| mmcc | ok - my branch to fix the race in controlpanel.backend that caused too many calls to get_credentials, and on my system, caused the eternal loading overlay at startup, is here: https://code.launchpad.net/~mikemc/ubuntuone-control-panel/fix-credentials-race/+merge/121969 | 00:25 |
|---|---|---|
| mandel | mmcc, awesome! | 00:33 |
| mandel | mmcc, I need to test that branch merged with the stale broker fix :) | 00:33 |
| mandel | oh, I'm back :) | 00:33 |
| mmcc | hi mandel, what time is it for you? | 00:34 |
| mandel | mmcc, is 2 am, but not to worry, lets try to fix all this bugs so that the only thing we do tom is release | 00:34 |
| mandel | mmcc, by the way, is this a public release or internal? | 00:35 |
| mmcc | mandel: public, ralsina mentioned a blog post | 00:35 |
| mandel | frack | 00:35 |
| mandel | !!! | 00:35 |
| mandel | mmcc, ok, we better do a good one then :) | 00:35 |
| mmcc | heh | 00:35 |
| mandel | mmcc, so far we are much better than the first windows one | 00:35 |
| mmcc | well, as soon as this build uploads I'm EOD, but I will probably check back a little bit tonight, please let me know where you leave things, maybe I can test or review a branch | 00:36 |
| mandel | mmcc, ok, if things work I'll send you an email with the branches to review | 00:37 |
| mmcc | mandel: sounds great | 00:37 |
| mandel | mmcc, ideally with the stale broker fix and your we have a working sso | 00:37 |
| mandel | mmcc, next would be the daemon, right? | 00:37 |
| mandel | mmcc, your branch, what is the expected behavior? | 00:43 |
| mmcc | here's the bundle with my branch: http://ubuntuone.com/3Ak3HKLhv7KETroVctKBkV | 00:43 |
| mandel | for me when getting the new creds from sso it gets stuck | 00:43 |
| mmcc | mandel: gets stuck where? did you already have creds? this was solving an issue when you already had creds, it would hang | 00:44 |
| mmcc | I have to go, will try to check back in a few hours, sorry | 00:44 |
| mandel | mmcc, ok, no problem then, it gets stuck in the wizard where I was working :) | 00:44 |
| mandel | mmcc, go go, have fun with the baby :) | 00:44 |
| mmcc | mandel: just thought of this: you probably don't see the issue my branch fixes unless you have merged from trunk since gatox' share links branch landed… | 01:32 |
| mandel | mmcc, oh, ok, will do | 01:32 |
| ralsina | hi mandel | 01:48 |
| ralsina | go to bed, start fresh tomorrow | 01:49 |
| ralsina | it's what 3AM there? | 01:49 |
| mandel | ralsina, is 4 am, I prefer to do the opposite, finish late with things working | 01:52 |
| mandel | ralsina, we need review by the way :) | 01:52 |
| mandel | ralsina, mmcc told me is a public release, has that been announced? | 01:52 |
| ralsina | mandel: it's "public" | 01:57 |
| ralsina | mandel: it's also alpha | 01:57 |
| ralsina | so, it will be announced internally, and if anyone outside wants to try it at his own peril, he can | 01:57 |
| ralsina | mandel: have links for the reviews? | 01:58 |
| ralsina | mandel: I suspect we never saw the credentials race condition because 1) On linux it doesn't happen 2) on windows the timing is different | 02:02 |
| mandel | ralsina, sure, here they are: https://code.launchpad.net/~mikemc/ubuntuone-control-panel/fix-credentials-race and https://code.launchpad.net/~mandel/ubuntu-sso-client/fix-stale-broker/+merge/121973 | 02:02 |
| ralsina | mandel: done mmcc's starting on yours | 02:03 |
| mandel | great | 02:03 |
| ralsina | mandel: with this one you can do the initial login, or that is something else? | 02:04 |
| mandel | ralsina, close, we are missing the signals parts we talked about, I'm trying to find out why Qprocess.finish signal is not raised | 02:04 |
| ralsina | mandel: ack | 02:06 |
| mandel | ralsina, can you please try this in your mac: paste.ubuntu.com/1175005 | 02:07 |
| ralsina | mandel: sure | 02:08 |
| mandel | ralsina, do I need to have a qapp running for that? | 02:08 |
| ralsina | mandel: let me mool | 02:08 |
| ralsina | look | 02:08 |
| ralsina | yes | 02:08 |
| ralsina | a QCoreApplication should be enough | 02:08 |
| ralsina | you do the start, then the app.exec_() | 02:09 |
| ralsina | mandel: I am not thinking clearly. I am going to bed, and will be back early and do reviews then | 02:11 |
| mandel | ralsina, no problen :) | 02:11 |
| ralsina | <murthaugh>I'm too old for this shit</murthaugh> | 02:12 |
| mandel | lol | 02:14 |
| mandel | ralsina, found a work around, the twisted implementation of the runner works.. later please refresh my mind on why we have to use the qt one.. | 03:54 |
| mandel | now to bed, is close to 6 am :P | 03:54 |
| mmcc | mandel - your pasted script gave me an idea. looks like another Qt ordering problem - you have to make sure that the QCoreApplication is initialized before you call start(). I wonder if that's a problem we're having… | 05:54 |
| mmcc | here's my version of your script that shows that: http://paste.ubuntu.com/1175209/ | 05:54 |
| mmcc | interesting, it really looks like it should work. I can reproduce the same issue in SSO + CP -- no finished signal getting sent | 06:14 |
| mmcc | but it looks like we check that the QApplication instance is not None before we call qt.spawn_program, so that check should avoid what I was seeing… | 06:15 |
| mmcc | well, I see the same thing, where if you use twisted runner, it works. that might be fine for a workaround, but if we need Qt for some reason, (i didn't find a reason in the logs), then I have a big clue -- | 06:30 |
| mmcc | the sso gui ubuntu-sso-login-qt is not spawned by controlpanel, it's spawned by ubuntu-sso-login, which is launched by controlpanel using tcpactivation… so we aren't using qt there, we actually use subprocess.Popen() | 06:35 |
| mmcc | ack, I really don't get why the twisted runner works, though. not thinking clearly here. going to sleep | 06:42 |
| JamesTait | Good morning all! :) | 08:23 |
| === _thumper_ is now known as thumper | ||
| === mpt_ is now known as mpt | ||
| === yofel_ is now known as yofel | ||
| mandel | mmcc, ralsina, I have a fix for the signing problem, I don't like it but is a single line, mainly using the qt runner on mac does nto work while the tx one works | 11:02 |
| mandel | mmcc, yes, the issue here is that ubuntu-sso-login is using the qtreactor, which is done so that we have support for proxies using QtNetwork, I don't see a reason why we should use the qt runner if it does not work | 11:06 |
| mandel | at least for this release we can go with this | 11:07 |
| ralsina | mandel: sure, if it works, do a branch that switches it and we merge that one manually for release | 11:07 |
| ralsina | and that's the last thing we needed fixed! | 11:07 |
| mandel | ralsina, I was thinking a single if sys.platform == 'darwin' | 11:08 |
| ralsina | mandel: sure | 11:09 |
| ralsina | mandel: just we shouldn't mergeit to trunk, I think | 11:10 |
| ralsina | mandel: with 24 hours to think I can surely fix it | 11:10 |
| mandel | ralsina, ok, but there are something I have been thinking about, QProcess is not thread safe, and this is being executed in the qtreactor, I don't know if the qtreactor is putting the guy in a thread and there we are having issues | 11:10 |
| mandel | ralsina, also, why using qprocess over the tx version if we are using a reactor, I understand using it on linux because there is no reactor.. but in the others | 11:11 |
| mandel | I find it a safer option | 11:11 |
| mandel | ralsina, if you want to test it: lp:~mandel/ubuntu-sso-client/send-signals Remove | 11:12 |
| mandel | ups, copy paste fail from lp :) | 11:12 |
| mandel | ralsina, has the shares tab ever worked on mac? | 11:17 |
| mandel | ralsina, I'm moving to see if I can get the daemon support | 11:23 |
| mandel | ralsina, please review: https://code.launchpad.net/~mandel/ubuntu-sso-client/fix-stale-broker/+merge/121973 | 11:32 |
| mandel | I want to start using trunk for the buils asap | 11:32 |
| alecu | hello, y'all! | 11:45 |
| mandel | alecu, morning! | 11:47 |
| alecu | mandel: got anything I can help with? | 11:47 |
| mandel | alecu, got questions for you :) | 11:47 |
| alecu | mandel: it wasn't me, I swear it! | 11:48 |
| mandel | alecu, first, I noticed that the qt runner from ubuntu-sso-utils is not running on mac, why do we use it over the tx implementation? | 11:48 |
| mandel | alecu, I know it was not you, I always do a bzr blame hehe | 11:48 |
| alecu | mandel: probably because we do the same elsewhere | 11:48 |
| alecu | mandel: hehehe | 11:48 |
| mandel | alecu, ok, so I have a dirty fix at least for the release which is a sys.platform == 'darwin' :) | 11:49 |
| mandel | alecu, also, can you tell me where you guys where stuck with the daemon? mmcc told be it was not working, right? | 11:49 |
| alecu | mandel: a sys platform check sounds fine for this. | 11:54 |
| alecu | mandel: regarding the daemon, I've been peeking at how macfsevents and twisted interact, and I don't like it at all. Small testing scripts that use both end up segfaulting randomly. | 11:55 |
| mandel | alecu, and the fsevents daemon? I want to fix that guy today, I remember mmcc telling me you helped him a little with that | 11:56 |
| alecu | mandel: did I helped with the root daemon? | 11:58 |
| mandel | alecu, that is what I understood, I could be wrong.. | 11:58 |
| mandel | alecu, I'm going to try to make things work using the root daemon to see if we don't have those segfaults | 11:59 |
| mandel | if that is the case, we need to fix fsevents.. | 11:59 |
| mandel | alecu, this fixes the runner problems: https://code.launchpad.net/~mandel/ubuntu-sso-client/send-signals I don't know if it should land in trunk or not.. | 12:00 |
| alecu | mandel: we should test if that works better than the other one. | 12:02 |
| mandel | alecu, it works on mac, which the other doesn't I'd like to understand better the qtreactor + yield and what it really does, also QProcess is not thread-safe but reentrant | 12:05 |
| alecu | mandel: what do you want to know about "qtreactor + yield"? | 12:06 |
| alecu | mandel: reentrant usually means "thread safe" | 12:06 |
| mandel | alecu, no perse, you can access its vars but on qt bad things happen with the signals and slots | 12:08 |
| mandel | alecu, you have to connect them in the right thread where you want the information.. | 12:08 |
| dobey | reentrant means "you need to know what you're doing and use the proper locking APIs for this to work in threads" | 12:08 |
| mandel | I wonder if we are not getting one of the signals due to that.. but again, I don't know enough about the qtreactor | 12:09 |
| ralsina | mandel: you can connect signals across threads just fine | 12:11 |
| ralsina | mandel: Qt should be smart enough and use a queued signal in those cases. However thoe have to be Qt threads, not python threads | 12:11 |
| ralsina | mandel: which is probably not the case here | 12:11 |
| alecu | mandel: well, the qtreactor should handle all that by itself. I think it's much better to use the tx runner than the qt one, but as always, it's better to test it irl | 12:11 |
| alecu | mandel: both inside our code, and on small test scripts. | 12:11 |
| ralsina | alecu: we may even switch to the tx runner everywhere if it works everywhere and remove some complexity | 12:12 |
| alecu | ralsina: we can't use it on sso on linux, since we don't have any reactor there. | 12:13 |
| mandel | ralsina, alecu, everywhere as in windows and mac right, 'cause we do not have a reactor on linux for sso or do we? | 12:13 |
| alecu | mandel: exactly. | 12:13 |
| ralsina | alecu: but there we are not using the qt one either, right? | 12:13 |
| mandel | alecu, he, stop reading my mind! | 12:13 |
| ralsina | so the qt runner runs out of uses | 12:13 |
| alecu | mandel: anyway: IRL tests, tests, tests. | 12:13 |
| mandel | ralsina, for the proxy support we use qt | 12:13 |
| mandel | alecu, ^ | 12:13 |
| ralsina | grmbl | 12:14 |
| ralsina | ok, out of scope for today, but let's keep it as something to think about | 12:14 |
| mandel | ralsina, ditto and I said the same at 6 am :P | 12:14 |
| ralsina | mandel: no, the shares tab never worked on mac yet | 12:14 |
| ralsina | mandel: I am going to disable+hide it for this release | 12:14 |
| ralsina | mandel: and fix it for the next one | 12:14 |
| mandel | ralsina, the small one liner fixes the problem, at least for the release (about the runner) | 12:14 |
| alecu | mandel, ralsina: can you try running a couple of times this script? | 12:15 |
| ralsina | mandel: awesome | 12:15 |
| mandel | alecu, sure, url? | 12:15 |
| alecu | mandel: doh: pastebin.ubuntu.com/1174680/ | 12:15 |
| mandel | ralsina, with what we have we might have a nice release.. but I'd like to talk with mmcc and try to get it using the root daemon | 12:15 |
| ralsina | alecu: sure | 12:15 |
| mandel | also land want we have around in private branches | 12:15 |
| ralsina | mandel: probably not on this one :-( | 12:15 |
| * alecu hates new firefox that strips http:// from urls copied from the url bar. | 12:15 | |
| ralsina | alecu: yes, that's deeply stupid | 12:16 |
| alecu | ralsina, mandel: here, that script dies 4/10 times with some memory corruption. | 12:16 |
| mandel | alecu, the exception is expected? | 12:17 |
| mandel | alecu, whqt I get: paste.ubuntu.com/1175739 | 12:18 |
| alecu | mandel: yes, the exception is expected. | 12:19 |
| alecu | mandel: you are running with your latest macfsevents, right? | 12:20 |
| mandel | alecu, yes | 12:20 |
| ralsina | alecu: no memory corruptions, just the exceptions | 12:20 |
| mandel | alecu, shall we try passing process_asap = True to the observer? | 12:20 |
| alecu | mandel, ralsina: can you try running it a couple of times? | 12:21 |
| alecu | mandel, yes, please. | 12:21 |
| mandel | alecu, I'm in the process | 12:21 |
| ralsina | alecu: on the 10th time, no memory issues | 12:21 |
| alecu | mandel, ralsina: also: add a line with a "raise", just after the log.err() | 12:21 |
| ralsina | alecu: ok | 12:21 |
| mandel | alecu, and latency=0 too? | 12:22 |
| mandel | alecu, is the other setting we use | 12:22 |
| ralsina | alecu: now I got "called with null pointer" | 12:22 |
| alecu | mandel, sure. | 12:23 |
| ralsina | alecu: is that the error you wanted? | 12:23 |
| alecu | ralsina: can you paste it? | 12:23 |
| ralsina | alecu: sure | 12:23 |
| mandel | same here | 12:23 |
| alecu | ralsina, mandel: ok: I got that first, so I thought it might be the exception being missed in the fsevents callback. | 12:24 |
| alecu | but then I saw that I got similar corruption issues when I was handling the error inside the callback too. | 12:24 |
| mandel | alecu, with process_asap=True it never finishes.. | 12:25 |
| mandel | it keeps saying hola! | 12:25 |
| alecu | so, as far as I can tell, this is a race condition between the macfsevents thread and extension, and the reactor we are using on mac. | 12:26 |
| ralsina | alecu: http://ubuntu.pastebin.com/1175753 | 12:26 |
| alecu | ralsina: thanks. | 12:26 |
| alecu | ralsina: you've typed that url, right? :-) | 12:26 |
| ralsina | alecu: yes, paste from vnc into ubuntu is not working | 12:26 |
| ralsina | alecu: letme check ;-) | 12:27 |
| alecu | ralsina: I just fixed it in my browser :-) | 12:27 |
| ralsina | alecu: right, pastebin.ubuntu.com/same number ;-) | 12:27 |
| alecu | ralsina: so, yes: that's similar to the errors I've been getting. | 12:27 |
| mandel | ralsina, alecu, same errors here | 12:28 |
| alecu | ralsina, mandel: so, my options at this point: | 12:29 |
| alecu | 1) we switch to some other reactor supported on mac for SD | 12:29 |
| alecu | 2) we move the macfsevents stuff to a different process | 12:29 |
| alecu | 3) we find a magical way to fix this. | 12:30 |
| alecu | ... | 12:30 |
| mandel | sounds reasonable.. I vote for the magic unicorn fix.. | 12:31 |
| mandel | alecu, we can try the reactor that uses the main loop from cocoa, that might be useful and is in theory a support reactor in twisted.. | 12:31 |
| ralsina | +1 for magicalunicorns | 12:31 |
| mandel | but yet again, the IOCompletionPorts one too | 12:31 |
| ralsina | how complex is swithing reactors going to be? | 12:32 |
| ralsina | I am tempted to do a "non-root-daemon" using macfsevents | 12:32 |
| ralsina | and armonize it with the root daemon, conceptually | 12:32 |
| mandel | in theory it should just mean changing the import to install the new one.. | 12:33 |
| mandel | theory.. | 12:33 |
| alecu | ralsina: +1 to a user-daemon | 12:33 |
| alecu | mandel: remember that some of the osx reactors are "Unmaintained": http://twistedmatrix.com/documents/current/core/howto/choosing-reactor.html#auto1 | 12:33 |
| ralsina | twisted reactors are like a minefield, only without the fun explosions | 12:34 |
| ralsina | alecu: so, current status is, this crashes sd and then sd restarts a lot, right? | 12:34 |
| mandel | alecu, maybe kqueue? | 12:34 |
| alecu | ralsina: right | 12:34 |
| alecu | mandel: test-test-test | 12:34 |
| mandel | yes yes yes | 12:35 |
| ralsina | alecu: ok, so let's try to switch reactors now, and the plan is to do a daemon and avoid this crap | 12:35 |
| alecu | mandel: let's start by doing the corresponding imports in the test script: | 12:35 |
| alecu | like: | 12:35 |
| alecu | from twisted.internet import kqreactor | 12:35 |
| alecu | kqreactor.install() | 12:35 |
| alecu | and so on | 12:35 |
| ralsina | alecu: have 2' for a quick mumble? | 12:36 |
| mandel | with the kqreactor we have the same PyCObject_AsVoidPtr problem | 12:36 |
| mandel | alecu, ^ | 12:36 |
| alecu | ralsina: sure | 12:36 |
| mandel | I'm off to have lunch | 12:37 |
| mandel | alecu, cfreactor seems to work.. anyways, off to get food | 12:38 |
| alecu | mandel: how did you end up installing pyobjc? | 13:11 |
| alecu | mandel: I was unable to compile pyobjc with the brewed python. | 13:39 |
| mandel | alecu, I cannot remember.. | 13:39 |
| alecu | and if I use the 10.8 python with the 10.8 pyobjc, cfreactor dies after the second exception is raised, without printing anything useful. | 13:40 |
| mandel | alecu, wait, is not meant to do that.. I though it was.. | 13:41 |
| mandel | mierda.. | 13:41 |
| mandel | alecu, I have that problem too then | 13:41 |
| dobey | gah | 13:59 |
| dobey | dbus is crashing and i don't know why :( | 13:59 |
| ralsina | dobey: dbus as a whole? | 14:02 |
| dobey | no, dbus lib making a call from libsyncdaemon | 14:05 |
| dobey | so, probably libsyncdaemon bug | 14:05 |
| dobey | yup | 14:07 |
| * dobey wonders if that's one of the existing bugs | 14:08 | |
| dobey | wow, there isn't a bug for this crash | 14:20 |
| dobey | i guess it's hard to hit when you just get a giant dialog with a huge empty white box | 14:20 |
| dobey | and it's such a trivial fix too :( | 14:26 |
| dobey | briancurtin, ralsina: https://code.launchpad.net/~dobey/ubuntuone-client/fix-1043868/+merge/122076 please :) | 14:28 |
| ralsina | dobey: trivial fixes for which you propose branches should always be along a ":-)" | 14:29 |
| dobey | ralsina: it's a :( because this bug should have never existed | 14:29 |
| briancurtin | dobey: looking once the diff generates | 14:29 |
| dobey | it's a silly missing "+ 1" for an allocation call | 14:29 |
| ralsina | yes, silly off-by-ones | 14:29 |
| ralsina | briancurtin: don't bother, I got it | 14:29 |
| dobey | are we doing mumble or hangout? | 14:30 |
| ralsina | briancurtin: BTW, look at https://bugs.launchpad.net/ubuntuone-client/+bug/1043359 | 14:30 |
| ubot5 | Ubuntu bug 1043359 in Ubuntu One Client "[windows] Does not launch control panel if Windows Firewall service is down" [Medium,Confirmed] | 14:30 |
| ralsina | dobey: mumble | 14:30 |
| ralsina | briancurtin: after you are done with the rush of py3 stuff, we should do a release that fixes that | 14:30 |
| briancurtin | ralsina: yeah i saw that one come in yesterday. need to think of a way to test if that's up first | 14:30 |
| ralsina | briancurtin: http://installbuilder.bitrock.com/docs/installbuilder-userguide/ar01s09.html | 14:31 |
| ralsina | briancurtin: it's even the example :-) | 14:31 |
| briancurtin | bam | 14:31 |
| dobey | ralsina: ok; i will need to leave promptly at 11:30 today, and if we can cut it a few minutes earlier, even better :) | 14:32 |
| ralsina | dobey: sure | 14:32 |
| ralsina | dobey: you talk first to be sure | 14:32 |
| ralsina | dobey: before we ramble on mac stuff | 14:32 |
| ralsina | thisfred: you're excused from today's call if you want | 14:33 |
| thisfred | I'll take that :) | 14:33 |
| ralsina | thisfred: we have a ton of mac stuff to discuss and AFAIK you are doing server-side stuff this week, right? | 14:33 |
| thisfred | yeah 99% server side | 14:34 |
| dobey | ralsina, briancurtin: and https://code.launchpad.net/~dobey/ubuntuone-client-gnome/no-more-eds/+merge/122080 which is a fair bit less trivial. :) | 14:45 |
| ralsina | dobey: ack | 14:45 |
| ralsina | dobey: global +1 on the other one BTW | 14:45 |
| dobey | thanks | 14:45 |
| briancurtin | dobey: will look | 14:46 |
| dobey | may need design approval on the -gnome one, but UI freeze is today so was hoping to get it in since I finally got a chance to do that work last night/this morning | 14:46 |
| ralsina | dobey: if you have screenshots we can run it through design quickly | 14:51 |
| dobey | i can nab one | 14:52 |
| mmcc | hi folks. catching up now… | 14:56 |
| ralsina | good morning mmcc! | 14:56 |
| dobey | ralsina: http://ubuntuone.com/4SPMPsP5FysA55FQK5MU8V | 14:58 |
| dobey | also added to comment | 14:58 |
| ralsina | dobey: have a before shot? | 14:58 |
| mandel | ralsina, dobey or thisfred, can I have a very simple review: https://code.launchpad.net/~mandel/ubuntuone-client/ignore-ds-store/+merge/122084 | 14:59 |
| mandel | the bug was really annoying me :) | 14:59 |
| ralsina | mandel: sure thingie! | 14:59 |
| mandel | ralsina, also need to rest my brain from harder bugs hehe | 14:59 |
| ralsina | mandel: yeah, we needed that fixed before release, really :-) | 14:59 |
| dobey | ralsina: one second | 15:00 |
| mandel | ralsina, the ds_store one, cool I had no idea it was urgent, it was annoying me that is all | 15:00 |
| ralsina | mandel: well, it was going to annoy everyone :-) | 15:00 |
| dobey | ralsina: https://launchpadlibrarian.net/85567180/u1-sharing.png | 15:00 |
| ralsina | alecu, mandel, mmcc, dobey, briancurtin: mumbles! | 15:00 |
| dobey | ralsina: is attached to the one "omg confusing" bug | 15:00 |
| mandel | ralsina, there already | 15:01 |
| ralsina | dobey: oh, great | 15:01 |
| ralsina | dobey: I'll get design signoff after the call | 15:01 |
| dobey | ok | 15:01 |
| ralsina | joshuahoover: call if you are free :-0 | 15:01 |
| ralsina | that was meant to be a riendly ":-)" | 15:02 |
| joshuahoover | ralsina: heh, joining... | 15:02 |
| ralsina | mandel: on https://code.launchpad.net/~mandel/ubuntuone-client/ignore-ds-store/+merge/122084 the indentation of the new line is off, can you fix that? | 15:09 |
| mandel | ralsina, is not off that is an lp problem, open the file :) | 15:10 |
| ralsina | mandel: ok | 15:10 |
| mandel | ralsina, see: http://bazaar.launchpad.net/~mandel/ubuntuone-client/ignore-ds-store/view/head:/data/syncdaemon.conf#L80 | 15:11 |
| mandel | stupid lp | 15:11 |
| dobey | mandel: it is off because you mixed tabs and spaces | 15:14 |
| mandel | dobey, fixing it | 15:17 |
| mandel | dobey, done | 15:19 |
| mandel | mmcc, the branch that fixes it: https://code.launchpad.net/~mandel/ubuntu-sso-client/send-signals | 15:22 |
| mandel | mmcc, little and crappy but works | 15:22 |
| mandel | dobey, please approve the branch before you go so that is not blocked :) | 15:23 |
| dobey | mandel: bueno. +1 :) | 15:24 |
| mandel | dobey, gracias!!! | 15:24 |
| mmcc | mandel: I'll try to propose it soon, but it'll be after a new test bundle for sure. | 15:41 |
| mandel | mmcc, no problem, I found something I can have fun with.. and is mac related :) | 15:41 |
| mmcc | mandel: we don't copy the plist - it gets embedded into the daemon executable as a linker __TEXT section, and the launchdadd process gets it from there and puts it in a system location we can't touch | 15:41 |
| mmcc | mandel: cool, what? | 15:41 |
| mandel | mmcc, pync :) | 15:42 |
| mandel | mmcc, notification center support | 15:42 |
| mmcc | mandel: great | 15:42 |
| chaselivingston | mandel: oooh, I would love that | 15:42 |
| mmcc | will it work with growl in 10.7 also? | 15:42 |
| mandel | mmcc, I'll add support for both, is very small code that needs to be added | 15:42 |
| mandel | thisfred, when are the notifications shown in u1? | 15:53 |
| thisfred | mandel, that question is a little too vague for me to answer | 15:53 |
| mandel | thisfred, is there an action I can perform to trigger a notification? | 15:54 |
| ralsina | mandel: why nt start with growl and then use that free app for nc? | 15:54 |
| thisfred | mandel, you can drop a file into your u1 folder | 15:54 |
| mandel | ralsina, because we can have both very cheap | 15:54 |
| ralsina | mandel: ok | 15:54 |
| mandel | ralsina, no need for a 3 party app and is like 20 lines of code for each | 15:55 |
| thisfred | mandel, notifications are / should be shown whenever u1 starts or finishes uploading or downloading | 15:55 |
| thisfred | uploading is obviously the easiest one to trigger | 15:55 |
| mmcc | mandel: you're going to run into exceptions from the on_download_finished remote call… | 15:56 |
| mandel | mmcc, true.. I'll have to fix that too, is easy, or should be easy.. | 15:56 |
| mmcc | just a heads-up. gatox fixed some of those for the sync indicator but not all. you can see what he did. it's either removing or adding an asterisk | 15:56 |
| mandel | ralsina, mmcc, please don't hate me, but can we push this in the release: https://code.launchpad.net/~mandel/ubuntuone-client/fix-wrong-fs-action/+merge/122099 | 16:07 |
| mandel | is kind of important.. | 16:07 |
| mmcc | mandel, is fsevents.IN_ATTRIB == 4? | 16:07 |
| mandel | mmcc, yes | 16:07 |
| mmcc | hooray | 16:08 |
| ralsina | mandel: I don't mind, but I have no idea what that does, so it's up to mmcc | 16:08 |
| mmcc | that's the bug I ran into two days ago | 16:08 |
| mmcc | +1 | 16:08 |
| mandel | :) | 16:08 |
| mmcc | well, in spirit ;P -- when does IN_ATTRIB get sent, exactly? | 16:09 |
| mandel | mmcc, in my case it got when I did a copy from one location to another | 16:09 |
| mmcc | really | 16:10 |
| mmcc | well, we do need to handle that event, and I figure sending modify is a good start… besides we need to have a hard look at fsevents.py after the release anyway | 16:11 |
| mandel | mmcc, yes and yes | 16:11 |
| mmcc | ok, fsevents sends IN_ATTRIB if it sees a file that has the same modified time but different ctime -- "the time of most recent metadata change". So we might be sending an event when the content hasn't actually changed -- will that be a problem up the stack? | 16:14 |
| ralsina | mmcc, mandel: branch to hide the share tab completely -- lp:~ralsina/ubuntuone-control-panel/ninja-share-tab | 16:14 |
| mmcc | this IN_ATTRIB was added upstream in fsevents trunk -- we could just ignore it | 16:14 |
| mmcc | ralsina: grat | 16:14 |
| mandel | mmcc, no, we will check the hash and if it the same we do nothing | 16:15 |
| mmcc | mandel ok then I will approve that branch | 16:15 |
| mandel | mmcc, we do the same on windows | 16:15 |
| mandel | mmcc, is a little ugly, I agree | 16:15 |
| mmcc | anyone else want to review that? | 16:16 |
| ralsina | mmcc: merge it | 16:17 |
| ralsina | consider that a virtual +1 | 16:17 |
| ralsina | and... lunch! | 16:17 |
| mmcc | done | 16:17 |
| mmcc | wait ralsina - did you do the IRL test of mandel's fix-stale-broker? | 16:17 |
| mandel | ralsina, mmcc, we have notification center support on 10.8 :) | 16:17 |
| ralsina | mmcc: no, I didn't | 16:18 |
| mmcc | ralsina: ok, cool | 16:18 |
| ralsina | mmcc: stupid me, I should have :-( | 16:18 |
| ralsina | mmcc: so you do it ;-) | 16:18 |
| mmcc | ralsina, ack - mandel, can you elaborate on how you tested that? the instructions on the branch weren't totally clear | 16:21 |
| mandel | mmcc, what you can do is the following, start sso in a terminal, ok | 16:22 |
| mmcc | which sso | 16:22 |
| mandel | mmcc, ubuntu-sso-login | 16:22 |
| mandel | the daemon | 16:22 |
| mmcc | ok | 16:22 |
| mandel | mmcc, then in a second terminal start control panel | 16:22 |
| mmcc | right, and I just have to be quick about killing sso-login right after I start cp? | 16:23 |
| mandel | yes, so that it gets the wrong remote object | 16:23 |
| mandel | mmcc, then work normally and things should just 'work' | 16:24 |
| mmcc | when I tried this earlier it was hard to kill it before it finished and killed itself | 16:24 |
| mandel | mmcc, you can use a test script to connect so it does not kill itself | 16:24 |
| mmcc | I guess I never rebuilt the UI since the qss change for the tabs to be left-aligned showed up | 16:26 |
| mmcc | hrm. | 16:26 |
| mmcc | mandel, it's caching the credentials, so I don't think it needs SSO after it started up… I'm not sure I can test this, what am I missing? | 16:28 |
| mmcc | it's possible that I'm just tired here :) | 16:28 |
| mandel | mmcc, hehe, same in this side, I think a good way to test it is request bad creds, kill and retry | 16:28 |
| mmcc | mandel: how do I request bad creds? is that something that happens in real use? sorry, I'm being dense here | 16:30 |
| mandel | mmcc, yes, you can start without them in syncdaemon | 16:31 |
| mmcc | so, delete my creds, start SSO-login, then start CP, then kill SSO login while it's running the gui? | 16:31 |
| mandel | mmcc, mumble? | 16:31 |
| mmcc | mandel: sure, just a sec | 16:32 |
| mmcc | mandel: i'm there | 16:32 |
| mandel | one sec | 16:32 |
| mmcc | ok mandel, approved. that'll be going in the release | 16:44 |
| mandel | awesome! | 16:44 |
| mmcc | btw, why did we make the tabs left-aligned? is that for all platforms? | 16:46 |
| mandel | mmcc, yes, to be consistent | 16:51 |
| mmcc | mandel: wasn't it centered everywhere before? why left on all platforms instead of centered on all platforms? | 16:52 |
| mandel | mmcc, no AFAIK | 16:55 |
| mandel | anyway, EOD here I've got rugby training, laters o/ | 16:56 |
| mmcc | mandel: ok, thanks again | 16:56 |
| mmcc | bye | 16:56 |
| mandel | mmcc, ralsina, before I go, we have notification center and growl support (post 1.3) support, I'll create a bug and propose tom morning :) | 16:59 |
| mmcc | mandel: great | 16:59 |
| briancurtin | does anyone know how to tell where a package came from? like i want to know where the source was taken from to install twisted into python3.2/dist-packages | 17:02 |
| briancurtin | oh wait, i'm an idiot | 17:03 |
| briancurtin | i did that myself... | 17:03 |
| mandel | hahahaha | 17:03 |
| briancurtin | i was trying to blame it on something i apt-get'ed | 17:03 |
| mmcc | with the right combo of filesystem metadata and shell features, one day we might be able to have the utility that tells us what we did to create a given file. I'd call it 'wtf', as in 'where's this from'. | 17:06 |
| mmcc | 'wtf ~/bin/python' => "you installed it in june from ~/source/python-2.3 while you were working on the back port bug" or something like that | 17:07 |
| * mmcc done daydreaming | 17:07 | |
| * alecu will have lunch and do some errands | 17:08 | |
| mmcc | arg, tarmac not happy with client tests right now | 17:12 |
| mmcc | had to re-approve mandel's fsevents tweak | 17:12 |
| mmcc | ok, building a bundle with our tweaks for last-minute testing, then I'll figure out the certificate stuff and work on building a bundle for release with the right cert and an alpha version number | 17:21 |
| chaselivingston | mmcc: should I have two sf's running at once in activity monitor? | 17:31 |
| chaselivingston | *sd's | 17:31 |
| mmcc | chaselivingston: no, they're probably there as a result of a bug we fixed this week. | 17:31 |
| chaselivingston | mmcc: never mind, just got a popup that said sd error and asked me to terminate them | 17:32 |
| mmcc | what did it say the error details were? | 17:32 |
| chaselivingston | mmcc: I didn't look…. sorry | 17:32 |
| mmcc | I'm about to put a new build up, so killing everything and starting over is what I recommend… | 17:32 |
| mmcc | it's uploading - 56% | 17:32 |
| chaselivingston | mmcc: cool, will you post a link here? | 17:32 |
| mmcc | yes | 17:32 |
| chaselivingston | awesome, thanks | 17:33 |
| mmcc | Ok new mac client alpha release candidate build -- http://ubuntuone.com/7COx8IkZd5MInXLL3sb0BN | 17:35 |
| ralsina | mmcc: awesome! | 17:36 |
| mmcc | chaselivingston: please kill your syncdaemons and any other ubuntuone related processes, then delete the "Ubuntu One" item in Keychain Access before starting that. Also remove ~/Library/Preferences/ubuntuone/syncdaemon.conf | 17:36 |
| mmcc | when you start up you should be able to log in with no freeze and see an initial-setup wizard that walks you through picking which folders to sync | 17:37 |
| mmcc | if you could test that, then quit and restart the app a few times to see if anything stalls out on you, that'd be very helpful | 17:38 |
| mmcc | don't worry about killing syncdaemons for that last step, just interested in the UI | 17:38 |
| mmcc | ralsina: if you've got time, can you do the same? | 17:39 |
| chaselivingston | mmcc: cool, will do | 17:41 |
| ralsina | mandel: looks like wrong-fs-event bounced? | 17:42 |
| ralsina | mmcc: will do! | 17:43 |
| chaselivingston | mmcc: made it to folder selection, but I'm stuck at "getting information, please wait" | 17:45 |
| chaselivingston | mmcc: quit and restart and made it right to the cp | 17:46 |
| mmcc | hmm, not great. sorry, have to be a fk for ~10… will check on that in a minute | 17:49 |
| mmcc | er, when i get back. | 17:49 |
| mmcc | chaselivingston: how long did you wait before quitting and restarting? | 17:49 |
| chaselivingston | mmcc: ~2-3 mins | 17:49 |
| ralsina | chaselivingston: I have seen that on windows sometimes | 17:54 |
| ralsina | chaselivingston,mmcc: it's something to verify but probably can be explained away on the mail describing this release | 17:54 |
| chaselivingston | ralsina: agreed, and maybe it was just a weird case for me or something | 17:55 |
| chaselivingston | ralsina: seems to be syncing ok now | 17:55 |
| chaselivingston | ralsina: ok maybe I lied, don't see sd running... | 17:55 |
| ralsina | mmcc: we should make --with-icon the default on darwin! | 17:57 |
| ralsina | chaselivingston: confirmed some steps in the wizard lock | 18:00 |
| ralsina | chaselivingston: next & finish still work though ;-) | 18:00 |
| chaselivingston | ralsina: ah, interesting... | 18:00 |
| mmcc | ugh. ralsina which steps in the wizard? | 18:01 |
| mmcc | I only ever walked through with next and finish… | 18:01 |
| ralsina | mmcc: the first one after login, selecting folders to sync | 18:01 |
| ralsina | other than that, it works for me | 18:02 |
| chaselivingston | mmcc, ralsina: sd seems to be quitting on me after a few mins of closing the cp | 18:02 |
| ralsina | chaselivingston: yes | 18:02 |
| ralsina | chaselivingston: we need to start the cp so it stays alive | 18:03 |
| chaselivingston | so if I leave the cp up for now, it will continue running? | 18:03 |
| ralsina | chaselivingston: if it crashes, it will restart itself at least ;-) | 18:03 |
| chaselivingston | ralsina: haha, good to know, I'll just leave it in the bg for now | 18:04 |
| ralsina | mmcc: I'll do a branch to make --with-icon true by default on darwin | 18:05 |
| mmcc | ralsina: ok | 18:05 |
| mmcc | chaselivingston: can you make a bug for SD quitting, and attach your syncdaemon.log ? I'm wondering why it's quitting | 18:07 |
| chaselivingston | yeah, sure thing | 18:07 |
| ralsina | mmcc: the gc crash? ;-) | 18:07 |
| mmcc | ralsina: possibly, but I was hoping that wasn't so common | 18:09 |
| ralsina | mmcc: shouldn't be | 18:09 |
| chaselivingston | mmcc: https://bugs.launchpad.net/ubuntuone-control-panel/+bug/1043983 | 18:10 |
| ubot5 | Ubuntu bug 1043983 in Ubuntu One Control Panel "syncdaemon crashes when control panel is quit" [Undecided,New] | 18:10 |
| ralsina | mmcc: mine survives u1cp closing, has been up 14 minutes | 18:11 |
| mmcc | ralsina: ok | 18:11 |
| mmcc | I'm seeing the stalled loading overlay on the cloud-to-computer page, and the computer-to-cloud page also seemed stalled but was just really slow | 18:12 |
| chaselivingston | mmcc: mine isn't running even w/ the cp up right now... | 18:12 |
| ralsina | chaselivingston: ok, that's bad | 18:12 |
| chaselivingston | ralsina: maybe I should try a restart? | 18:13 |
| chaselivingston | ralsina: of my computer I mean | 18:13 |
| ralsina | chaselivingston: should not make a difference | 18:13 |
| ralsina | chaselivingston: maybe if we could see logs | 18:13 |
| ralsina | which I don 't recall where they are on darwin | 18:14 |
| chaselivingston | ralsina: that previous bug report I submitted has my sd logs | 18:14 |
| ralsina | chaselivingston: oh, awesome, I'll look then | 18:14 |
| mmcc | ralsina logs are in ~/Library/Caches/ubuntuone and ~/Library/Caches/sso | 18:14 |
| ralsina | mmcc, chaselivingston: yes, that looks like the GC failure to me | 18:15 |
| ralsina | why doesn't it happen for me at all? Could it be that I am on 10.7? | 18:15 |
| chaselivingston | ralsina: hm, and no fix for that yet? | 18:15 |
| ralsina | chaselivingston: not one we can package today :( | 18:15 |
| chaselivingston | ralsina: cool, no worries | 18:16 |
| mmcc | ralsina: are you just looking at the end? I don't recognize that non-traceback | 18:16 |
| ralsina | mmcc: that's a bare exception inside twisted | 18:16 |
| ralsina | mmcc: which is what I got with alecu's test script earlier | 18:16 |
| mmcc | ok | 18:17 |
| mmcc | chaselivingston: what happens if you press the disconnect button and re-connect from within controlpanel? | 18:18 |
| mmcc | does that restart SD for you? | 18:18 |
| chaselivingston | mmcc: let me try, one sec | 18:18 |
| chaselivingston | mmcc: it does start up when I open the cp, but apparently quits some time after that | 18:19 |
| mmcc | can you pastebin the end of the syncdaemon log from after that? | 18:19 |
| chaselivingston | mmcc: me? I'm not even sure I'd know where to look to find that :) | 18:20 |
| mmcc | chaselivingston: yes, you :) just the same syncdaemon.log -- I'm wondering if it quit the same way. you could attach it to the same bug or if it's more convenient, use http://paste.ubuntu.com | 18:21 |
| mmcc | since it shouldn't be too big - I just want the end to see how it died | 18:21 |
| chaselivingston | ah ok, one sec | 18:21 |
| chaselivingston | mmcc: remind me how the copy command works in Terminal? trying to copy it to the desktop from the folder it's in | 18:22 |
| mmcc | cp ~/Library/Caches/ubuntuone/syncdaemon.log ~/Desktop/ | 18:23 |
| chaselivingston | ok, also just posted a crash report I got from sd to that bug | 18:24 |
| mmcc | oh, an osx crash report? great | 18:24 |
| chaselivingston | mmcc: yeah, from console | 18:24 |
| chaselivingston | mmcc: will that do for now, about to go into a meeting, I can try to get the sd logs a bit later if you want | 18:25 |
| mmcc | chaselivingston: just curious, can you load one.ubuntu.com in a browser right now? | 18:25 |
| chaselivingston | mmcc: yep | 18:25 |
| mmcc | ok, thanks | 18:26 |
| dobey | ok, need second review of https://code.launchpad.net/~dobey/ubuntuone-client-gnome/no-more-eds/+merge/122080 asap | 18:33 |
| dobey | briancurtin: can you do it please? :) | 18:33 |
| dobey | i guess he's probably at lunch now | 18:34 |
| ralsina | mmcc: review for https://code.launchpad.net/~ralsina/ubuntuone-control-panel/darwicon/+merge/122122 please? | 18:35 |
| ralsina | also dobey maybe? | 18:36 |
| briancurtin | dobey: oh whoops, i hadn't finished that one yet. will finish in a few mins | 18:37 |
| dobey | ralsina: +1 from me | 18:39 |
| ralsina | dobey: thanks! | 18:39 |
| briancurtin | dobey: looks ok, approved | 18:42 |
| mmcc | ralsina: I am getting test explosions from that branch on darwin… | 18:42 |
| ralsina | mmcc: whoa | 18:42 |
| ralsina | mmcc: can you pastebin, my mac is not very dev-friendly right now | 18:42 |
| mmcc | of course it's mostly the same explosions as before, will have to re-add my hack for the "wrapped c/c++" blah and see if that's it | 18:42 |
| mmcc | give me a minute or two | 18:42 |
| dobey | briancurtin: thanks! | 18:42 |
| ralsina | mmcc: oh ok | 18:42 |
| mmcc | some looked new though :( | 18:43 |
| ralsina | mmcc: it changes a default so it's possible it needs tweaking a ton of places | 18:46 |
| mmcc | ralsina, all the failures are in test_share_links_search.py… http://paste.ubuntu.com/1176430/ | 18:47 |
| ralsina | mmcc: ok, I now can run the tests on mac, so I'll fix them | 18:48 |
| mmcc | I merged your darwicon into trunk, so this doesn't have the full disable branch, iirc | 18:48 |
| dobey | what is up with all the crazy ram usage from the u1client tests lately? | 18:48 |
| mmcc | er, I mean, doesn't have the branch that you put up today to fully disable share_links | 18:49 |
| ralsina | mmcc: it's ok, should be ok without it | 18:50 |
| ralsina | mmcc: unless you already get those errors in trunk | 18:50 |
| mmcc | ralsina: ok, I'll go check in trunk, just a sec | 18:53 |
| mmcc | ralsina: looks like the same errors happen in trunk: http://paste.ubuntu.com/1176455/ | 18:57 |
| ralsina | mmcc: ok, so not branch related | 18:58 |
| ralsina | mmcc: and probably explains why the share tab is broken on mac ;-) | 18:58 |
| ralsina | mmcc: could you do a quick IRL test and see if my branch works? I can try to fix these tomorrow | 18:58 |
| mmcc | ralsina: ok. | 18:59 |
| mmcc | I see the menu, and your new tests pass, so +1 | 19:02 |
| ralsina | yay | 19:02 |
| ralsina | mmcc: bug #1044012 | 19:04 |
| ubot5 | Launchpad bug 1044012 in Ubuntu One Control Panel "file share tab tests fail on mac (and the tab doesn't work)" [High,Triaged] https://launchpad.net/bugs/1044012 | 19:04 |
| ralsina | I don't expect that to take longer than tomorrow to fix | 19:04 |
| ralsina | so maybe we will do another release next week | 19:04 |
| mmcc | cool, yes | 19:04 |
| ralsina | school run! Will be back in about 30' | 19:13 |
| dobey | briancurtin, ralsina: https://code.launchpad.net/~dobey/ubuntuone-client-gnome/update-4-0/+merge/122131 basically same as earlier, just merging to stable to make a relaese and get it in ubuntu in the next 1.5 hrs. | 19:33 |
| briancurtin | dobey: looking | 19:33 |
| dobey | brb, gotta make a quick run | 19:38 |
| ralsina | dobey: this branch bounced because tarmac ran out of memory, right? https://code.launchpad.net/~mandel/ubuntuone-client/fix-wrong-fs-action/+merge/122099 | 19:59 |
| ralsina | mmcc: you may have to merge that one manually ... | 19:59 |
| mmcc | again, jeez | 19:59 |
| chaselivingston | mmcc: fyi, just attached another crash log to that bug I submitted earlier | 20:00 |
| mmcc | ralsina: how do I merge it manually? just push to lp:/ubuntuone-client/trunk ? | 20:00 |
| mmcc | chaselivingston: great, thanks | 20:00 |
| ralsina | mmcc: I meant more merging it when doing the release | 20:00 |
| ralsina | mmcc: like my branch about hiding the share tab | 20:00 |
| mmcc | ralsina: oh, ok sure | 20:01 |
| mmcc | I like that better anyway | 20:01 |
| ralsina | we'll get it in trunk eventually, anyway | 20:01 |
| mmcc | oh hey, I forgot, I already did that. the build I sent earlier today has that branch | 20:03 |
| ralsina | awesome | 20:05 |
| ralsina | so one more merge of trunk and ready to package? | 20:05 |
| mmcc | now fighting with Xcode to download developer-ID signed certificate from apple | 20:05 |
| mmcc | the future answer will be to add someone from IS to the team on apple's site and get them an official cert too. | 20:06 |
| mmcc | or potentially they can just use their certs, although I'm not sure exactly if it *has* to be signed by apple | 20:06 |
| mmcc | as root | 20:06 |
| mmcc | ralsina: merge of trunk for the darwicon branch ? am I forgetting anything else? | 20:07 |
| ralsina | mmcc: yes, for that one | 20:07 |
| mmcc | ok | 20:07 |
| dobey | ralsina: technically it bounced for that reason | 20:11 |
| mmcc | ralsina: building with the darwicon branch now, I will sign it with my Apple provided canonical group Mac Developer cert. I don't think this will fix the 10.8 warning, but it might. There is a single "developer ID" certificate that will definitely make 10.8 happy, even with "gatekeeper" enabled, but only urbanape has that one for now. | 20:30 |
| dobey | ralsina: do i need to get the patch to remove the search bits on the share tab, into quantal today? | 20:30 |
| ralsina | dobey: ? | 20:32 |
| dobey | ralsina: you disabled the search box in http://bazaar.launchpad.net/~ubuntuone-control-tower/ubuntuone-control-panel/trunk/revision/354 | 20:33 |
| ralsina | dobey: I don't think so | 20:34 |
| ralsina | dobey: that doesn't change UI really | 20:34 |
| ralsina | dobey: and I want to do a much better solution | 20:34 |
| dobey | oh, it makes the ui insensitive while it's searching? | 20:34 |
| dobey | or while it's building the index for which to search in? | 20:35 |
| ralsina | dobey: right, while it's building the index | 20:36 |
| ralsina | dobey: what I want to do is provide a second source of autocompletion where you type the path from the left | 20:37 |
| ralsina | dobey: which can be done without prepopulating a list | 20:37 |
| dobey | ah | 20:38 |
| dobey | how does that search work anyway, hits a REST API and stores a big list in memory? | 20:39 |
| ralsina | dobey: worse, scans the disk | 20:39 |
| ralsina | dobey: when it hit my 300K files it kinda sucked ;-) | 20:39 |
| dobey | yeah, that sucks | 20:40 |
| ralsina | dobey: the idea there is that if you want to publish a file, you type the name, choose it from the popup, then blam, it's published. It just needs some refining | 20:41 |
| dobey | yeah | 20:41 |
| dobey | like it shouldn't search shares to you | 20:42 |
| ralsina | dobey: it should also just scan UDFs you are subscribed to in this box | 20:45 |
| ralsina | dobey: so, lots of things to fix | 20:45 |
| mmcc | ralsina: build done, signed, writing a README with instructions on how to file bugs. care to look and tell me if I need to add anything? | 20:46 |
| ralsina | mmcc: wooooohoooo | 20:46 |
| ralsina | sure | 20:46 |
| dobey | ralsina: i presume you meant "shouldn't" there. i would think it should search all the files you've synced to U1, regardless of whether they are synced to the local machine or not. | 20:46 |
| mmcc | ralsina: https://docs.google.com/a/canonical.com/document/d/123__JHLgeuLrloyEtTB0qzhRnpwOcXPT_bz_vqgV0go/edit | 20:47 |
| dobey | well, hooray! | 20:48 |
| ralsina | dobey: to publish, you need to have it locally | 20:48 |
| dobey | just uploaded u1-client-gnome to ubuntu | 20:48 |
| dobey | ralsina: why? | 20:48 |
| ralsina | dobey: because it doesn't work if you don't? | 20:48 |
| ralsina | dobey: although I admit I have not tried it :-) | 20:48 |
| ralsina | dobey: and also because there'sno API to get the list of all the files on the server | 20:49 |
| ralsina | dobey: at least no reasonably quick one | 20:49 |
| dobey | well, ok; so technical limitations aside | 20:49 |
| ralsina | mmcc: looks good to me | 20:50 |
| dobey | the design should be that it lets you publish any of your files that are on the server, regardless of whether they are synced locally or not, and if the API doesn't let us do that yet, we should at least design the UI/code in a way that we can once the API is there; i think | 20:50 |
| mmcc | ralsina: do I need to add any license verbiage anywhere? I actually just realized we might be missing a license screen in the .app because we don't have any installer… (!) | 20:51 |
| mmcc | maybe just add a LICENSE file for now? | 20:51 |
| ralsina | mmcc: yes | 20:52 |
| ralsina | mmcc: add the file | 20:52 |
| ralsina | mmcc: maybe we'll add a way to show the license once per user, the page is there, at least. | 20:52 |
| ralsina | dobey: completion is modular, so we can plug further sources as needed/possible | 20:56 |
| ralsina | dobey: the logical place to publish a file that's on the servers and not in the local box seems to me to be the server though | 20:56 |
| ralsina | dobey: but yes, when it's doable, let's do it | 20:56 |
| mmcc | ralsina: I just remembered --installer. that does exactly what you just said, and it works too | 20:57 |
| ralsina | mmcc: yes but it does it *every time* | 20:57 |
| ralsina | mmcc: so we would need to make it so that the first time it runs it has that set, and the following ones it doesn't | 20:58 |
| ralsina | mmcc: oh, you are RIGHT | 20:58 |
| mmcc | huh, when I first tried it, I didn't see it, I had to delete my creds to see it | 20:59 |
| ralsina | mmcc: next time ;-) | 20:59 |
| ralsina | mmcc: so it only shows the license if you have no credentials. That's pretty much what we want, you are right. | 20:59 |
| dobey | ralsina: well, remember (though may have been before your time), we were also discussing "ghost files" support a long time ago :) | 20:59 |
| ralsina | dobey: yes, before my time | 20:59 |
| ralsina | mmcc: only niggle with --installer: if you delete the creds, you get the license again, not only on startup | 21:00 |
| ralsina | dobey: in fact, I don't know what those would be | 21:00 |
| ralsina | mmcc: I am taking a break to be a dad, go ahead packing and send me the binary, I'll send the announcement late | 21:01 |
| mmcc | ralsina: yes, agreed not ideal. probably will end up writing a file to prefs dir to check or something. for now, want to sneak it in to the release ("if installer or sys.platform == darwin")? | 21:01 |
| ralsina | feel free to test it as much as possible ;-) | 21:01 |
| ralsina | mmcc: nah | 21:01 |
| ralsina | mmcc: let's do it right | 21:01 |
| ralsina | mmcc: this time, just add a LICENSE.txt | 21:02 |
| mmcc | ok, and what's in the LICENSE.txt ? :) | 21:02 |
| mmcc | just the GPL, or ? | 21:02 |
| ralsina | mmcc: yes | 21:03 |
| mmcc | I could copy the text from the --installer pane | 21:03 |
| mmcc | basically just poitns to GPL but includes words about how basic is free but add ons might be paid | 21:03 |
| mmcc | sorry to delay your dad break… I'll just put in the GPL | 21:04 |
| mmcc | ok, uploading alpha 1 now | 21:07 |
| ralsina | mmcc: sure, that sounds good :-) | 21:09 |
| ralsina | mmcc: but put a GPL.txt too then | 21:09 |
| ralsina | we need to include the license. The license requires it :- | 21:09 |
| ralsina | ) | 21:09 |
| mmcc | oh, I just put the GPL in as LICENSE.txt | 21:14 |
| mmcc | here: http://ubuntuone.com/0V2rH0fXsHeUSaPmFk5qn1 | 21:14 |
| ralsina | mmcc: can you rename it? | 21:15 |
| ralsina | mmcc: that's more like 3.99a or something | 21:15 |
| mmcc | oh, sure. | 21:15 |
| mmcc | I thought I had asked about that already, sorry | 21:16 |
| ralsina | dobey: what version is our latest Q release? | 21:22 |
| ralsina | mmcc: we can match that one ^ | 21:23 |
| dobey | 3.99.90 | 21:24 |
| mmcc | so I should make it "3.99.90" and the human readable string "3.99.90 alpha 1"? | 21:27 |
| ralsina | mmcc: sure | 21:27 |
| mmcc | ok will do | 21:27 |
| mmcc | ok, uploading | 21:30 |
| ralsina | yay | 21:30 |
| dobey | alrighty. have a good evening all! | 21:35 |
| ralsina | bye dobey! | 21:35 |
| mmcc | UbuntuOne.app 3.99.90 alpha 1: http://ubuntuone.com/5blGZU9a93kufUxjOLnLxT | 21:38 |
| ralsina | mmcc: awesome! | 21:38 |
| ralsina | mmcc: I'll send the announcement tonight after I test it a bit | 21:39 |
| mmcc | ralsina: sounds good. | 21:39 |
| mmcc | I'm going to take a lunch break now… when I get back I'll do some more testing too | 21:39 |
| ralsina | cool, write here if you find any "new" bugs | 21:40 |
| mmcc | one bug - I noticed the menu says "file sync is disconnected" while the control panel's indicator is correct | 22:56 |
| mmcc | I also just got a new syncdaemon crash, might have a better crash log | 23:02 |
| ralsina | mmcc: release mail sent to warthogs! | 23:47 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!
