[09:02] <JamesTait> Good morning all! :)
[09:04] <rye> rodrigo_, hi, I've spent some time this weekend trying to make gir bindings for libsyncdaemon  work with python and tried to annotate signals. Having spent nearly an hour I found bug #631058 :)
[09:04] <ubot4> Launchpad bug 631058 in ubuntuone-client "libsyncdaemon source uses tabs and breaks gtk-doc and gobject-introspection (affects: 1) (heat: 6)" [Undecided,New] https://launchpad.net/bugs/631058
[09:05] <rodrigo_> rye, looking
[09:05] <rodrigo_> rye, ugh
[09:06] <rye> rodrigo_, i.e. the source uses tabs, but if I add comments aligned with the code that is registering the signals - it silently skips that block :(
[09:06] <rye> rodrigo_, i am still thinking about connect() but with gir rename function we may even stop worrying about renaming anything in the sources, but I have yet to succeed to return GSList from got-published-files signal - it is now just an opaque gpointer
[09:07] <rodrigo_> rye, ah, cool  about the renaming
[10:21] <rodrigo_> rye, in which .h file do you see the tabs? I have my emacs configured to use spaces for tabs, so it might be just in one spot
[10:22] <rodrigo_> although, gtk-doc and g-i are indeed too fragile if they get confused by this
[10:59] <rye> rodrigo_, syncdaemon-daemon.c
[11:00] <rye> rodrigo_, if i put signal annotation there preserving the indentation - it is skipped; however I am still not able to force gi-scanner to treat gpointer as GSList
[13:04] <rye> anybody here experiencing bug #629517 ?
[13:04] <ubot4> Launchpad bug 629517 in ubuntuone-client (Ubuntu) "Can't log in: required argument 'password' (pos 2) not found (affects: 4) (heat: 20)" [Undecided,Confirmed] https://launchpad.net/bugs/629517
[17:55] <md-llyr> when I invoke system>preferences>Ubuntu One no account info is shown, the only machine shown is <LOCAL MACHINE>, and any attempt to connect just does nothing. When I login to the website my computer is shown unde "Computers on your account." Can someone point me in a direction to resolve this?
[17:56] <rye> md-llyr, are you running Maverick Meerkat beta?
[17:57] <md-llyr> No. Sorry, I should have pointed out this is 10.04 and is completely up-to-date
[18:05] <rye> md-llyr, could you please log into https://one.ubuntu.com/account/machines and see whether your machine is listed. Another question - are you running nightlies release or stock ubuntuone-client version shipped in lucid-updates?
[18:12] <md-llyr> rye: I'm already logged in there and this computer is shown there, with its correct name, under "Computers on your account."
[18:13] <md-llyr> rye: I'm running the stock client.
[18:13] <rye> md-llyr, may I ask you to check what's the status of syncdaemon - could you please run u1sdtool -s in the terminal?
[18:16] <md-llyr> rye, is there a particular part of the output that I should report back on? I'm an Ubuntu newb
[18:16] <rye> md-llyr, you can copy & paste the output to http://paste.ubuntu.com and give the link to the paste here
[18:16] <md-llyr> rye: will do. One moment, please
[18:19] <md-llyr> rye: I thinnk I've done this correctly. . . output is at http://paste.ubuntu.com/489346/
[18:21] <rye> md-llyr, ok, just to test the token - could you please run u1sdtool --connect - that will ask syncdaemon to connect. After that (after 5 seconds or so) could you please re-paste the output of u1sdtool -s
[18:23] <rye> Chipaca, hi, may I ask you about bug #618225 ? I feel that if the users will end up starting the machines, waiting for local rescan and then failing to get syncdaemon connected - that would not be really great
[18:23] <ubot4> Launchpad bug 618225 in ubuntuone-client "ubuntuone-launch is useless, nearly exclusive HDD usage on login during metadata loading (affects: 1) (heat: 37)" [High,New] https://launchpad.net/bugs/618225
[18:23] <mkarnicki> rye: :D ---> u1sdtool --connect; sleep 5; u1sdtool -s ;)
[18:24] <rye> mkarnicki, should be "... ; u1sdtool -s ---> ;)"
[18:25] <rye> mkarnicki, but yes, that could be done easier, thanks... I really need to come up with a diagnostic gui for all that
[18:28] <md-llyr> rye: Got a bit confused with the other input ;-) u1sdtool --connect just gave me a prompt in response. u1sdtool --connect; sleep 5; u1sdtool -s gave me the output at http://paste.ubuntu.com/489352/ Same as before
[18:33] <rye> md-llyr, meh, not really good though it looks like... hm, could you please copy the contents of ~/.cache/ubuntuone/log/syncdaemon.log  - btw, to efficiently copy&paste to paste.ubuntu.com - install pastebinit package and put contents of http://paste.ubuntu.com/489360/ to ~/.pastebinit.xml replacing the values with your own
[18:35] <md-llyr> rye: installing pastebinit pkg now
[18:36] <rye> md-llyr, that would be handy in the future, manual pasting of the files is boooring
[18:36] <mkarnicki> rye: np, just throwing my 2 cents in :)
[18:37] <mkarnicki> rye: a diagnostic gui is a cool idea. it could even paste selected files into paste.ubuntu.com (with user permission) and copy the link into the clipboard :)
[18:37] <rye> mkarnicki, exactly, and with one-file download it will really make everything simpler
[18:42] <md-llyr> rye: Sorry for the delay. That log is at http://paste.ubuntu.com/489364/
[18:43] <rye> md-llyr, so, when you run u1sdtool --connect it simply returns the prompt w/o printing anything, right?
[18:43] <md-llyr> rye: Yes, correct
[18:45] <rye> md-llyr, could you please run the following - " ps auxww | grep ubuntuone | pastebin "  - that would automatically pastebin processes info that have ubuntuone in their names, looks like u1sdtool cannot control ubuntuone syncdaemon process so we will try killing all ubuntuone processes to get a fresh start
[18:45] <Chipaca> rye: what was that re #618225 ?
[18:50] <md-llyr> rye: I ran that and got "pastebin: command not found"
[18:51] <rye> md-llyr, ah, sorry -  ps auxww | grep ubuntuone | pastebinit
[18:52] <rye> Chipaca, i start my machine, i log into gdm, nautilus starts, nautilus plugin starts, contacts syncdaemon, syncdaemon eats my hdd for a minute or so and in the end it is not connected. How would the user feel?
[18:52] <Chipaca> rye: how many files do  you have?
[18:53] <rye> Chipaca, 376 nodes (files + folders)
[18:54] <md-llyr> rye: output is at http://pastebin.com/5gyEmT4C
[18:54] <Chipaca> rye: across all udfs?
[18:55] <rye> Chipaca, yup
[18:55] <rye> Chipaca, heh, and rtg@buzz:~/.local/share/ubuntuone/syncdaemon/fsm$ find  | wc -l -> 4974 O_O
[18:56] <Chipaca> rye: you're telling me you have only 100 nodes outside of the "from system user" ones?
[18:56] <rye> Chipaca, i don't have "from system user" shares :(
[18:57] <Chipaca> rye: can you run the metadata dumper tool and pastebin its output?
[18:57] <rye> Chipaca, metadata dumper - ?
[18:57] <Chipaca> rye: yes, the metadata dumper
[18:58] <Chipaca> rye: dump_metadata.py
[18:58] <Chipaca> facundobatista: any reason dump_metadata.py isn't in lp:ubuntuone-client ?
[18:59] <facundobatista> Chipaca, it is!
[18:59] <rye> Chipaca, https://pastebin.canonical.com/36790/
[18:59] <rye> facundobatista, Chipaca - it is, in contrib/ folder, in the sources only
[18:59] <Chipaca> facundobatista: yay :)
[19:01] <rye> Chipaca, well, i believe i have too much metadata but that does not change the issue - in case there are a lot of files syncdaemon will start disconnected
[19:01] <Chipaca> rye: yes, that is true
[19:02] <rye> Chipaca, btw, is there any cleanup script to get rid of old and bad metadata files?
[19:03] <Chipaca> rye: not afaik. facundobatista, verterok? ^
[19:03] <verterok> rye: what is an old and bad metadata file? :)
[19:03] <rye> verterok, i have 4K metadata files and only ~600 files - is that ok?
[19:04] <Chipaca> I'd call that 5k, not 4k :)
[19:04] <verterok> rye: how do you count the md files?
[19:05] <facundobatista> rye, that 'find' also counts directories
[19:05] <rye> verterok, Chipaca wait, i have ... 4974 files & folders in the metadata, ... syncdaemon/fsm$ find -type d | wc -l -> 4369
[19:05] <Chipaca> heh
[19:05] <rye> 4369 firectories.. omg
[19:05] <Chipaca> empty subdirs
[19:05] <rye> directories
[19:05] <rye> and all they are scanned
[19:05] <rye> booo
[19:05] <Chipaca> so, easy to fix
[19:05] <rye> Chipaca, yup
[19:06] <Chipaca> find -type d -depth -print0 | xargs -0 rmdir -p
[19:06] <facundobatista> rye, 'scanned'?
[19:06] <verterok> scanned ?
[19:06] <Chipaca> or something like that
[19:06] <Chipaca> rye: can you try, and let me know if that speeds things up? it might :)
[19:06] <facundobatista> Chipaca, it doesn't, we already measured that up
[19:06] <Chipaca> darn
[19:06] <rye> Chipaca, verterok, facundobatista, let me restart to have the latest readings first and to get the proper logs
[19:07] <facundobatista> verterok, do you remember that? I was in Perú at that time
[19:07] <Chipaca> facundobatista: rye has very few files, but syncdaemon still takes a minute to scan the metadata, which makes it lose the request to connect
[19:07] <facundobatista> rye, restart syncdaemon or your computer?
[19:07] <facundobatista> Chipaca, the metatada is not scanned, it is loaded
[19:07] <Chipaca> I think next cycle we need to revisit the file shelf performance, with an eye on startup this time as well
[19:08] <rye> md-llyr, could you please run the following -  kill 15642  23418 23656 and then run preferences window again - i will return in 5 minutes to continue
[19:08] <verterok> Chipaca: file shelf isn't the problem, building the fsm index is the startup bottleneck
[19:08] <Chipaca> damn again. I knew that.
[19:08] <verterok> :)
[19:19] <rye> facundobatista, verterok, Chipaca - you will be surprised: http://paste.ubuntu.com/489383/
[19:21] <verterok> rye: I'm quite sure isn't related to the deleted directories
[19:22]  * verterok need to get some food, bbiab
[19:22] <rye> verterok, 1. rebooted cleanly - ubuntuone-client started, finished into READY state. copied the lines to the text file. Removed directories; repeated - no other changes
[19:23] <rye> verterok, and i don't really see how 4K folders being scanned during metadata loading would not make a difference...
[19:23] <verterok> rye: syncdaemon doesn't scan the metadata
[19:23] <rye> verterok, how is the loading performed?
[19:24] <Chipaca> in the limit, just traversing a directory with too many nodes is slower
[19:24] <verterok> rye: really need to get some food, but basically it's doing a os.walk
[19:24] <rye> Chipaca, btw, ubuntuone-launch will fail with sso tokens - no ubuntuone-realm https://ubuntuone.com there
[19:25] <rye> verterok, exactly, os.walk() walks the directories and finds the folders, files, if I walk /tmp it would be much faster than /, right?
[19:25] <verterok> rye: you can easily write a script to do os.walk over a lot of dirs vs just a few, and you'll not see a significant difference
[19:26] <rye> verterok, ok, waiting for you after you return back and we will continue fighting :)
[19:26] <verterok> :)
[19:27] <rye> Chipaca, aargh i konw why syncdaemon is not started
[19:27] <Chipaca> rye: see? I told you it was your fault all along
[19:27] <rye> Chipaca, no
[19:27] <Chipaca> i didn't?
[19:28]  * Chipaca tries to rewrite history, and fails
[19:28] <rye> Chipaca, it won't work - syncdaemon is started by nautilus, and the check is_running() succeeds leaving syncdaemon in READY state 4EVER!
[19:28] <Chipaca> rye: el qué?
[19:30]  * rye is testing testing testing
[19:30] <Chipaca> rye: you rock. But i didn't quite get it.
[19:31] <rye> Chipaca, don't worry, i don't get it too but I am testing to find out what's wrong
[19:32] <Chipaca> rye: yes, nautilus is starting syncdaemon early. Three or four minutes later u1sdtool -c should be run by the ubuntu one startup thing. That's a bug, btw, but one that should make your issue better, not worse.
[19:32] <rye> Chipaca, erm, it is not u1sdtool -c, it is /usr/bin/ubuntuone-launch which is running
[19:32] <md-llyr> rye: I did as you suggested and killed those processes then reinvoked the Preferences module. This time I got a web screen asking me to verify the machine. However, I think NoScript was blocking something there, so I disabled NoScript, but am now back to the same place I was at the beginning. WOuld it be worthwhile to simply remove then reinstall Ubuntu One?
[19:33] <Chipaca> rye: same difference :)
[19:33] <rye> md-llyr, if noscript prevented you from connecting then try killall ubuntuone-login, close and reopen ubuntuone-preferences
[19:34] <rye> Chipaca, nope, first of all ubuntuone-launch will not work with SSO tokens since they do not have ubuntuone-realm: https://ubuntuone.com in them and that is a separate bug i am going to file
[19:34] <Chipaca> rye: correct. Yes, thank you.
[19:37] <Chipaca> rye: do you have handy the url to apachelogger's post about u1 and kde?
[19:38] <rye_> wifi died
[19:39] <rye_> bug #631822
[19:39] <ubot4> Launchpad bug 631822 in ubuntuone-client "[maverick] ubuntuone-launch is not compatible with SSO tokens (affects: 1) (heat: 6)" [High,Confirmed] https://launchpad.net/bugs/631822
[19:41] <rye_> stage_two fails......
[19:41] <rye_> no it does not...
[19:42]  * rye_ needs to make up his mind
[19:44] <duanedesign> afternooon rye_
[19:44]  * duanedesign is getting a late start on the day :)
[19:44] <rye_> duanedesign, hi, is u1sdtool starts up in READY or connected state for you?
[19:45]  * rye_ remembers asking that but now it is a subject of the channel...
[19:45] <Chipaca> duanedesign: the answer to that is "it depends", because there is a timing issue
[19:45] <rye_> Chipaca, syncdaemon does not connect now even though i run /usr/bin/ubuntuone-launch manually
[19:45] <rye_> Chipaca, by now i mean now as we speek
[19:46] <Chipaca> rye_: and you have an old-style token in your keyring?
[19:46] <rye_> Chipaca, i do, i have a time machine for this purpose
[19:46] <md-llyr> rye_: Same problem until I disabled the ABE (Application Boundaries Enforcer) module in NoScript. It seemed to be related to "Prevent Internet sites from requesting LAN resources." Even though I've reenabled that module all appears fine for now
[19:46]  * Chipaca 's sarcasm-o-meter blows up in a haze of sparks
[19:46] <rye_> to those who wonder about what I've just said - http://people.canonical.com/~roman.yepishev/ubuntuone-scripts/ubuntuone-old-auth.py
[19:47] <Chipaca> ah, cool
[19:47] <Chipaca> rye_: how does it fail?
[19:47] <duanedesign> State: QUEUE_MANAGER
[19:47] <duanedesign> queues:IDLE
[19:47] <md-llyr> rye_: Thank you VERY much for your help. Much appreciated. Discovering this may actually help me solve another problem I'm having.
[19:48] <rye_> md-llyr, another problem - what is that?
[19:48] <rye_> duanedesign, so your one got connected
[19:48] <duanedesign> rye_: yep
[19:48] <rye_> and mine is stuck in READY the same way md-llyr's one got... hmmm
[19:48] <duanedesign> rye_: my computer has been non for several days. I shall disconnect and reconnect
[19:49] <duanedesign> s/nen/on
[19:49] <duanedesign> ugh, s/non/on  :)
[19:50] <rye_> Chipaca, ok, not a clean-room experiment. Rebooting to find out more
[19:50]  * Chipaca buys rye_ a box of tissues
[19:51]  * rye_ feels that this is precisely what he had been doing last year to debug appplet dbus issue...
[19:51]  * rye_ rebooooot!
[20:00] <rye> Chipaca, duanedesign - yes, timing issue; after I got rid of those 4K old folders my syncdaemon connected on start; I believe this calls for a bug report and a cleanup script
[20:02] <Chipaca> rye: yes
[20:03] <Chipaca> rye: however, note the performance of the indexes at start *is* known to be a issue for people with many files. Such as me.
[20:04] <rye> Chipaca, yes, i have filed at least 2 bugs about that, participated in the discussion and cried about that all over #ubuntuone channel
[20:04] <Chipaca> rye: ok, good.
[20:05] <Chipaca> rye: now that that time is actually longer than server rescan, we can probably get it fixed :)
[20:05]  * rye needs to come up with a clever cleanup script...
[20:06] <rye> Chipaca, files are now uploaded faster than metadata loading times! Server rescan is nearly unnoticeable...
[20:06] <Chipaca> rye: :) yes
[20:06]  * Chipaca distracts rye away from the pain for a few seconds
[20:09]  * rye is easily distracted now
[20:38] <rye> duanedesign, could you please run find ~/.local/share/ubuntuone/syncdaemon/fsm/ -type d | wc -l ?
[21:28] <duanedesign> rye: ping
[21:29] <duanedesign> rye:  4368
[23:19] <seif> hey guys
[23:19] <seif> i keep getting
[23:19] <seif> Your Ubuntu One storage is full. Follow the link below to upgrade your subscription
[23:19] <seif> every time i log into ubuntu on mavrick
[23:19] <seif> but i only have 200 kb on ubutnu one
[23:26] <duanedesign> hello seif
[23:27] <seif> hey guys
[23:27] <seif> :)
[23:27] <seif> hi duanedesign
[23:47] <duanedesign> seif: are all your files in your Ubuntu One folder? Do you have any other folders in your Home Directory set to sync?
[23:58] <seif> duanedesign, i have nothign set ot sync