=== mmcc_ is now known as mmcc | ||
mpt | Where do I report bugs in the U1 music store? | 09:54 |
---|---|---|
mpt | oh, https://bugs.launchpad.net/ubuntuone-music-store | 09:54 |
gatox | good morning! | 11:11 |
alecu | good morning! | 11:49 |
gatox | alecu, good morning to you sir! | 11:51 |
=== zyga is now known as zyga-food | ||
gatox | alecu, i have a branch stuck in approved since yesterday.... it has a commit message, and i didn't see any reason why it doesn't land...... but it's stuck there.... | 11:53 |
alecu | gatox: I'm sure you've already tried putting it as "needs review" and back to "Approved", right? | 12:04 |
gatox | alecu, no..... i didn't know that trick..... trying... | 12:05 |
ralsina | good late morning! | 12:49 |
=== zyga-food is now known as zyga | ||
mmcc | good morning everyone. | 13:46 |
gatox | brb | 13:46 |
ralsina | morning mmcc! | 13:50 |
gatox | back | 13:53 |
mmcc | Any opinions on the version number for the mac app? For test builds, should we have some per-build number, or include the date in it somewhere? There's both a numerical version number and a localizable string that we can put things like "build 20120808-12" into | 14:37 |
gatox | back..... i'm having some issues with my internet connection | 14:38 |
ralsina | either tests are getting slower every day or I am getting more impatient every time I run them | 14:45 |
ralsina | mmcc: is there a separate version and build number? | 14:45 |
ralsina | mmcc: trunk is supposed to be 3.99.x but something like a date is a better idea at this moment | 14:47 |
gatox | ralsina, i don't know which one it is.... but i'm having the same problem | 14:48 |
ralsina | gatox: it seems to happen every few months, then someone wastes a week making them go fast again | 14:49 |
mmcc | sorry, had to step away for a sec. | 14:51 |
mmcc | ralsina: there's a version number that has to be period-separated digits, then a "version string" that shows up in finder's get info box, that can be anything - convention is that it includes the version number | 14:52 |
mmcc | I just now made it throw the contents of revnos.txt into that version string, so it looks like this: "3.0.0 (ubuntu-sso-client: 988, ubuntuone-client: 1289, ubuntuone-control-panel: 343, ubuntuone-storage-protocol: 159)" | 14:53 |
mmcc | A little messy for distribution but seems handy for testing | 14:54 |
gatox | ralsina, do you know how to run specific qt tests for u1-cp?? i forgot something..... or something change.... i'm trying to do: u1trial --reactor=qt4 path/to/test.py | 14:59 |
briancurtin | me | 15:02 |
alecu | me | 15:02 |
gatox | me | 15:02 |
ralsina | me | 15:04 |
ralsina | thisfred, mmcc: say me | 15:04 |
mmcc | me | 15:04 |
thisfred | ni | 15:04 |
ralsina | gatox: --gui ? | 15:04 |
gatox | ralsina, don't know.... will try | 15:04 |
mmcc | ralsina, gatox it's both - --gui and --reactor=qt4 | 15:05 |
briancurtin | DONE: installer upgrade testing, 1-1, installer sent for testing | 15:05 |
briancurtin | TODO: metaclass/next porting, 2to3 runs to see what else is left | 15:05 |
briancurtin | NEXT: alecu | 15:05 |
gatox | mmmm it's failing with boths..... | 15:05 |
alecu | DONE: reviews, finished a branch for bug #1031197, needing reviews itself | 15:06 |
alecu | TODO: more tests of the ported protobuf with python3, catch up with py3k and briancurtin | 15:06 |
alecu | BLOCKED: no | 15:06 |
alecu | NOTE: I'm taking off this friday | 15:06 |
alecu | NEXT: gatox | 15:06 |
ubot5` | Launchpad bug 1031197 in ubuntuone-client (Ubuntu) "DownloadFinished signal triggers when file has not yet completed writing" [Undecided,Confirmed] https://launchpad.net/bugs/1031197 | 15:06 |
ralsina | gatox: I'll help you after the standup | 15:06 |
gatox | DONE: | 15:06 |
gatox | Propose u1-client branches for sync menu (need reviews on this one: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/ipcmenu/+merge/118621) | 15:06 |
gatox | TODO: | 15:06 |
gatox | Keep working on u1-cp implementation. | 15:06 |
gatox | BLOCKED: | 15:06 |
gatox | No | 15:06 |
gatox | ralsina, go | 15:06 |
ralsina | DONE: calls calls calls (to the tune of Motley Crue's girls girls girls) reviews, some bug triaging, finished contract wrangling (twisted port on the way!) TODO: fix a bug or two, reviews BLOCKED: no NEXT mmcc | 15:06 |
gatox | TODO: 1-1 with ralsina | 15:06 |
mmcc | DONE: cert loc code, reviews, looking at filemonitor bug | 15:06 |
mmcc | TODO: package for test builds, bug filing | 15:06 |
mmcc | BLCK: none | 15:06 |
mmcc | NEXT: thisfred | 15:06 |
thisfred | DONE: cosas conflict resolution TODO: same BLOCKED: 1/2 (API decisions need to be made that I don't feel comfortable making all by myself, but I can continue on other aspects) | 15:06 |
alecu | ralsina: "(twisted port on the way!)" yay! | 15:07 |
ralsina | thisfred: happy to give biased unfounded opinions on API matters | 15:07 |
thisfred | ralsina, in that case see the backlog in the internal u1db channel ;) | 15:07 |
ralsina | thisfred: will do | 15:08 |
ralsina | thisfred: it's LONG though | 15:08 |
thisfred | yep :) | 15:09 |
alecu | mmcc: did you find anything in the filemonitor bug? | 15:11 |
mmcc | alecu, well, I noticed that if you kill and restart syncdaemon, the files get uploaded. Still looking at what might be going wrong. I am suspicious of the RPC errors about on_upload_finished. | 15:12 |
ralsina | thisfred: the idea of the extra argument in query calls feels natural, but it also feels strange in that, when would the user *not* want to know of conflicts? | 15:12 |
ralsina | thisfred: although I am probably misunderstanding because I read it quickly | 15:13 |
gatox | can anyone try this on u1-cp: u1trial --gui ubuntuone/controlpanel/gui/qt/tests/test_filesyncstatus.py ??? and tell me if that works for you? | 15:13 |
gatox | i'm probably doing something wrong.... but i don't know what | 15:13 |
gatox | i thought that should work | 15:13 |
ralsina | gatox: --gui without reactor is not going to wrk for qt tests | 15:14 |
gatox | ohhhhhhh long time without using both!!!! | 15:14 |
gatox | ralsina, thx!!! | 15:14 |
gatox | that's why you are the manager!!!! \o/ jeje | 15:15 |
ralsina | gatox: jeje | 15:15 |
ralsina | gatox: besides, alecu already told you that a ew minutes ago ;-) | 15:15 |
gatox | oh.... totally miss that meesage.... sorry alecu | 15:16 |
ralsina | Any pending reviews? | 15:16 |
gatox | ralsina, yes please | 15:16 |
thisfred | ralsina, right. Problem is we can't do the extra argument, because the queries have a *arg parameter | 15:16 |
gatox | ralsina, this one: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/ipcmenu/+merge/118621 | 15:16 |
thisfred | which means no keyword args allowed | 15:16 |
ralsina | thisfred: ugh | 15:16 |
ralsina | thisfred: ok, no good suggestions then | 15:17 |
alecu | gatox: ralsina, it was mmcc that said it, not me :-) | 15:17 |
thisfred | ralsina, discussing with aq now, but I think we're steering towards just always including the information on whether there are conflicts | 15:17 |
gatox | sorry mmcc then :P | 15:17 |
ralsina | thisfred: that makes more sense | 15:17 |
ralsina | thisfred: like I said, why would the caller not want to know that? :-) | 15:17 |
alecu | mmcc: hmmm... if sd finds the files after being restarted, it sounds like some events were missed the first time. | 15:18 |
* gatox lunch | 15:18 | |
=== gatox is now known as gatox_lunch | ||
mmcc | gatox :) no prob | 15:18 |
ralsina | alecu, mmcc: could be that argument error is making it drop some events? | 15:18 |
thisfred | ralsina, I agree. The only reason it might not want to know if the app never syncs. Then why use u1db ;) | 15:18 |
ralsina | thisfred: and willful ignorance is easy to achieve, just ignore the data ;-) | 15:19 |
alecu | mmcc: that happesn because after the restart sd has a chance to rescan the filesystem. | 15:19 |
thisfred | and it's easy to ignore, you'd just pay a tiny performance price | 15:19 |
thisfred | right | 15:19 |
mmcc | alecu: on the first run, it sees the file create, it generates a MakeFile, but no corresponding Upload log entry ever shows up. | 15:19 |
alecu | ralsina: which "argument error" ? | 15:19 |
mmcc | alecu, right- the restarted daemon sees the files in the local rescan | 15:19 |
alecu | mmcc: oh, ok. Were you able to reproduce that consistently? | 15:20 |
ralsina | alecu: the "requires 2 arguments, received 3" ior something, it was in mmcc's log | 15:20 |
ralsina | alecu: it was on on_upload_finished, IIRC | 15:20 |
* alecu needs fosfovita | 15:20 | |
mmcc | alecu, ralsina: 2012-08-07 12:12:46,689 - ubuntuone.SyncDaemon.Pb - WARNING - Could not emit signal 'on_upload_finished' to <twisted.spread.pb.RemoteReference instance at 0x10512ad40> due to 'on_upload_finished() takes exactly 2 arguments (3 given)' | 15:20 |
ralsina | OMG *my* fosfovita worked :-) | 15:21 |
ralsina | translation: fosfovita is a minerals/vitamins pill that claims to improve memory because it has phosphorus | 15:21 |
* ralsina would rather lick a matchbox | 15:21 | |
mmcc | mmmmmm phosphorus | 15:22 |
alecu | ralsina: remember to lick the matches, not the box! | 15:22 |
ralsina | alecu: all the phosphorus on safety matches is in the box :-) | 15:22 |
ralsina | alecu: that's why you can't light them in your boot like a cowboy | 15:22 |
alecu | ralsina: then why do we still call them "fósforos" ? | 15:23 |
ralsina | alecu: tradition | 15:23 |
* alecu tries licking a lighter | 15:23 | |
ralsina | alecu: and because "little sticks with some thing that burns on the tip but is not phosphorus" is not marketable | 15:23 |
mmcc | ralsina: you've heard of "I can't believe it's not butter," right? | 15:24 |
ralsina | mmcc: I even tried it! | 15:24 |
ralsina | mmcc: I totally believed it! | 15:24 |
alecu | mmcc: I don't think that error has anything to do with the missed events problem. | 15:24 |
ralsina | mmcc: http://boingboing.net/2012/07/04/walmarts-i-cant-believe.html | 15:25 |
mmcc | alecu - ok, so it just gets logged and ignored, and doesn't leave anything wedged? | 15:25 |
ralsina | alecu: the other possibility that comes to mind is a semantic difference or out-of-order events | 15:26 |
ralsina | alecu: which is scary | 15:26 |
alecu | mmcc, ralsina: right | 15:26 |
alecu | ralsina: remember that macfsevents does not get events for individual files, but for some changes in a folder | 15:27 |
ralsina | alecu: then maybe there is a bug in the code to figure out the diff | 15:27 |
alecu | ralsina: and it's synthezising (wow, that was hard) some events from the changes it sees. | 15:27 |
ralsina | right | 15:27 |
alecu | hmmm\ | 15:28 |
alecu | but, mmcc says that the event was received. And it's just the Upload that's missing.... | 15:29 |
mmcc | alecu - well, I'm not sure it's just the upload missing - I don't know what *should* be there. here's the log: https://pastebin.canonical.com/71687/ | 15:29 |
ralsina | if there is a makefile, and no upload, maybe it thinks it's still open | 15:29 |
mmcc | I'm going to go log in to my u1 account in ubuntu and try the same thing and see what log messages I get so I know what's up | 15:30 |
ralsina | and/or the nanny is preventing the upload | 15:30 |
mmcc | ralsina: if the nanny was blocking it, would there be a log message? there's no 'nanny' string in the log paste | 15:30 |
ralsina | mmcc: no | 15:30 |
briancurtin | alecu: if you have a chance, would you mind running the tests on lp:~brian.curtin/ubuntu-sso-client/py3-unicode-part-3 to see what you get. this is that SyncTimestampChecker that is failing | 15:33 |
mmcc | IIRC you guys did a sprint a while back with presentations on syncdaemon, right? are those presentations around anywhere for me to refer to? | 15:33 |
alecu | mmcc, it was more like a workshop, and there were no slides, sorry. | 15:35 |
mmcc | alecu: ok. | 15:36 |
alecu | we should definitely record it next time | 15:36 |
briancurtin | mmcc: i have the hand-written posters and diagrams if you want them, but like alecu said, it was more of a workshop. there was a lot of talk going along with reading code | 15:36 |
mmcc | briancurtin: hand-written posters and diagrams sounds great, got to be better than nothing | 15:36 |
briancurtin | mmcc: i'll see if i can take good pictures of them or something | 15:37 |
mmcc | briancurtin: oh, I thought you already had pics :) ok, but take your time - and don't worry too much about quality | 15:38 |
alecu | briancurtin: here, right? test_server_date_sends_nocache_headers | 16:17 |
briancurtin | alecu: yep, that's the one | 16:17 |
alecu | briancurtin: I get the same 500. I'll take a further look. | 16:17 |
briancurtin | alecu: the odd thing is that i do what that test does in the interpreter just fine. i can also use urllib to hit that URL just fine. it only seems to fail when run in the test suite :/ | 16:18 |
alecu | briancurtin: what is the url? | 16:18 |
ralsina | lunchtime | 16:19 |
briancurtin | alecu: "http://one.ubuntu.com/api/time" | 16:19 |
alecu | briancurtin: I think we should not be hitting that from the tests. We should be using a mock webserver if needed. | 16:20 |
briancurtin | alecu: i don't disagree, but this should work. it works fine on trunk | 16:20 |
alecu | briancurtin: self.patch(SyncTimestampChecker, "SERVER_URL", self.ws.get_iri()) | 16:21 |
briancurtin | it also works if you just open up an interpreter, import that, and call get_server_time just fine | 16:21 |
alecu | briancurtin: so, we are not using that url in the test, as it's being patched. | 16:21 |
briancurtin | alecu: ah | 16:22 |
alecu | briancurtin: one more thing: I *think* this code is not used anymore. | 16:25 |
alecu | briancurtin: oauth_headers and the timestamp_checker module instance of SyncTimestampChecker | 16:26 |
alecu | briancurtin: right. I think I forgot to get rid of the SyncTimestampChecker when it was no longer needed. I surely seem to have gotten rid of the AsyncTimestampChecker. | 16:27 |
briancurtin | alecu: ah, well then that makes this easier. im doing a double check that we dont need it. i think you're right | 16:28 |
alecu | briancurtin: I've done a quick grep in u1-client and u1cp and they are not used either. | 16:29 |
=== gatox_lunch is now known as gatox | ||
briancurtin | alecu: cool | 16:29 |
alecu | briancurtin: I think we should do a branch to delete both, and see if anything breaks. And if it breaks, it should be using the new webclient anyway. | 16:30 |
briancurtin | alecu: i'll go that route and take SyncTimestampChecker out outside of the unicode branch | 16:31 |
alecu | this netsplitting is just awesome. | 16:46 |
alecu | briancurtin: don't know if this reached you: | 16:46 |
alecu | <alecu> briancurtin: I think we should do a branch to delete both, and see if anything breaks. And if it breaks, it should be using the new webclient anyway. | 16:46 |
alecu | <alecu> briancurtin: would you mind creating a bug for this and assigning it to me? | 16:46 |
alecu | <alecu> briancurtin: you can skip those fixes in this branch. Thanks for finding it! | 16:46 |
briancurtin | alecu: now i got it, and i'll create the bug shortly | 16:47 |
alecu | thanks :-) | 16:47 |
gatox | alecu, dobey is on vacation? | 16:48 |
alecu | gatox: calendar says he returns tomorrow | 16:49 |
alecu | hey! it's joshuahoover's birthday today! | 16:49 |
gatox | alecu, ok..... usually he knows with the branches are not landing :P | 16:49 |
alecu | gatox: we should probably ping sidnei about tarmac when dobey is not around | 16:50 |
alecu | lunch time for me. | 16:53 |
ralsina | gatox: 1-1? | 17:32 |
gatox | ralsina, right! | 17:32 |
gatox | mumble or irc? | 17:33 |
ralsina | mumble | 17:33 |
gatox | ralsina, ack | 17:33 |
mmcc | ok, early lunch here. back in a bit | 17:39 |
ralsina | mmcc, briancurtin, gatox, alecu: any reviews pending? | 18:08 |
gatox | ralsina, nop | 18:09 |
briancurtin | none at the moment | 18:09 |
alecu | ralsina: https://code.launchpad.net/ubuntuone/+activereviews | 18:09 |
gatox | ralsina, ahhhh yes, tihs one: https://code.launchpad.net/~diegosarmentero/ubuntuone-client/ipcmenu/+merge/118621 | 18:09 |
ralsina | alecu: prefiero el human touch ;-) | 18:09 |
gatox | ralsina, i thought you already approve it | 18:09 |
alecu | ralsina: https://code.launchpad.net/~alecu/ubuntuone-client/fix-download-finished/+merge/118759 | 18:09 |
briancurtin | alecu: i created bug #1034528 for that SyncTimestampChecker. working on it right now | 18:09 |
ubot5` | Launchpad bug 1034528 in Ubuntu Single Sign On Client "Remove SyncTimestampChecker" [Undecided,New] https://launchpad.net/bugs/1034528 | 18:09 |
ralsina | gatox: no, seen it but not reviewed it entirely yet | 18:09 |
ralsina | alecu: got it | 18:10 |
gatox | ralsina, ahh ok | 18:10 |
alecu | briancurtin: ok, great. I can work on that if you want to keep going over the unicode branches. | 18:10 |
ralsina | gatox: will +1 as soon as I run the tests | 18:10 |
gatox | ralsina, ack | 18:10 |
* alecu has just conflated py3 and unicode. | 18:10 | |
alecu | I wonder how much of that is true :P | 18:11 |
briancurtin | alecu: it seems easy enough. i just started on it a few mins ago | 18:11 |
alecu | briancurtin: great then. | 18:11 |
ralsina | alecu: unicode in python2 is gatox, so you can conflate them only in some cases ;-) | 18:12 |
ralsina | grrrr I have to check the assertIn argument order every time I run into it | 18:19 |
ralsina | yes, it's arg1 in arg2 but then that reads "assertIn needle, haystack" which makes my eyes stumble | 18:19 |
alecu | ralsina: I always remember that bit (thanks fosfovita), but I never type it right. | 18:28 |
ralsina | gatox: +1 | 18:29 |
gatox | ralsina, thx | 18:30 |
ralsina | alecu: have to love a 5-line patch with 55 lines of tests, right? :-) | 18:31 |
alecu | lol | 18:31 |
alecu | I had to refactor one of the tests, that's why :-) | 18:32 |
alecu | and yes... 11:1 is too high. Usually it's 3:1 or 4:1 | 18:33 |
ralsina | at least the problem seems well contained :-) | 18:35 |
ralsina | alecu: +1 ran tests only on linux though | 18:35 |
alecu | thanks | 18:36 |
mmcc | ralsina: I have another high test-to-code branch for you: https://code.launchpad.net/~mikemc/ubuntuone-storage-protocol/fix-1025950-cert-locs/+merge/118603 | 18:50 |
mmcc | (because there weren't tests before) | 18:50 |
ralsina | mmcc: got it | 18:50 |
ralsina | mmcc: he | 18:50 |
ralsina | mmcc: storage protocol is mostly very old code before we got religion | 18:50 |
gatox | ralsina, as some parts of u1-client....... | 18:53 |
ralsina | gatox: on some parts of u1-client I expect to see *mold* | 18:54 |
gatox | jeje | 18:54 |
* mmcc just wrote "import tracebacon". | 19:06 | |
beuno | mmcc, did it work? | 19:08 |
alecu | it's just a whiff away | 19:11 |
mmcc | sadly, no… it did not work. | 19:16 |
briancurtin | alecu: so i'm looking at this SyncTimestampChecker removal...it doesn't look like TimestampChecker is intended to be used by itself, so it doesnt seem as easy as just replacing its use | 19:26 |
briancurtin | unless im misunderstanding it | 19:27 |
briancurtin | it looks like its consumed by BaseWebClient, but i dont think the places which used a timestamp checker want a BaseWebClient (or other web client) | 19:27 |
mterry | Are U1 oauth credentials stored in my home directory? | 19:28 |
mterry | Ah, in the keyring I bet | 19:30 |
gatox | alecu, when you have a moment: https://code.launchpad.net/~diegosarmentero/ubuntuone-control-panel/refactor-sync-status/+merge/118809 - https://code.launchpad.net/~diegosarmentero/ubuntuone-client/ipcmenu/+merge/118621 | 19:33 |
alecu | hey all, ralsina has just smsd me: his house ran out of power | 19:43 |
briancurtin | that sucks | 19:43 |
gatox | :S | 19:44 |
alecu | briancurtin: afaik SyncTimestampChecker is not used anymore. | 19:45 |
alecu | briancurtin: only TimestampChecker is used nowadays. | 19:45 |
briancurtin | alecu: there were a few things that seem to have been using SyncTimestampChecker this whole time. they were getting it from utils.timestamp_checker, which pointed to an instance of SyncTimestampChcker | 19:46 |
alecu | briancurtin: I can't find any usage of that timestamp_checker other than the tests for it | 19:54 |
alecu | briancurtin: in ussoc, in u1c nor in u1cp | 19:54 |
alecu | and u1sp | 19:54 |
alecu | briancurtin: in which module did you find that usage? | 19:55 |
briancurtin | alecu: maybe i'll post the branch in a minute. i removed STC and its tests, but things are still depending on utils.timestamp_checker | 19:55 |
alecu | ack | 19:55 |
briancurtin | alecu: i guess it's just line 207 in ubuntu_sso/utils/tests/test_common.py - it patches that utils.timestamp_checker (which would be SyncTimestampChecker) | 19:58 |
briancurtin | alecu: and the oauth stuff in ubuntu_sso.utils then uses timestamp_checker, which is the SyncTimestampChecker | 19:59 |
alecu | briancurtin: right. And that oauth_headers method is not used anymore either. | 20:01 |
alecu | briancurtin: I rolled up both into the new webclient. | 20:02 |
briancurtin | alecu: alright, i'll remove that then see what happens | 20:02 |
alecu | briancurtin: previously every webclient user would check the timestamp and would sign oauth. | 20:02 |
alecu | briancurtin: with the new one, this is done automatically by the webclient. | 20:03 |
briancurtin | alecu: so i should then remove the entire SignWithCredentialsTestCase? because that was still using oauth_headers | 20:04 |
alecu | briancurtin: it looks like it should be gone, yes. | 20:11 |
briancurtin | alecu: ah there we go, i think it's good now | 20:13 |
alecu | great | 20:13 |
briancurtin | i guess that was easier than i thought...it seemed like a whole bunch of important stuff to just be ripping out | 20:14 |
gatox | eod here!!! bye all! | 20:16 |
ralsina | hello again, sorry, was out of power & internet | 20:20 |
=== mbarnett` is now known as mbarnett | ||
alecu | hello back there, boss. | 20:28 |
briancurtin | alecu: when you have a minute: https://code.launchpad.net/~brian.curtin/ubuntu-sso-client/remove-SyncTimestampChecker/+merge/118826 | 20:49 |
alecu | briancurtin: looking | 20:55 |
alecu | briancurtin: small needsfixing | 20:57 |
alecu | briancurtin: we should also remove the unused imports | 20:57 |
alecu | ok, EOD here. | 21:00 |
alecu | cheers, all! | 21:00 |
ralsina | bye alecu! | 21:12 |
ralsina | EOD for me as well. bye! | 21:22 |
mmcc | ok, I'm done for today. see you all tomorrow | 22:56 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!