[16:16] <cjwatson> Such is life, I'm afraid.
[16:16] <oSoMoN> fair enough
[16:16] <oSoMoN> thanks for the explanation!
[16:17] <cjwatson> (For completeness, it can also happen if LP loses connectivity to the builder for long enough, or some other similar issue.)
[16:59] <rbasak> I'm expecting lp.git_repositories.getByPath to be callable, but instead I get a KeyError: http://paste.ubuntu.com/25904265/
[16:59] <rbasak> What am I missing?
[17:00] <rbasak> Using python3-launchpadlib    1.10.3-3ubuntu0.1
[17:00] <rbasak> On Xenial.
[17:00] <rbasak> len(lp.git_repositories) is 0, and next(iter()) on it gives me an immediate StopIteration.
[17:01] <rbasak> len(lp.distributions) says 37, so presumably something works.
[17:02] <rbasak> My paste is wrong. It was "import launchpadlib.launchpad" not just "import launchpadlib".
[17:03] <rbasak> (but I still have the problem)
[17:23] <rbasak> I think I figured it out.
[17:24] <rbasak> version='devel'. Not just 'devel'.
[17:24] <rbasak> nacc: ^
[17:24] <rbasak> (FYI)
[17:37] <nacc> rbasak: ah sure
[18:10] <cjwatson> Yep.  (Also, the reason len(lp.git_repositories) is 0 is that iterating over all Git repositories in LP isn't a thing we consider sensible to do on the webservice, but the collection has to exist as a place for all-Git-repositories-related methods to live.)
[18:13] <nacc> cjwatson: is there a sensible way for me to go from a user or team to all git repositories for that user/team? Or do I need to do an iteration over all the git repositories? In essence, I want to iterate the list of current usd-import-team git repositories so I can see what source packages are already imported and determine what their default git repo s
[18:20] <cjwatson> We could export something for that if needed, although we might want to come up with something more targeted; iterating over 20k+ or whatever entries isn't going to be particularly quick.
[18:21] <nacc> rbasak: --^ maybe we're thinking about it wrong, though; given that we have the phasing constants, whitelist and blacklist available, I can figure out the list of source packages we 'expect' to ahve imported currently (in theory, this is a one-time operation anyways) and we can iterate any of those with usd-import-team repositories?
[22:16] <Saviq> is it possible with launchpadlib to find if a person is an indirect member of a team? I can only see https://launchpad.net/+apidoc/devel.html#team_membership mentions a ParticipationTable, but can't find it anywhere else
[22:20] <Saviq> also found mention of "IPerson.inTeam", but bug #377579 :/
[22:23] <Saviq> or do we have a better way to check if someone's signed the CLA, or is employed by Canonical, to check if their contributions can be appropriated correctly? maybe a webservice of some kind?
[22:25] <cjwatson> The best way I can think of OTTOMH (inefficient but I believe it works) is to get the target team and iterate over its .participants
[22:29] <cjwatson> I'm not totally sure why the answer to that bug isn't to just put the method on the team rather than the participant, though
[22:34] <cjwatson> I guess that wouldn't help with private subteams