[03:23] <ajmitch> hads: I'd like to go, maybe I'll have to see if I can stay at my sister's place :)
[03:24]  * ibeardslee just had the word from above .. it'd have to be under my own steam
[03:24] <ajmitch> shame
[03:25] <ibeardslee> aye
[03:25] <ajmitch> ticket+accomodation+travel+days off work tends to add up a bit
[03:26]  * ibeardslee really needs to find more interesting things to talk about
[03:28] <ojwb> possible tip - osdc last year was in canberra and from wellington i found it was much cheaper (and actually quicker) to fly to sydney and get the bus which runs from sydney airport to central canberra
[03:28] <ojwb> i saw something suggesting there might be direct flights by lca though
[03:29] <ojwb> which would at least make that a slower option
[03:32] <ajmitch> ojwb: yeah I've taken the bus from sydney airport before, it often works out pretty cheap
[03:33] <ojwb> the one annoying feature is it's cheaper if you say exactly which bus you'll be on, which is a bit hard to predict with customs, etc
[03:33] <ajmitch> unless you want to wait around for some time
[03:33]  * ajmitch has always found australian customs to be fairly quick
[03:34] <ojwb> yeah, usually isn't bad, but it'd be annoying to have to buy a whole new ticket
[03:35] <ajmitch> do you have a NZ passport?
[03:36] <ajmitch> in the last few years it's been made quicker with the automation, you don't really need to talk to a customs person
[03:36] <ojwb> no, uk one
[03:36] <ojwb> not quite been here long enough to get an nz one
[03:45] <hads> It does add up, ticket price is only a moderate percentage.
[03:45] <hads> I've missed the last two though so wasn't missing a third.
[05:35] <G> ajmitch: Customs is only quicker etc, if you have absolutely nothing to declare, living rurally throws that out of the water so to speak
[05:37] <G> 30-60 minutes in customs hall at Brisbane last time (includes the queue when I arrived in school holidays morning), vs straight through the MAF checkpoint at Auckland at 11:30pm with quite a few others
[05:59] <ojwb> MAF seemed very slack coming in a few months back
[05:59] <ojwb> no interest in walking boots at all
[08:39] <G> ojwb: I have my entry to Australia routine down good "yeah, I live on a farm, but nothing I've got in my bag has been on the farmland, only the driveway" they umm and ahhh for a minute then decide to just put me through the dog queue
[19:13] <ibeardslee> morning
[19:20] <Atamira> morning
[19:20] <Atamira> raining up here
[19:24] <ibeardslee> supposed to be raining here this morning
[19:24] <ibeardslee> haven't seen it yet
[19:28] <ojwb> morning
[19:39] <mwhudson> morning
[19:48] <ajmitch> morning
[20:25] <chilts> morning
[20:27] <ojwb> every time I go to report a bug in ubuntu I find it harder to find the link
[20:28] <ojwb> aha - https://bugs.launchpad.net/ubuntu/+filebug/?no-redirect
[23:19] <thomi> lifeless: you around? Got a minute to answer some questions about testtools & unicode?
[23:21] <thomi> I'm trying to figure out what I have to do to make testtools work with an assertion message that contains unicode characters, currently I get this: http://pastebin.ubuntu.com/1259064/
[23:22] <lifeless> thomi: ok, so thats not testools thats blowing up :)
[23:22] <lifeless> UnicodeDecodeError: 'ascii' codec can't decode byte 0xa1 in position 425: ordinal not in range(128)
[23:23] <lifeless> note that its a *decode* error,
[23:23] <lifeless> but the line of code referenced is doing an *encode* operation.
[23:23] <thomi> ... huh
[23:24] <lifeless> thomi: you have a bytestring (type(foo) is str), which is non-ascii, and is being passed in to a thing that is expecting a string (type(foo) is unicode))
[23:24] <thomi> hmmm
[23:24] <lifeless> thomi: I bet you that if you put a debug statement there, or run under pdb or something
[23:24] <lifeless>  you'll find text is a str, not unicode
[23:25] <thomi> hmmmm
[23:25] <lifeless> thomi: now, testtools can deal with arbitrary binary data, but you're in the text attachment codepath
[23:25] <thomi> ahhhh
[23:25] <lifeless> because backtraces are strings, not bytestrings (normally)
[23:25] <thomi> the penny drops!
[23:25] <lifeless> thomi: note that if you do this: raise MyException(u"foo\x8")
[23:25] <lifeless> under python2.x
[23:26] <lifeless> you'll fuck things up royally
[23:26] <thomi> yeah, I'm attaching content using text_content
[23:26] <thomi> ...which, n this test, will have odd unicode characters in it, and I'm using bytestrings :(
[23:26] <lifeless> because python2 's __unicode__ and __str__ for Exception is broken.
[23:26] <lifeless> thomi: so, either attach them as non-text
[23:27] <lifeless> thomi: or attach them as text_content with actual strings, not bytestrings
[23:27] <thomi> this is python 2, so my understanding is that there's just str and unicode, right?
[23:28] <lifeless> effectively
[23:29] <lifeless> there is a bytes type in 2.7, but its just an alias to str
[23:29] <lifeless> so its not terribly useful in catching confused types early.
[23:29] <lifeless> thomi: can you paste the attaching code?
[23:30] <thomi> sure, I was using this:
[23:30] <thomi> stdout, stderr = process.communicate()
[23:30] <thomi> self.addDetail('stdout', text_content(stdout))
[23:30] <thomi> where 'process' is a subprocess.Popen object
[23:47] <lifeless> ah yes, so thats not going to work so well :)
[23:47] <lifeless> as it could be anything ;>
[23:48] <lifeless> you can either decode stdout
[23:48] <lifeless> if you know, for instance, that its utf8
[23:48] <lifeless> stdout = stdout.decode('utf8')
[23:48] <lifeless> then it will be text that you're adding
[23:49] <lifeless> or you can use the lower level layers to add what you've got with a more conservative encoding (e.g. cp1280 or whatever it is that accepts 256 bytes as legitimate, never crashes.
[23:50] <lifeless> thomi: e.g.
[23:52] <lifeless> self.addDetail('stdout', Content(ContentType('text', 'plain', {'charset': 'iso-8859-1'}), lambda:[stdout]))
[23:55] <thomi> lifeless: cool, that works - thanks :)
[23:56] <lifeless> thomi: np