=== Quintasan_ is now known as Quintasan
=== yofel_ is now known as yofel
=== Vorpal_ is now known as Vorpal
=== medberry is now known as med_out
[20:08]  * AuroraBorealis sighs
[20:08] <AuroraBorealis> for some reason when bazaar calls gpg it doesn't call pinentry and it just fails =(
[20:13] <AuroraBorealis> it seems that bzr explorer needs to use --no-tty
[20:13] <AuroraBorealis> for the gpg command
[21:02] <jelmer> AuroraBorealis: please file a bug :)
[21:06] <AuroraBorealis> k
[21:16] <AuroraBorealis> submitted. https://bugs.launchpad.net/bzr-explorer/+bug/847388
[21:16] <ubot5> Ubuntu bug 847388 in Bazaar Explorer ""gpg: cannot open `/dev/tty': No such device or address" on Ubuntu when signing commits" [Undecided,New]
[21:17] <jelmer> AuroraBorealis: thanks
[21:17] <AuroraBorealis> i just commit using the regular terminal command and it works, so i think its just a --no-tty issue
[21:17] <AuroraBorealis> however i dont know where in the code the command gets called so i can't test it xD
[21:18] <jelmer> AuroraBorealis: it's in bzrlib/gpg.py
[21:27] <AuroraBorealis> it seems that the tty environment variable is not set in ubuntu
[21:27] <AuroraBorealis> and thats whats causing it to freak out
[21:28] <AuroraBorealis> not sure if thats bad or not
[21:39] <AuroraBorealis> it also appears that --no-tty fixes it, but i'm unsure of how to integrate that :3
[21:48] <AuroraBorealis> annnnnd i seem to of figured out a solution. but now how to i generate a patch or something :o
[21:49] <jelmer> AuroraBorealis: I'd recommend running the gpg related tests and fixing anything that breaks: ./bzr selftest --no-plugins gpg
[21:50] <AuroraBorealis> after i make my changes?
[22:02] <jelmer> AuroraBorealis: yep
[22:03] <AuroraBorealis> kk
[22:03] <jelmer> AuroraBorealis: well, ideally you should fix the tests to expect what you want to happen and then fix bzrlib.gpg
[22:03] <jelmer> but this works too, and makes more sense given you've already changed bzrlib.gpg.
[22:03] <AuroraBorealis> where are the tests located?
[22:03] <AuroraBorealis> and this is kinda hard to test as it requires bazaar explorer
[22:04] <AuroraBorealis> (as os.environ("TTY") has to not be set)
[22:04] <jelmer> AuroraBorealis: this shouldn't require bzr explorer - I just meant tests that make sure that --no-tty is specified
[22:06] <AuroraBorealis> yeah, i'm assuming these are unit tests. but the way to check that --no-tty is only specified correctly is to be using it where the environment variable for tty is not set, like in a GUI
[22:06] <AuroraBorealis> unless i'm misunderstanding something
[22:11] <jelmer> AuroraBorealis: sure, but in the unit test you can override the environment variable and see if the right parameters are specified
[22:11] <AuroraBorealis> ah ok.
[22:11] <jelmer> AuroraBorealis: see bzrlib.tests.TestCase.overrideEnv
[22:22] <AuroraBorealis> i cant seem to run this bazaar branch i checked out from launchpad because it can't import "shlex_split_unicode"
[22:26] <AuroraBorealis> any idea how to fix that jelmer so i can run the tests? xD
[22:35] <Noldorin_> hi jelmer
[22:39] <jelmer> hi Noldorin_
[22:39] <jelmer> AuroraBorealis: I'm not sure - that should only be necessary on Windows I think
[22:39] <AuroraBorealis> well the test cases are working i guess
[22:39] <AuroraBorealis> so i'm trying to figure how how i borked those :<
[22:46] <Noldorin_> jelmer, i am finding many bzr bugs these days...
[22:46] <Noldorin_> while researching this issue
[22:47] <jelmer> Noldorin_: what exactly?
[22:47] <Noldorin_> jelmer, for a start, https://bugs.launchpad.net/bugs/846122
[22:47] <ubot5> Ubuntu bug 846122 in Bazaar "bzr thinks working tree is out of date after uncommit" [Undecided,Confirmed]
[22:48] <jelmer> Noldorin_: is the history empty after that operation perhaps?
[22:48] <Noldorin_> what history?
[22:48] <jelmer> Noldorin_: what's the output of "bzr revno"
[22:49] <Noldorin_> jelmer, 46
[22:49] <jelmer> Noldorin_: that's odd - is this in a bzr-git tree?
[22:50] <Noldorin_> nope
[22:50] <jelmer> Noldorin_: bound branch?
[22:51] <Noldorin_> jelmer, maybe
[22:51] <Noldorin_> i forget
[22:57] <AuroraBorealis> is there a way to use 'print' when you are using these test cases? it seems that its not actually printing them (and causing other tests to fail)
[23:03] <jelmer> AuroraBorealis: I generally use bzrlibtrace.mutter
[23:03] <jelmer> AuroraBorealis: I generally use bzrlib.trace.mutter
[23:04] <jelmer> AuroraBorealis: that will be printed as part of the test output if a test fails
[23:05] <AuroraBorealis> also another question
[23:05] <AuroraBorealis> bzrlib.TestCase.overrideEnv says that the environment variable will be reset after each test
[23:06] <AuroraBorealis> is a test the entire class that extends TestCase or are they the individual methods?
[23:06] <lifeless> the method
[23:07] <lifeless> the class is a way to group related tests and (sometimes) test helpers that are specific to that group
[23:07] <AuroraBorealis> ok. so i guess i have no idea why the environment variable for 'tty' would be none
[23:12] <AuroraBorealis> as i thought it would be None if you were running like a user interface, but its none even if you invoke bzr from a command line
[23:12] <AuroraBorealis> so i have no idea how to distinguish when you should add the --no-tty switch or not
[23:37] <AuroraBorealis> well, it appears that just adding --no-tty works, so how do i submit this as a patch to the bug report i opened?