[08:03] <JamesTait> Good morning all! :)
[11:37] <mandel> ralsina, ping!
[11:37]  * mandel is back :)
[13:08]  * mandel lunch
[13:37] <daveonearth> Ola Ubuntu Oners :)
[13:38] <daveonearth> Just saw this feature in dropbox
[13:38] <daveonearth> http://www.jotform.com/dropbox/
[13:38] <daveonearth> And Identified it as a brilliant and must have for UbuntuOne
[13:40] <dobey> daveonearth: tell the jotform developers to get on it? :)
[13:44] <daveonearth> dobey, I am just about to get in touch with them :) I also thought it might be a great task to add directly in UbuntuOne?
[13:45] <dobey> daveonearth: i don't see any way to do that, which doesn't involve building a web forms service
[14:16] <daveonearth> Dobey my suggestion is only that U1 offer a URL to it's user that can be cloaked and has a form to upload to a dir in U1 folder?
[14:20] <dobey> daveonearth: any web service could use the REST API to store files in a user's U1 files account
[14:21] <dobey> we don't have fine-grained access control though, so any service you authorize to access your files can access all your files
[15:04] <mmcc> hi folks. quiet backlog! no standup again today?
[15:22] <dobey> mmcc: indeed; also, you've only be on-line for 10 minutes :)
[15:22] <dobey> there's only 4 of us working today
[15:30] <mmcc> heh, I was looking at the web irclogs too, but I guess that has some lag :)
[15:31] <dobey> ah
[15:44] <dobey> ok, must get lunch. bbiab
[16:15] <mmcc> brb, coffee
[16:51] <rye> dobey: u1.to has the upload facility... when it works
[16:52] <rye> does anybody know how to get ubuntuone pairing record on 12.04 for couchdb?
[16:57] <dobey> rye: "don't" but also you can install desktopcouch-ubuntuone though it might be broken
[17:01] <rye> dobey: well, contacts syncing was not said to be discontinued, so it is expected to be working
[17:02] <dobey> rye: it is not installed/enabled by default, we don't provide a ui to install/enable it, and we announced we would be dropping support for it. until we turn off the server, it will continue to randomly work or not work, depending on an innumerable number of variables over which no man has any control :)
[17:07] <rye> dobey: AttributeError: 'module' object has no attribute 'DBUS_CRED_PATH'
[17:07] <rye> dobey: with nightlies :(
[17:08] <dobey> eh?
[17:08] <dobey> rye: what version exactly?
[17:09] <dobey> rye: or is that from desktopcouch?
[17:10] <rye> dobey: that's from desktopcouch, frankly speaking i don't see any announcement regarding the future of contacts syncing
[17:10] <dobey> https://lists.ubuntu.com/archives/ubuntu-desktop/2011-November/003474.html
[17:11]  * dobey needs to poke about getting desktopcouch pulled from the archive for 12.10
[17:13] <dobey> rye: just saw this too: http://askubuntu.com/questions/161640/what-should-i-do-if-i-didnt-recieve-the-verification-email-for-ubuntu-1
[17:18] <rye> dobey: contact ISD
[17:20] <rye> dobey: answerred
[17:20] <rye> r
[17:21] <dobey> thanks
[17:21] <mmcc> dobey: if you have some time, I have a Needs-Information for your Needs-Fixing on https://code.launchpad.net/~mikemc/dirspec/add-exefind/+merge/113782 (and I made some changes to that branch)
[17:24] <dobey> ok, will re-review and answer shortly
[17:25] <mmcc> thanks.
[18:50] <dobey> it's brain thinking time
[19:04] <mmcc> lunch eating time, over here.
[20:17] <dobey> hrmm
[20:17] <dobey> on win32/darwin, is sys.frozen a boolean or what?
[20:19] <mmcc> checking...
[20:21] <dobey> searching the interwebs is not helpful
[20:22] <mmcc> nope, it is not ...
[20:22] <dobey> and all the examples i see are using hasattr()
[20:23] <mmcc> right - it's apparently not defined in the sys module, it's set by py2exe and py2app's bootloaders.
[20:23] <dobey> right
[20:23] <mmcc> py2app sets it to a string, either 'macosx_app' or 'macosx_plugin'
[20:24] <mmcc> cx_freeze sets it to a boolean
[20:24] <dobey> ah-hah
[20:24] <mmcc> (True, obv)
[20:24] <dobey> sys.frozen only exists in the executable. It is set to "console_exe" for a console executable, to "windows_exe" for a console-less gui executable, and to "dll" for a inprocess dll server.
[20:24] <mmcc> and apparently, pyinstaller also sets it to a bool
[20:25] <dobey> ok
[20:25] <dobey> so falling back to None should be fine then
[20:25] <mmcc> yeah, I think so
[20:25] <dobey> was just making sure it wasn't getting added as an attribute that was set to None :)
[20:26] <dobey> because that would be fun...
[20:26] <mmcc> I guess we could simplify the platform check a bit, and just check sys.frozen against 'console_exe' or 'macosx-app' or whatever, but actually, then we'd get two cases for windows, so...
[20:26] <mmcc> yeah
[20:27] <dobey> yeah i'm not sure we want to do that exactly
[20:28] <mmcc> yeah, no, probably not
[20:41] <dobey> mmcc: ok, finally got some comments added to that proposal. hopefully what i said makes sense, and i don't sound like an oppressive tyrant :)
[20:42] <mmcc> ha, dobey. thanks for looking at it carefully. If we're going to try to unify things to pay down tech debt, it's worth getting it right.
[20:42] <mmcc> now I'll go read it and we'll see how oppressed I feel
[20:42] <dobey> heh
[21:05] <mmcc> I'm beginning to think the cross-project exe finder is a bad idea
[21:06] <mmcc> I'm writing up what the API would look like if we avoid passing in a module, which I agree is kind of gross
[21:10] <dobey> well, it's better than writing the same code N times i guess :)
[21:15] <mmcc> dobey, I guess that depends on what you think of how it looks with explicit parameters - I just updated the merge with another comment with some details.
[21:16] <mmcc> if N = # projects instead of # callsites, it might be OK, and the code would definitely be simpler
[21:17] <mmcc> a good deal of this complexity is because we want to use it for sso and u1_client / controlpanel, and we want dirspec to avoid knowing anything about its clients
[21:18] <mmcc> if I wrote one version for sso and one for ubuntuone/platform/ipc, each one could be simpler - you could use the module's __file__ and a local app_names dict, and it wouldn't be so weird
[21:19] <dobey> well it wouldn't be weird, because those aren't libraries that people might use for finding stuff.
[21:19] <dobey> so they can be very specific
[21:20] <mmcc> exactly.
[21:20] <dobey> and obviously putting in specific things to a generalized place, will seem weird
[21:20] <dobey> but part of those specific things, is also the duplication of some generic code
[21:20] <dobey> and just read your comment
[21:21] <dobey> will need to do a bit more brain thinking on it
[21:21] <mmcc> sure, while you do that I'll think about your other point - why we need to / shouldn't do the 'python ' prepending...
[21:35] <mmcc> dobey, I wrote more words about that issue, on the MP. basically, I agree it shouldn't be in the module. we probably shouldn't have to do it at all, but I'm not sure it's worth fixing/replacing buildout before I can move forward on this...
[21:41] <dobey> mmcc: ok, i'm not sayinf we should fix/replace buildout as a means of blocking this branch. i do think it needs further discussion though, and i'd like to avoid diverting the core issue of the branch (unifying the find_me_an_executable stuff), by also having that in the branch. perhaps prefixing python in the calling method when needed, for now. mostly i don't understand why this code would ever launch anything that wasn't instal
[21:42] <dobey> anyway, i will also think about that some more as well
[21:44] <mmcc> oh, I didn't think you were saying fix buildout first, I was just trying to give my understanding of why it's currently necessary to prepend python when running from source.
[21:45] <mmcc> the reason it'll launch things that aren't installed is e.g. when I'm running controlpanel from source while trying to fix bugs, it will need to run syncdaemon from source - at least on darwin there's no good 'general' installed location to default to looking in - either it's all in the same packaged wrapper or it's all in wherever we put the source trees.
[21:45] <dobey> right, i understand the /usr/bin/python thing and python path thing. what i don't understand, is why this is happening at all; iow, what is running stuff from the tree where this is a problem?
[21:46] <dobey> ah ok
[21:46] <mmcc> s/packaged wrapper / packaged app bundle/
[21:46] <mmcc> and the app bundle takes a long time to build, so you really want to be able to run it from the source trees
[21:47] <dobey> right. thanks
[21:47] <dobey> i will think on it a bit more, knowing that
[21:47] <mmcc> cool. thanks
[21:48] <dobey> and am out of here for now. i'll comment tonight or in the morning, whenever i think of more useful words to put there :)
[21:48] <dobey> later
[21:51] <mmcc> ok, bye dobey. thanks again