[05:40] <mmcc> mmcc: just solved that Qt image plugin packaging issue. We might be out of the woods...
[08:29] <JamesTait> Morning all! :)
[08:46] <mandel> morning!
[09:33] <rye> anybody here uses U1 with proxy?
[09:59] <rye> mandel: I have several users who repetedly enabled and disabled file sync status for a folder from the control panel because the latter was seemingly not reacting to the click (and the user has just confirmed that the flag appeared after 5 seconds). Is it known?
[10:04] <mandel> rye, I can test thing if you need me regarding proxy, and the second question..
[10:05] <mandel> rye, what is the later? what OS?
[10:05] <mandel> rye, is that ubuntu? is the falg the one from nautilus?
[10:10] <rye> mandel: no, the control panel, on Ubuntu
[10:10] <rye> mandel: i got this once but i was not able to reproduce. Apparently, some succeed at this
[10:12] <mandel> rye, and when you say the later what does it mean 'I have several users who repetedly enabled and disabled file sync status for a folder from the control panel because the latter was seemingly not reacting to the click'
[10:12] <mandel> rye, sorry I'm a little lost in translation :)
[10:12] <mandel> rye, and my context switch from mac to this is not that good hehe
[10:13] <rye> mandel: i am context-switching every 5 minutes and already have problems recalling things, i understand. So, you click a folder checkbox in control panel. Nothing appears to happen. You click again, nothing (during this time SD has enabled and then disabled folder syncing) but CP does not show anything.
[10:14] <rye> karni: hi, is there a reason why we prohibit moving the application to SD card for those who have less than 0 free bytes on the internal flash :) ?
[10:15] <mandel> rye, ah, ok, and the checkbox does not show the 'check' correctly ?
[10:17] <rye> mandel: it shows the check but after a looooooong delay
[10:18] <mandel> rye, ok, so that operation is async, that is, the message gets sent to sd when sd process the action it then sends a signal over dbus to control panel
[10:19] <mandel> rye, so, we can have the delay because a: dbus message from cp to sd is slow, b: the action is push to the reactor thread and there stays a little 'til processed, c the signal is slow
[10:19] <mandel> rye, and I have had no info about this happening before, is a first to me
[10:22] <rye> mandel: e.g. when it is syncing something... Grrr. I need to have my account fixed
[10:23] <mandel> rye, well uploading/downloading/hashing are done in diff threads but maybe there is something blocking there
[10:23] <mandel> rye, or is a ui problem.. it can be that the main loop qt does not get the signal on time from the deferred and there is an issue there
[10:52] <rye> mandel: i thought that meta queue items are still executed consequently, so e.g. if there are lots of MakeFile/MakeDirs, adding a new request to get volume list (or whatever enabling share is using) will be added to the end of the queue...
[10:53] <mandel> rye, yes, I think that meta queue is in the reactor main thread
[10:53] <mandel> rye, all event processing is there too
[10:53] <karni> rye: Yes, multiple support requests of the need to log in whenever they mount USB storage/restart their phone, or pictures not syncing. People just don't get the limitations of apps running on SD card.
[10:54] <karni> rye: It wasn't an easy decision and it took me time to decide to revert to internal storage only, but it was thought out and intentional. Sorry if it upset a few folks. Certainly less than the trouble associated with having app2sd enabled.
[10:57] <rye> karni: and with MTP-only access on the newer phones. I understood that :)
[10:58] <karni> rye: right, newer phones have large internal storage only, so that's less of an issue.
[11:10] <gatox> good morning!
[11:22] <rye> karni: another quesiton - are the changes to make the files already downloaded and changed on the phone to be uploaded automatically planned for u1f?
[11:24] <karni> rye: yes, sync is indeed on the roadmap. (I'll finally stop complaining, how people expect sync, while we never said the Android app had it, anywhere ;))
[11:25] <karni> rye: FWIW, keeping files from U1 with those on the phone (re-downloading) is pretty darn easy to implement. I didn't add it this week only because I knew if we have sync, it has to be _sync_ and not a partial solution.
[11:28] <rye> karni: the user actually edits the files on the phone
[11:29] <rye> karni: oh, while I got your attention - any reason why our music app does not stream flac files?
[11:29] <mandel> gatox, morning!
[11:30] <mandel> gatox, can you give me a hand with something?
[11:30] <gatox> mandel, of course.... shoot
[11:31] <karni> rye: Right, noticing local changes will either require rescan or FileObservers on each of the files "of interest"
[11:31] <mandel> gatox, if you have the following (char *)&(kevent_arg->data.str) where data is a union, what do you think str should be, and the answer is not a char*
[11:31] <mandel> gatox, I'm getting a little confuse with all this address/pointer operations :(
[11:31] <karni> rye: Not that I know of. I'll sync with Chad when he's around.
[11:32] <gatox> mandel, let me see
[11:32] <mandel> kevent_arg is an array of a struct that has two ints and the data union
[11:33] <gatox> mandel, "str" is something like __repr__
[11:33] <gatox> ?
[11:34] <mandel> gatox, let me paste the struct definition, one sec
[11:35] <gatox> mandel, maybe i'm wrong, but i assume what you are getting a reference to some structure... and then obtaining a pointer to the first byte of that structure
[11:35] <mandel> gatox, so, if I wanted to recreate the structure (bloody tests!!!) how would it be done?
[11:36] <mandel> gatox, going data.str = "blah" is not correct
[11:37] <gatox> mandel, asking the sizeof of the union and creating a char* array as big as that? starting from the first byte as you have now?
[11:38] <mandel> gatox, http://paste.ubuntu.com/1016244/
[11:38] <gatox> mandel, if i'm talking nonsense, show me some code
[11:38] <gatox> that :P
[11:38] <mandel> gatox, that is the struct def
[11:39] <gatox> ahhhh i understand better now what you mean
[11:40] <gatox> mandel, and *str is not ending in \0?
[11:41] <mandel> gatox, is ac string, so it should
[11:41] <gatox> mandel, that would be a way to obtain the complete string, doesn't it?
[11:42] <mandel> gatox, the problem I have is that I want to create such a struct to be used in the tests yest when I do data.str = 'blah' is not correct and get some random mem info
[11:43] <mandel> gatox, and the kernel sends the string that way  (char *)&(kevent_arg->data.str) so I cannot remove it because that makes the daemon to segfault
[11:51] <gatox> mandel, i'm trying to think some options..... but..... i'm kind of rusty :S
[11:52] <mandel> gatox, same here!
[11:52] <mandel> gatox, I might as well ask dobey, I'm sure he is better than us at this
[11:52] <gatox> mandel, yep..... he is the guy
[12:17] <alecu> rye, mandel: regarding the delay when marking folders to be synced: it sounds like the user is transferring a lot of files (more than 200?) so SD buffers are full and the "suscribe" operations need some download to finish before they are executed on the server.
[12:18] <rye> alecu: oh, that's fun
[12:18] <mandel> alecu, nice!
[12:18] <alecu> rye, mandel: they also could happen when more than the limit of concurrent transfers (10 is the default), but in that case the wait would be much smaller.
[12:18] <rye> alecu: i guess we need to somehow show the operation is in progress. There were 4 users today who spent a several days enabling/disabling the folders
[12:18] <alecu> rye, in any case we *should* fake that a little bit better on the UI side.
[12:19] <alecu> I think that we should be ticking the checkbox as soon as the user clicks on it... or mark it as grayed, or something.
[12:20] <alecu> rye, did you got any more info on the "chrome download gone" that was in fact a "mind map file gone"?
[12:21] <rye> alecu: ticking immediately looks good unless restarting the control panel shows it is not flagged. In any case we should not queue the request for toggling and do only action that is intended
[12:21] <rye> alecu: I am afraid i am swampted with other requests and I am physically not able to follow up on some of these tickets
[12:21] <alecu> rye, I understand :-(
[12:22] <mandel> alecu, question, do you know what is str mean to be here: http://paste.ubuntu.com/1016244/ in the daemon code we do  (char *)&(kevent_arg->data.str) but in the tests when a do data.str = "char" I'm not getting the correct string out, any idea?
[12:22] <mandel> alecu, my pointers knowledge is rusty (at least at this level)
[12:23] <alecu> mandel, never venture into C land with a rusty knowledge o' pointers!
[12:24] <gatox> alecu, i'm back on fsevents (finish with the u1-cp and u1-client high issues), can you remind me where is (or just the name) of the function in linux where we wait when a new folder is added before start listening to fsevents?
[12:24] <mandel> alecu, yes, I know.. my fault, I understand what is doing but I'm puzzled on how to recreate the strcture for the tests
[12:25] <alecu> mandel, please show me where you do "(char *)&(kevent_arg->data.str)" in our code.
[12:26] <mandel> alecu, sure, give me a sec
[12:28] <mandel> alecu, https://bazaar.launchpad.net/~mandel/ubuntuone-fsevents-daemon/add-run-tests/view/head:/objc/FsEvents/FsEvents/FSEvent.m#L108
[12:28] <mandel> alecu, I want to recreate the arg structure to have tests to ensure that is correctly parsed
[12:29] <gatox> alecu, "new_local_dir"?
[12:32] <mandel> alecu, I'm off to have lunch, I'll be back asap
[12:33] <alecu> mandel, got a sec?
[12:33] <mandel> alecu, yes, sure
[12:34] <alecu> mandel, (char *)&(kevent_arg->data.str) is wrong
[12:34] <alecu> mandel, it should be: (char **)&(kevent_arg->data.str)
[12:34] <alecu> mandel, kevent_arg->data.str is the pointer to the string
[12:35] <alecu> mandel, &(kevent_arg->data.str) is a pointer to the pointer to the string
[12:35] <mandel> alecu, yes, and I have tried to simple remove (char *)& because arg->data.str is already a char* but if I do that, when I run the daemon I get a segfault accesing the wrong memory
[12:36] <alecu> mandel, so, with &(kevent_arg->data.str) you are getting a pointer that points inside your  kfs_event_arg struct (probably to the 4th byte, since the first four are the two uint_16s)
[12:36] <alecu> mandel, well, I need to "C" that "see" code.
[12:37] <mandel> alecu, puf.. terrible joke hehehe
[12:37] <alecu> mandel, yes, probably my worst this week.
[12:37] <alecu> mandel, but it surely beats your best this month :-)
[12:38] <mandel> alecu, ha, I've got great jokes! hehe
[12:38] <mandel> alecu, let me have lunch and I'll take a look at the original c code and find out what that is really doing...
[12:39] <mandel> alecu, maybe my struct in the test is wrong and breaking everything
[12:41] <alecu> mandel, go ahead, have lunch.
[12:41] <mandel> alecu, ok, I'll be back asap
[12:44] <ralsina> good morning!
[12:44] <mandel> ok, really going now.. will be back in a few mins
[12:45] <ralsina> alecu: I am pretty sure of it, but bug #732255 is a non-issue now, right?
[12:45] <ralsina> alecu: Also, why is the UBuntu Foundation bug's bt changing the priority of our old bugs?
[12:46] <ralsina> that was not alecu-specific :-)
[12:48] <alecu> ralsina, it looks very old, and looks like it should be fixed already.
[12:49] <ralsina> alecu: what I thought
[12:49] <ralsina> if it still happened, I would have seen it in the last year
[12:49] <alecu> ralsina, I didn't know there was a "ubuntu foundation" tough :P
[12:49] <ralsina> alecu: and it has robots!
[12:49] <dobey> ralsina: because the priority was probably unset, and something got fixed in the bot, so it sets crashers to medium by default like used to happen
[12:50] <ralsina> dobey: interesting
[12:50] <dobey> also, note it's only setting the severity on bugs filed against the ubuntu packages, not the upstream project
[12:50] <dobey> and yes, there's an "ubuntu foundations" team :) (though i think maybe it's called platform or somethng now)
[12:51] <rye> alecu: PROXY!
[12:51] <rye> alecu: i found that proxy does not work with syncdaemon
[12:51] <alecu> rye, tell me more
[12:52] <rye> alecu: STR: install win7, install squid on another machine. prohibit windows from accessing the internet directly, configure proxy in windows internet explorer -> control panel works, SD fails to connect
[12:53] <rye> alecu: i can give you rdp access to the instance if you want to have a look
[12:53] <ralsina> rye: that happens when you don't have DNS on the client box, yes, known bug
[12:53] <rye> ralsina: DNS works
[12:53] <rye> or
[12:53] <rye> wait
[12:54] <dobey> i guess it's a problem with the SRV magic
[12:54] <dobey> which the http stuff doesn't use
[12:54] <rye> dobey: no, SRV fallbacks to a host...
[12:55] <rye> ralsina: how was the proxy support tested then?
[12:55] <rye> ralsina: dns works
[12:56] <ralsina> rye: I tested it but I had a local resolving DNS
[12:57] <ralsina> rye: works? Ok then it's not that bug
[12:57] <alecu> rye, can you give me rdp access?
[12:57] <gatox> only 4 erros in test_filesystem_notifications in mac \o/ (happy news for me :P)
[12:57] <ralsina> gatox: yay
[12:58] <ralsina> I am tempted to do a blanket WONTFIX for desktopcouch
[12:58] <ralsina> it has 75 bugs which we won't, you know, fix.
[12:59] <gatox> rye, did you have the chance to read my comment in the Bug #1006329
[12:59] <dobey> ralsina: unfortunately, we do sort of have a contract to support it for some versions of ubuntu
[12:59] <ralsina> dobey: we could provide emotional support...
[12:59] <dobey> i don't think that's what LTS means
[13:00] <ralsina> dobey: and I am pretty sure the contract doesn't say we are going to fix any bugs
[13:00] <ralsina> ok, it has 156 bugs, launchpad was lying by omission a bit there
[13:00] <dobey> and changing them to wontfix, won't stop you getting e-mail for them
[13:01] <ralsina> the 1st page doesn't show the total bug count, says "75 bugs" instead of "75 of 156 bugs", but it works on the second. Reporting...
[13:02] <ralsina> worse, it says "75 of results"
[13:02] <dobey> eh?
[13:02] <dobey> it works fine here
[13:05] <ralsina> dobey: http://www.flickr.com/photos/ralsina/7307958708/in/photostream
[13:08] <dobey> ralsina: what browser/os?
[13:08] <ralsina> dobey: ubuntu + chromium
[13:08] <ralsina> dobey: but the number is not there in the HTML, so it's not a browser issue
[13:09] <ralsina> OTOH, it's JS so it may be
[13:11] <dobey> hrmm
[13:11] <dobey> chromium also is broken there
[13:11] <dobey> for me
[13:11] <dobey> oh weird
[13:12] <dobey> it's there, and then gets hidden, in chromium and epiphany
[13:13] <dobey> that is really weird
[13:13] <dobey> because it's not in a separate div or anything
[13:14] <dobey> not sure how this isn't a webkit bug though
[13:15] <ralsina> dobey: if any thing that looks off in a page is to be fixed upstream, web devs would have it much easier than they do ;-)
[13:23] <dobey> ralsina: it /could/ be a YUI bug perhaps, but again, looking at the source for the page, i don't really see how
[13:24] <ralsina> dobey: it's reported and now I know it's there, so I don't care anymore
[13:24] <ralsina> we have enough bugs with our names on them :-)
[13:43]  * mandel bacj
[13:43] <mandel> ups, back
[13:43] <ralsina> rye: forbug #1006899 how did you test DNS connectivity?
[13:43] <ralsina> rye: for bug #1006899 how did you test DNS connectivity?
[13:45] <rye> ralsina: nslookup google.com -> LOTS OF IPs
[13:45] <ralsina> rye: ok
[13:48] <rye> ralsina: sent you rdp details privately
[13:49] <ralsina> rye: thanks!
[14:17] <ralsina> dobey: you remember the bug # for the empty dict in dbus signature you fixed?
[14:18] <ralsina> dobey: I think it has 2 dups, 1006932 and 951406
[14:18] <dobey> the ubuntuone-login issue?
[14:18] <ralsina> dobey: I think so
[14:19] <dobey> oh, control panel
[14:20] <dobey> hmm
[14:22] <dobey> ralsina: 711162, marked both as dup
[14:23] <ralsina> thx dobey!
[14:32] <alecu> ralsina, I was able to reproduce the bug that rye reported re: proxy. I get very different behaviour on a pristine vm vs my dev env.
[14:32] <ralsina> alecu: yikes
[14:32] <alecu> probably different versions of twisted, or something like that.
[14:32] <alecu> ralsina, my dev-env connects; the pristine vm with 3.0.1 does not.
[14:32] <ralsina> alecu: well, if it's that, it's easy to fix at least
[14:33] <ralsina> alecu: you can ask briancurtin for the versions used in the package
[14:33] <briancurtin> alecu: i'll get them right now - which versions do you need?
[14:33] <rye> yay
[14:33] <rye> bugs
[14:33] <rye> fixs
[14:34] <alecu> briancurtin, I'd like to know which version of twisted is used to build the package.
[14:34] <briancurtin> alecu: twisted 11.1.0
[14:35] <alecu> Hmmm... I'm on 11.0.0. I don't think that should be the cause.
[14:55] <ralsina> alecu, dobey, briancurtin, gatox, thisfred, mmcc, mandel: remember team call in 5 minutes
[14:55] <gatox> ack
[14:55]  * alecu is already on mumble
[14:55] <mmcc> aaaaaaack
[14:56] <thisfred> ackack
[14:56] <dobey> remember remember, it's not november
[14:56] <thisfred> camembert
[14:56] <dobey> is there a duck in the room?
[14:57] <mandel> ack
[14:57] <thisfred> at least we dispelled the myth that duck's quacks don't have echoes
[15:01] <ralsina> joshuahoover: you're welcome to our call too!
[15:13] <ceed^> Hi, I've set autouploa of pictures from my phone to a folder within the Ubuntu One folder. Will that cause any problems?
[15:13] <ceed^> autoupload even
[15:33] <dobey> karni, qengho: ^ can you answer ceed^ please?
[15:34] <karni> yes, just a sec
[15:34] <ceed^> I can wait :)
[15:34] <ceed^> Kinda difficult moving it all from Dropbox and make it work right.
[15:35] <qengho> ceed^: that sounds normal to me, if I understand you.
[15:37] <ceed^> qengho: I created a folder within the Ubuntu One folder called "From Phone" and set the Android app to upload there. But since the Ubuntu One folder is auto synced I wonderd if there would be problems.
[15:40] <karni> I'm here, sorry.
[15:40]  * karni reads up
[15:41] <karni> ceed^: Of course, no problem with auto uploading to a folder within ~/Ubuntu One
[15:41] <karni> ceed^: Why don't you give the latest U1F a try :) http://t.co/gs7SVdf8 (QR: http://t.co/gs7SVdf8.qr )
[15:42] <ceed^> Thanks. I know all little details about Dropbox, but U1 is still a little akward for me
[15:42] <karni> ceed^: What do you find akward about it? :)
[15:43] <qengho> ceed^: nothing "owns" that folder, and it shouldn't pull files down to your device and then try to send it back up.  No worries.
[15:43] <ceed^> What happens if you create folders within a folde already set to autosync. Will that new folder be autosynced by default?
[15:43] <qengho> I think we may need to define "autosync" better, karni.
[15:43] <karni> ceed^: There's to sync, nor auto sync :)
[15:44] <karni> It is Photo Auto-Upload
[15:44]  * briancurtin brb, need to run to the copy shop to fax something...gas company won't let me email this form and i want hot water again
[15:44] <karni> we never ever used sync anywhere around Android, but (sadly) users still fail to understand Android app is quite limited compared to what the PC version is.
[15:44] <karni> So, ceed^, let me explain :)
[15:45] <karni> ceed^: Auto-Upload allows to automatically upload photos from (selected) folders from your phone to Ubuntu One
[15:45] <mmcc> mandel, line 108 in FSEvent.m doesn't look right, shouldn't it be string_value = kevent_argument->data.str;
[15:45] <karni> ceed^: Currently Ubuntu One Files will not automatically download nor _sync_ any files
[15:45] <ceed^> Yes, all pictures in folders on the phone are uploaded to a folder in U1. If that upload folder is within a folder set to sync will the upload folder also be synced to the computer?
[15:45] <mandel> mmcc, apparently no, if you do that you get a segfault, and that is exactly the thing I have been fighting with
[15:45] <karni> ceed^: If it notices a change in a file in the cloud (say, you made a change on your computer), it will remove the old file from your phone, if it has been downloaded, so it doesn't get used by the user.
[15:46] <karni> ceed^: Yes, that is correct :)
[15:46] <karni> ceed^: If you upload your pics to ~/Ubuntu One/My mobile pictures, everything under ~/Ubuntu One cloud folders gets synced to your comptuers
[15:46] <karni> qengho: Hi, nice to meet you :) Are you a community person, or just Canonical?
[15:46] <qengho> ...and web site.
[15:47] <karni> Well, yes. The web site represents whatever should be on the computers :)
[15:47] <ceed^> Thanks! that's what I wanted to know. Now I just have to wait to get all my 80gb pics uploaded. Gonna take days... :)
[15:47] <karni> Hahahahah
[15:47] <qengho> karni: you already know me.
[15:47] <ceed^> I also wonder why you can't share folders, only files
[15:47] <karni> ceed^: Do you have 80 GB available in U1?
[15:47] <mmcc> mandel: hrm. but data.str is a char*. data.vp is a struct vnode* vp. if you want a char* you should use .str  -- the segfault is probably something different. ergh, wish I had this building on 10.6
[15:48] <ceed^> karni: yes, purchased tonse of storage
[15:48] <ceed^> tons even
[15:48] <alecu> mandel, I agree with mmcc
[15:48] <karni> ceed^: Awesome :)!
[15:48] <ceed^> Keeping my Dropbox until I know it works :)
[15:48] <alecu> mandel, stringWithUTF8String takes a char*, so .str is the right thing to pass.
[15:48] <karni> ceed^: Are ou asking about sharing from the Android app, or within U1 in general?
[15:49] <mmcc> alecu, mandel, of course, .vp is the same address, so casting it to (char *) is basically the same thing! :) just uneccessary
[15:49] <dobey> mandel: can we get my branch unblocked please? :)
[15:49] <karni> ceed^: You can share whole folders to your friends, and (currently) only publish files from the Android client. The API the Android app uses does not expose shared folders yet.
[15:49] <qengho> ceed^: we're always eager to hear suggestions or feedback.  email or irc.
[15:49] <mandel> dobey, sure, can I have the mp link please
[15:49] <ceed^> And I miss the photo album functionality in Dropbox
[15:50] <mandel> alecu, mmcc it is a union, so the addresses are indeed the same
[15:50] <ceed^> Consider that a feature request :)
[15:50] <dobey> mandel: https://code.launchpad.net/~dobey/ubuntuone-installer/new-logo/+merge/107280
[15:50] <mmcc> mandel, is that code in FSEvents.m:108 segfaulting?
[15:50] <karni> ceed^: Considered! :)
[15:50] <qengho> ceed^: we have smart people thinking about that right now.
[15:50] <karni> ;)
[15:50] <dobey> mandel: thanks
[15:50] <ceed^> And sharing of folders although I found a python script that will do it
[15:50] <alecu> mmcc, is there a way that line 108 can segfault? I can't find it.
[15:51] <mandel> mmcc, the code as it is works, if you use the ->data.str directly it segfaults
[15:51] <mmcc> mandel, what exactly is the code that will segfault?
[15:51]  * ceed^ likes smart people to think when his data backup is on the line
[15:51] <alecu> mmcc, oh, if kevent_argument has an invalid value.
[15:51] <karni> qengho: You mind telling ceed^ how to share a folder with somebody? (I totally forgot xD)
[15:52] <mmcc> alecu, yeah but it'd segfault in the switch if kevent_argument was invalid, so i guess 108 can't
[15:52] <mmcc> mandel, I mean, can you show me the exact line that causes a segfault?
[15:52] <mandel> alecu, mmcc, if it points to a random address we go bananas, so it indeed can segfault
[15:53] <alecu> mmcc, you are right on the switch segfaulting first.
[15:53] <mandel> mmcc, I have not managed to get the daemon to run as root using lldb so I can only guess
[15:53] <ralsina> can I get a second review for https://code.launchpad.net/~ralsina/ubuntuone-control-panel/decrypt-errors-3-0
[15:54] <ceed^> I would also like a U1 icon in the notification area like Drobox :)
[15:54] <ceed^> so I can see what's going on
[15:54] <qengho> ceed^, karni, on linuxes, u1sdtool is a command-line tool for doing things like that.  Sharing something is (probably) just two steps, of looking up its ID and then sharing that ID.
[15:54] <ceed^> ah ok
[15:54] <mmcc> mandel, oh ouch. well at least there should be a core dump somewhere... let's see - look in /cores ?
[15:54] <dobey> ralsina: i need to take an extra long lunch today. meeting ran long, and i have an appointment in 1hr, so i'll be back in about 1.75-2 hrs, and work a little longer tonight
[15:54] <karni> qengho: I meant.. with a mouse ;D
[15:54] <karni> qengho: Is it under the context menu?
[15:55] <ralsina> dobey: np
[15:55] <karni> qengho: Or only of synced folders context menu?
[15:55] <dobey> bbiaw
[15:55] <mmcc> mandel, actually, if kevent_argument->type is FSE_ARG_VNODE, then kevent_argument->data won't be a string - that code shouldn't be expecting a string anywhere
[15:56] <mmcc> er, I mean kevent_argument->data.str is undefined... you shouldn't use .str if that's the type...
[15:56] <mandel> mmcc, kevent_argument->data will the the union pointing to the vnode struct
[15:56] <mmcc> mandel right - and the vnode struct isn't a char* but that code is treating it like it is
[15:57] <mandel> mmcc, yes, that is one of the errors in the code.. the question is, do we care about vnodes.. I think we don#t
[15:57] <mandel> don't
[15:58] <mmcc> hmm
[15:59] <qengho> karni: any place in your home directory, you can right-click in Nautilus and "synchronize".  Once the server knows of it that way, you can share contents.
[15:59] <karni> qengho: Thanks! ceed^ ↑
[15:59] <mandel> mmcc, I'll look into cleaning that code in the next branch that adds tests for the event arguments
[16:05]  * alecu will have some lunch
[16:06] <mmcc> mandel: ok.. I'm still curious about exactly what the code is when it doesn't work. is it  = (char *)&(kevent_argument->data.str) or just = kevent_argument->data.str ? because you shouldn't need the '&' - .str is already a char*, and with the & you're getting a char**, the address of the address of the char...
[16:07] <mmcc> ... so does line 117 work? it's doing the same thing and looks like it should be creating a string from the wrong place.
[16:09] <ceed^> the music streaming has tough competition comapred to Google Music (which also rund on Linux)
[16:10] <mandel> mmcc, does work, it shouldn't
[16:11] <mmcc> weird, I see fslogger also did that, so I guess I'm confused?
[16:11] <mandel> mmcc, you are not the only one, I'm too
[16:12] <mandel> mmcc, for me it is getting a char ** and not a char * since str is a char* and therefore applying & gets you the pointer to the pointer
[16:12] <mandel> mmcc, I'm going to fill the things with nNSLog and see wtf is going on
[16:13] <ceed^> I have one more question, is there a way to see what is going on with U1 like a log?
[16:14] <mmcc> mandel right, so when you then send that pointer to stringWithUTF8String, it should get garbage. unless we're really (un)lucky with memory layout. I suppose that if the storage for the string is malloc'd right after the storage for the struct, you might get something that almost works, depending on the utf8 value of the address in the struct :\
[16:14] <mmcc> but I could be making that up
[16:15] <mandel> mmcc, is something I grabbed form fslogger and I'm working on cleaning up + adding tests
[16:16] <mmcc> mandel, right, I am looking at the code in fslogger now. I'm just puzzled. You're on it, so I'll leave you alone :)
[16:16] <mandel> mmcc,  :)
[16:16] <mandel> mmcc, I'm going to EOD soom, so I'll ping you tom with what I have and we can take a closer look at the mem allocation
[16:17] <mmcc> mandel: one more thing: did you try just running lldb as root?
[16:17] <mandel> mmcc, yes, bad things happened
[16:17] <mmcc> mandel ok hrm, I thought that should work. oh well. gdb might also still work...
[16:18] <mandel> mmcc, yes, I have to try with gdb, specially because I know the commands :P
[16:22] <ralsina> LUNCH!
[16:32] <elopio> hey team. How can I help translating the control panel?
[16:32] <elopio> on launchpad it says that there are no translations.
[16:32] <mmcc> mandel BTW, for interpreting C variable type decl gibberish, http://cdecl.org can be useful... (cc thisfred )
[16:33] <thisfred> thx
[16:34] <thisfred> mmcc: is that the same one that is in the kernighan ritchie book? :)
[16:34] <mandel> mmcc, haha good one!
[16:35] <thisfred> I think I got it mostly, and vim usually warns me when I get it wrong
[16:35] <mmcc> thisfred: yeah, it's just a thing that's useful to know exists...
[16:36] <thisfred> sure
[16:36]  * gatox lunch
[16:36] <mmcc> it handles blocks, which is nice for macos stuff, and I've occasionally needed it when passing function pointers as void*...
[16:36] <mmcc> and then recasting them
[16:36] <mandel> mmcc, specially when you have (char *)&(char*) or at least is what I understand :P
[16:37] <mandel> ok, EOD (mainly walking the dog)
[16:37] <mandel> catch you all tom!
[16:37] <mmcc> bye, tom
[16:37] <mandel> mmcc, I hope my morning is useful for the funny mem thing :)
[16:38] <mmcc> mandel, me too. I've been running fslogger for a while now - no VNODE args ever showup.. maybe we don't care. but the author's site shows a trace with them in there, soo....
[16:38] <mmcc> and at least in freebsd source, struct vnode* does not start with a char*, so ... Amit Singh knows something I don't :)
[16:39] <mmcc> anyway, have a good walk, see you tomorrow
[16:43] <mmcc> for future reference -- lsyncd ignores any arguments that aren't a string, and has a different structure, it checks the argtype and does different things with a string argument depending on the argtype. It's GPLv2, and the code is here: http://lsyncd.googlecode.com/svn/trunk/fsevents.c
[17:23]  * mmcc lunches
[17:37] <briancurtin> i wonder if HR will ever mail me back about anything...
[17:37] <ralsina> briancurtin: aything I can help you with?
[17:38] <briancurtin> ralsina: one is just a benefits question, the other is that half-day i entered but didn't use. i just mailed for the second time about the day off
[17:38] <ralsina> ok
[17:38] <ralsina> about the half-day we can always just handle it unofficially as an exception
[17:39] <briancurtin> ralsina: yeah that was my next plan. if i dont hear back from them in the next few days i'll just schedule a half-day in the near future so i dont forget it
[18:33] <dobey> sigh. that took a little longer than expected
[19:18] <ralsina> school run!
[20:04] <gatox> ok...... eod for me people!! see you on monday!! and some of you see you on UbuConLA :D
[20:05] <mmcc> bye gatox, have a good trip!
[20:33] <ralsina> EOD for me
[20:33] <ralsina> see you all tomorrow!
[22:17] <mmcc> ok, heading out. everyone quiet down
[22:17]  * mmcc leaves