=== facu is now known as facundobatista [02:11] hello :> trying out ubuntu one in 9.10 again...i've noticed that although it seems that i can paste files into the "ubuntu one" folder and have them update successfully, when i connect to the website using my launchpad account, the files don't show up :< [03:08] Good day (night) all. I have a question to developers. Do you plan to implement the use of symbolic links in a synchronization directory? [13:56] OH HAI [13:58] o hai jblount [15:00] desktop+ meeting begins. aquarius, CardinalFang, dobey, jblount, mandel, rodrigo, rtgz, teknico, urbanape, vds: you know how it works: say "me" to get a turn, then in your turn say your done/todo/blocked status. [15:00] me [15:00] * rtgz wants meating [15:01] rtgz: say "me" then [15:01] me [15:01] me [15:01] me [15:02] humbug! [15:02] me-ish [15:02] me [15:03] * Chipaca suspects rmcbride wants to have a little chat with rtgz [15:03] well I may have a question :) [15:03] rmcbride, I did not make any segfaults today, I swear! [15:04] Nope, no segfaults. I'll get to my question when it's my turn :) [15:04] dobey: urbanape: ping [15:04] dobey is at the post office [15:05] um, rain [15:05] I'm in the patio [15:05] bbiab [15:06] Well then, let's go! [15:06] DONE: JS for account-assistance, thinking about public files in the web ui, chasing layout bugs [15:06] TODO: #498284, #451649, maybe some hacking on public files stuff, need to do my "bug day" duty, had slow internet at the coffee shop yesterday [15:06] BLOCKED: Nope [15:06] aquarius: Go! [15:06] ⚀ DONE: submitted initial fake music store and music store for review; worked out file delivery API [15:06] ⚁ TODO: package rhythmbox plugin; work out why decorators don't work on HttpResponseRedirects; allow OAuth to web UI; make music store views better; make workitems of outstanding todo items; make tomboy first-sync experience nicer [15:06] ⚂ BLOCKED: [15:06] rtgz, you're up [15:06] DONE: Tested UI on Maemo Microb and Android browsers (in emulator). [15:06] TODO: Update ubuntuone-client-diagnose with the latest bug info: proxy detection, auth loop due to AUTH_FAILED status. Prepare branch with Important emblem for u1conflict files. Hack "Shared with Me" folder emblem support. [15:06] BLOCKED: Need to get a fur-tree [15:06] jmp vds [15:07] DONE: branch to fix #498324 landed, working on branch to configure funambol to send sms messages #418048 [15:07] TODO: finish this branch to setup funambol to send sms messages [15:07] BLOCKED: no [15:07] CardinalFang: > [15:07] me [15:08] rmcbride: you go, then [15:09] CardinalFang is scanning local metadata, I suppose... [15:09] OK [15:09] DONE: Face, Reviews, chat about instrumenting server, worked on Bug #499301 and Bug#499595. TODO: catch up on bug-triage from yesterday. Pick one to work on. [15:09] BLOCKED: None [15:09] Bug 499301 on http://launchpad.net/bugs/499301 is private [15:09] I don't have a DONE: etc block today but I do have a questioni [15:09] Sorry. Juggling stuff here. [15:09] rtgz: I'm getting a build failure when doing client packages that I think you might have an idea on: [15:09] I get make[3]: *** No rule to make target `ubuntuone-marshallers.list', needed by `ubuntuone-marshallers.c'. Stop" [15:09] * statik makes the bug public [15:10] working #499850 to fix. [15:10] statik, er, which bug? [15:10] me [15:10] CardinalFang: bug 499301 [15:10] Bug 499301 on http://launchpad.net/bugs/499301 is private [15:11] ☺ DONE: Face day, Mailed motu-council, Reviews, New client code, Bug day [15:11] ☹ TODO: New Client Code [15:11] ☹ BLCK: None. [15:11] rtgz: basically it looks like some of the marshaller stuff isn't getting picked up during make dist. Thought perhaps it might be something you knew of a fast fix for, otherwise I'm going to pester dobey :) [15:11] rmcbride: hrmm [15:11] dobey: FWIW make distcheck fails the same way [15:12] rmcbride: distcheck will fail anyway for other reasons, but i'll check it [15:12] dobey: awesome [15:12] rmcbride, hm, are there any PPA build logs? [15:13] rtgz: yea one moment. [15:13] * rmcbride hugs notes sync [15:13] I don't think there are others in line, so I'll go now: [15:13] DONE: Brute forcing the checking of manifest differences. [15:13] TODO: Get CSS done on ubuntuone-servers branch and land it. Gonna be off this afternoon and tomorrow. Still need to make bugs for my plans. [15:13] BLOCK: None [15:14] and, that's it [15:14] i would like to make https://bugs.edge.launchpad.net/ubuntuone-servers/+bug/499863 public, launchpad won't let me. sorry world. [15:14] hmm that change didn't sync [15:14] Error: This bug is private [15:14] thanks all! desktop+ meeting ends [15:15] rmcbride, basically it may mean that .generated Makefile is somehow broken. And I guess that this is the high time for me to understand how PPAs are built... :) [15:15] statik: bug #499863 is public now [15:15] http://launchpadlibrarian.net/37098535/buildlog_ubuntu-karmic-amd64.ubuntuone-client_1.1.0%2Br298-0ubuntu1~ppa1~karmic_FAILEDTOBUILD.txt.gz [15:15] Launchpad bug 499863 in ubuntuone-servers "need a self-contained 503 error page for haproxy to serve during site maintenance" [High,Triaged] https://launchpad.net/bugs/499863 [15:15] rtgz: that was my understanding. It's been a long time since I've worked with c code building to any degree [15:16] i have a feeling i'm going to have to go back to the post office tomorrow morning too [15:16] grmbl [15:17] rmcbride, building revision 298 from lp:ubuntuone-client, right? [15:17] you'd think i was living in some country where the postal service is more a guarantee that your mail won't get delivered [15:17] dobey: ooh, i wonder if i couldn't see the edit widget for making the bug public because i'm using chrome [15:17] rtgz: that's correct [15:18] statik: maybe [15:18] statik: maybe in https://bugs.edge.launchpad.net/ubuntuone-servers/+bug/499863/+edit ? [15:18] Ubuntu bug 499863 in ubuntuone-servers "need a self-contained 503 error page for haproxy to serve during site maintenance" [High,Triaged] [15:18] rtgz, rmcbride: i think i have a fix already [15:18] rtgz: there's a bit of magic involving pulling in the debian info from another branch involved in building actual packages, but since the generated makefiles appear to be the issue, that probably isnt' necessary. [15:18] dobey: awesome [15:19] rmcbride: is there a bug for this problem? [15:19] dobey: bug #499850 [15:19] Launchpad bug 499850 in ubuntuone-client "PPA test packages fail with "make[3]: *** No rule to make target `ubuntuone-marshallers.list', needed by `ubuntuone-marshallers.c'. Stop."" [Critical,In progress] https://launchpad.net/bugs/499850 [15:19] rmcbride: great, thanks [15:20] bbiab, shopping [15:20] dobey: sure thing. I kind of grokked what the issue was when I figured it out, but was struggling with the appropriate fix. I haven't dealt with many C-based build problems since my gentoo days many years ago [15:21] statik: are you using Chromium now? [15:21] also ignore my make distcheck comment, since you say it doesn't work anyhow [15:21] rmcbride: yeah, i added a comment to the bug explaining that :) [15:21] urbanape: yep, i switched at UDS. using the daily PPA on lucid [15:22] the making bugs public thing doesn't work with WebKit. [15:22] it's a known issue [15:22] dobey, my distcheck fails with "cp: cannot stat `./ubuntuone-icons.rendercache': No such file or directory"... [15:22] the button doesn't show up. [15:23] urbanape, thanks! yep, firefox let me make it public [15:23] i got caught by that, too. [15:24] hm, i don't want it to fail due to my dear marshallers... [15:24] oh nice [15:24] it does show up, but then gets hidden [15:24] with webkit [15:25] how does the design team use launcpad then? surely safari has this problem ;) [15:25] haha [15:25] Heh [15:25] dobey, they... don't mark bugs as public/private? [15:26] rtgz: i would hope they do, since some of them are working on the LP UI :) [15:27] anyway [15:27] https://code.edge.launchpad.net/~dobey/ubuntuone-client/marshaller-dist/+merge/16541 [15:27] rmcbride: ^ that fixes your issue :) [15:28] dobey, THANKS! [15:29] dobey: hi [15:29] totally slipped my mind that was required, when reviewing your branch [15:30] verterok: hi. aren't you supposed to be enjoying the summer or something? [15:30] dobey: I am! :) [15:30] verterok: what's up? :) [15:31] dobey: about to go to the market to get supplies for the asado ;) [15:31] ooh [15:31] i want some asado :) [15:31] rmcbride, could you please give some hints as per how PPA for ubuntuone-client get built? I.e. there is trunk - the code, it gets pulled from bzr. Where does debian/ directory come from then? [15:31] dobey: I was wondering if you got a chance to do the automagic-thingy for the logging stuff [15:31] dear rain: you're hurting my wifi. Please stop. [15:32] verterok: i didn't. but your branch landed yesterday, so i'll look at it today :) [15:32] dobey: I'll propose a branch to change the debug level to debug as lucio requested to have debug logging in the ppa, we can make it automagic later, are you ok with this? [15:32] s/debug/logging/ [15:32] rtgz: I have a script that automates most of that. I have to go AFK for a bit but I'll follow up with you later. [15:33] rtgz: I think I need to move the branch somewhere public and I can't do that right this second [15:33] rmcbride, ok, will wait for you here :) [15:33] and dobey thanks for the fix [15:48] verterok: yeah that's fine [15:49] verterok: could you perhaps actually set the review on my branch to approve please? :) [15:49] verterok: looks like you forgot to change the drop-down :) [15:52] dobey: heh, sure [15:53] thanks :) [17:17] dobey, could you please check https://code.edge.launchpad.net/~rtgz/ubuntuone-client/stop-killing-thumbnails/+merge/16315 again, it stays in Needs Fixing just because it depends on a branch that is already landed :) [18:16] rtgz: yes was planning to do that too :) [18:16] * rtgz switches off "annoying" mode === CardinalFang changed the topic of #ubuntuone to: https://one.ubuntu.com | https://launchpad.net/ubuntuone | Please honk if you want a music store [18:34] Deglymphte, hi. I'm not officially "on call" any more, but I'll answer what I can. What's up? [18:35] brrrr [18:36] rmcbride: are you rebuilding nightlies now? [18:36] dobey: yep [18:37] rmcbride: great. poke me if there are any more issues like that :) [18:37] dobey: will do :) [18:47] Heya Cardinal, thanks. I have an issue with notes in Ubuntu one, when I log into the web interface and type a note, after ca 10 seconds this warning appears: There was a problem with saving your note! [18:47] and everything that's typed afterwards is not saved anymore [18:49] Did you already reauthorize your account? (In the past week or so?) [18:50] vergil66: that only affects using the tomoby syncing. using the web site should still work regardless of that :) [18:51] Okay...I'll slink away ... :) [18:53] it's just strange that everything works fine the first 10 seconds or so [18:53] until that note appears [18:57] Just tried it also...got the same message. [19:00] Cheers. So looks like this could be a general issue. [19:03] hm, the g_object_weak_unref looks reasonable.. [19:05] rtgz: yeah, i'm not sure if it will cause issues though [19:05] since it would cause the unref callback to try to remove the object from the hash table... which has already happened [19:05] but should be ok i think [19:06] basically, this is not visible since _finalize gets called only when nautilus shuts down, i guess... [19:07] yeah [19:07] it's better to clean that up, since if I put the callback there then I am responsible to take it from there as well, thanks for the review! [19:07] dobey, ^ [19:09] rtgz: sure. i don't think i've ever used weak refs before, so just making sure we get the memory mgmt right [19:50] grrr, the g_hash_table_new_full's GDestroyNotify does not allow uon object to be passed in, therefore I can not remove the weak pointer (callback routine address and user_data is matched) :-/ [19:51] of course, I can stuff in the uon object into the observed GHashTable as an attribute... [19:53] hrmm [19:53] oh, does weak unref require that too [19:54] oh it requires the callback and the object [19:54] dobey, yep, it would be OK if the object was NULL but we cannot afford that, since then it will be useless. [19:55] if it's already NULL it doesn't matter :) [19:55] and on the second thought even if I stuff uon object into the hash table, then ... I need to get the reference to the hash table using one of hash table values only, and I can not do that... Other than stuffing uon object into the NautilusFileInfo, which is.. .weird [19:55] hrmm, i guess it doesn't matter [19:56] dobey, okay, i will reverse the weak ref assignment, hash table insert for the time being... [19:57] rtgz: well, it's only going to happen when nautilus is shutting down, and at that point the objects are being unreffed anyway [19:57] the weak ref doesn't keep the object from being destroyed i don't think [19:58] dobey, yup, it does not prevent the object from being destroyed... And we only keep the hash value which is only a pointer... [19:58] * rtgz is checking whether finalize is actually called... [19:59] if the weak notify is called, then finalize was already called on the object [19:59] according to the docs [20:01] dobey, no, the ubuntuone_nautilus_finalize [20:01] dobey, btw, how exactly should nautilus behave for Shared With Me folder links? [20:05] rtgz: ubuntuone_nautilus_finialize gets called at exit (when the plug-in gets unloaded) [20:06] rtgz: how do you mean re: shared with me? [20:07] dobey, i mean since we are requesting syncdaemon to provide us with the info for /home/someone/Ubuntu One/Shared With Me/share from someone/file we will not get any response [20:07] either we or syncdaemon should translate this into the real path [20:09] and when we receive response, then 1) patch is needed for nautilus to get NautilusFileInfo object for the real file path and apply the emblem to the target file (it gets picked up by Nautilus then), or 2) Translate the path by s{user/.local/share/ubuntuone/shares}{Ubuntu One/Shared With Me/}g. [20:10] * rtgz does not like second variant [20:10] I mean I no longer like that, I used to like that :) [20:11] hm, cannot make our finalize routine to get called :-/ [20:16] ran under gdb, nope, does not want to get called :-( [20:20] heh [20:20] the nice thing about utime() is that it already works in this case :) [20:21] hrmm, not sure why you aren't getting the finalize call [20:21] it's obvious we're getting it called, since in the 0.95/96 range of releases, we got some bug reports about crashes in finalize :) [20:21] dobey, O_O [20:22] verterok: hrmm, i think i will need your help with the logging config changes [20:22] verterok: to split the logging config to a separate conf file [20:24] configglue is complicated :) [20:31] dobey: :) [20:32] i wonder how much else i can really get done today [20:32] dobey: for syncdaemon POV, config.get_config_files() should return an extra item between system and user config file, with the path to this global logging config file [20:33] hm [20:33] verterok: i got that far :) [20:33] verterok: i'm confused about whether syncdaemon will write out a single syncdaemon.conf with the logging values though, after i move them to a different file [20:34] dobey: the user config? [20:34] hmm, good point [20:34] it looks like it combines all incoming config into one thing, and then writes it all to a single file when writing out [20:35] dobey: actually, the _Config class don't write any logging stuff to syncdaemon.conf [20:35] since each file isn't a ConfigParser object, but just a string of the path, afaict [20:35] dobey: only throttling values [20:35] verterok: ok, so it should be safe? [20:35] verterok: i'm also worried about how moving that config interacts with the command line option for it [20:36] dobey: yes, as syncdaemon doesn't handle runtime changes to any other config values (only throttling) [20:36] ok, cool [20:36] dobey: if get_config_files returns: [system_config, system_logging_config, other_configs] the CLI should be ok [20:37] dobey: configgule will merge all the files and build the OptionParser for them [20:37] verterok: ok, great [20:37] i'll get my branch together and test it then [20:38] dobey: ok, I don't know if I'm going to be online later, but I'll take a look during the holidays [20:40] verterok: yeah. i'm not exactly expecting that a whole lot more will get landed or such today :) [20:42] erm [20:42] dobey, ubuntuone_nautilus_class_init does not get called now [20:43] rtgz: then the module isn't getting loaded [20:43] dobey, yup, but it is happily assigning emblems [20:43] something is reeeally really wrong, cleaning up my system... [20:43] rtgz: i'm not sure what you're doing then [20:44] :) [20:46] okay, now only finalize does not get called :) [20:48] hm [20:48] looks like ubuntuone extension is not immediately initialized, on nautilus startup (unlike terminal, thunar, etc...) [20:49] * rtgz sets -v on himself [20:49] it's all ok, init happens as needed [20:49] heh [20:50] thunar? [20:50] there's a nautilus thunar extension? [20:50] file-roller, sorry, XFCE background :( [20:51] heh [20:51] okay, so, in the end only finalize do not get called and there is something weird when nautilus is started - no emblem is applied to Ubuntu One folder, requires refresh for it to appear. But finalize does not get called :'-( [20:53] i found the bug #406788 describing the crash on nautilus exit but I can't understand HOW should I exit nautilus even more than nautilus -q (ok, killing does not count) [20:53] Launchpad bug 406788 in ubuntuone-client "nautilus crashed with SIGSEGV in g_type_check_instance_is_a()" [Critical,Fix released] https://launchpad.net/bugs/406788 [20:55] rtgz: log out [20:56] good bye, cruel world! [21:02] dobey, no, finalize does not get called. Shutting down nautilus-open-terminal extension\nShutting down nautilus-gdu extension\n, but no my message [21:04] ok, I will check other extensions and will get back once I get something interesting [21:05] (GClassInitFunc) nautilus_open_terminal_class_init, [21:05] (GClassFinalizeFunc) nautilus_open_terminal_class_finalize, [21:09] rtgz: you haven't added any g_print() or something to the ubuntuone finalize have you? [21:38] verterok: i have no idea how these tests are passing now... looks like they expect log_level to be 10, and not "DEBUG" [21:40] and ... nautilus_module_shutdown is not called as well... great, awesome. I had 1 feature to add, now I have 2 bugs to investigate :) [22:09] rtgz: thanks for all the work you've been doing! [22:09] i'm off for the year but i might poke in every once in a while to check that stuff is not going insane :) [22:09] later! [22:09] dobey, thank you! [22:10] dobey, have a nice year finalization :) [22:10] you too :)