[07:42] <didrocks> good morning
[08:28] <MacSlow> good morning everybody
[08:34] <htorque_> good morning, everyone! nux won't build due to a "present but cannot be compiled" warning (http://paste.ubuntu.com/631142/). is this known or should i report a bug?
[09:03] <bschaefer> htorque: did you try installing the missing header its telling you at the bottom?
[09:04] <bschaefer> htorque: sudo apt-get install google-mock
[09:07] <htorque> bschaefer: yes (see line 2). the warning refers to http://www.gnu.org/s/hello/manual/autoconf/Present-But-Cannot-Be-Compiled.html but that doesn't help me a lot
[11:33] <andyrock> DBO, around?
[11:34] <andyrock> or didrocks...
[11:34] <didrocks> sure
[11:35] <andyrock> about this bug (https://bugs.launchpad.net/unity/+bug/727903)
[11:36] <andyrock> is a compiz bug...
[11:36] <andyrock> didrocks, so it is writtern in the proposed solution...
[11:39] <andyrock> didrocks, so we should add the project compiz, right?
[11:40] <didrocks> andyrock: right, should be fixed in compiz itself. If you can do it, it's awesome this one ruined some presentation I was giving on unity! :)
[11:41] <didrocks> andyrock: shold detect a drag in progress + hover to collapse
[11:41] <didrocks> to avoid people dropping the file to fast
[11:41] <didrocks> and then blocking compiz
[11:41] <didrocks> andyrock: I think that smspillaz will be able to give you some guidance
[11:43] <andyrock> didrocks, i can just build scale plugin right?
[11:43] <andyrock> without building all compiz...
[11:44] <didrocks> andyrock: right, if you cmake .. and go the right directory and just call "make"
[11:44] <didrocks> that should work :)
[11:44] <andyrock> didrocks, compiz coding style is weird btw :)
[11:45] <didrocks> +1
[11:45] <didrocks> it's the Xorg one :)
[11:45] <didrocks> I'm really not used to it
[11:53] <andyrock> didrocks, weird is an "euphemism" :)
[11:55] <jono> lol
[11:59] <didrocks> heh
[11:59] <didrocks> it's like "I like mixing tabs and spaces AND I KNOW IT!" :)
[12:01] <andyrock> muahahahahah
[12:04] <jcastro> heya gord
[12:04] <jcastro> https://bugs.launchpad.net/ubuntu/+source/unity/+bug/741603
[12:04] <jcastro> what information do you need here for this not to be incomplete??
[12:21] <gord> jcastro, it was a long time ago (in ubuntu years) so bare with me :)
[12:21] <gord> jcastro, so i assigned ayatana-design which means that i'm waiting for input on them on what the correct solution there is
[12:35] <njpatel> jcastro, why you up so early?
[12:43] <gord> sleep-ayatana-ing, it happens ;)
[13:57] <andyrock> JohnLea, ping
[14:02] <czajkowski> anyone in Millbank today ?
[14:16] <andyrock> in bug https://bugs.launchpad.net/unity/+bug/727903
[14:16] <andyrock> i read:
[14:16] <andyrock> Open two firefox windows. Drag a .html or .png file to the Firefox icon in the Launcher. As soon as the spread appears quickly drop the file to one of the firefox windows. The file should open into the window it was just dropped on but nothing happens.
[14:17] <andyrock> i think that the dnd in scale plugin work in this way:
[14:20] <andyrock> if we doesn't move the mouse a timeout starts otherwise it stops....
[14:20] <andyrock> when the timeout finish the spread disappers and we can drop on hovered/selected window
[14:21] <andyrock> someone could give me design feedback?
[16:02] <JohnLea> andyrock, pong
[16:48] <andyrock> JohnLea, are you still around?
[16:48] <JohnLea> andyrock; I'm still here ;-)
[16:48] <andyrock> JohnLea, about this bug https://bugs.launchpad.net/ubuntu/+source/unity/+bug/741603
[16:49] <andyrock> i read:
[16:49] <andyrock> Open two firefox windows. Drag a .html or .png file to the Firefox icon in the Launcher. As soon as the spread appears quickly drop the file to one of the firefox windows. The file should open into the window it was just dropped on but nothing happens.
[16:50] <andyrock> are you sure that the intended behavior is not this?
[16:50] <andyrock> Open two firefox windows. [...] As soon as the spread appears
[16:51] <andyrock> move the mouse on the wanted window a wait for a shor timeout
[16:51] <JohnLea> that's a bug
[16:51] <andyrock> as soon as the wanted window is show drop the file in the window...
[16:51] <JohnLea> the whole 'drag and drop via the Launcher' functionality is less than 20% built and very buggy
[16:52] <JohnLea> you should be able to drop a file on to either the launcher icon or a window in the spread and it get's loaded into the selected application/window
[16:53] <JohnLea> but very incomplete atm
[16:54] <andyrock> i read the scale pluing code... this is not the wanted behavior (at least i'm pretty sure about it)
[16:54] <JohnLea> andyrock; but what does this have to do with bug 741603 ?
[16:54] <andyrock> wrong bug
[16:54] <andyrock> in bug https://bugs.launchpad.net/unity/+bug/727903
[16:55] <JohnLea> andyrock; ahh yes, that bug
[16:55] <andyrock> i don't understand if the bug is:
[16:56] <andyrock> a) you could not drop a file in a window in the spread
[16:57] <andyrock> b) the timeout never finish since if you have not a touchpad just 1 px mouse potion restart the timeout
[16:57] <andyrock> c) both
[16:58] <JohnLea> andyrock; yes, you should be able to drop a file on to any window in a spread and it performs exactly the same action as dropping the file in the same location in the window when it is not spread
[16:59] <JohnLea> andyrock; but at the moment you cannot drop a file on to a exact location in a window while the window is displayed as part of a spread
[16:59] <JohnLea> andyrock; does that answer your question?
[17:03] <andyrock> JohnLea, and what happens to the timeout?
[17:03] <andyrock> JohnLea, we don't need a timeout if we want just drop...
[17:05] <JohnLea> andyrock; we still want the timeout, so that if you continue to hover over one of the windows you exit the spread and that window jumps to the front of the z-order stack
[17:06] <andyrock> JohnLea, so the answer to my question is c)... ;)
[17:07] <JohnLea> andyrock; yes ;-)
[17:07] <JohnLea> andyrock; that would have been a easier answer ;-)
[17:08] <andyrock> JohnLea, thx... i think that DBO when wrote the "proposed solution" considered just the timeout bug ;)
[17:09] <DBO> I do
[17:09] <DBO> its a timeout issue
[17:09] <DBO> if you want to fix the issue of being able to drop without ending the spread...
[17:09] <DBO> well... good luck!
[17:09] <DBO> I have a theory on how that might be possible
[17:09] <DBO> but I fear it wont ever really be stable
[17:10] <JohnLea> andyrock; cool, thx!
[17:12] <andyrock> DBO, i know... I was going to ask smspillaz, but I had already figured out the answer «forget it»! :)
[17:12] <DBO> andyrock, excellent :)
[17:13] <andyrock> DBO, i am just asking «Why DBO and Ubuntu guys tagged it as bitesize one??? O.o»
[17:14] <DBO> it requires no real architecturing of code
[17:14] <DBO> you just need to find the relevant code
[17:14] <DBO> and tweak it
[17:14] <DBO> (hopefully)
[17:15] <DBO> basically bitesize to me has always been "no significant new code" or at least "no new code structures"
[17:15] <andyrock> DBO, another bug with the spread-dnd thing is that when the timeout finish and wanted window appears, you need an extra motion otherwise it doen't code
[17:15] <andyrock> *code->work
[17:15] <DBO> yep
[17:15] <DBO> not much you can do about that either
[17:15] <DBO> welcome to the XDND protocol baby
[17:15] <DBO> if you dont move your mouse, it assumes nothing has changed
[17:16] <jcastro> which bug are you guys talking about?
[17:16] <andyrock> for timeout thing we can use "rectangle thing" after the target receive xdndposition and send a xdndstatus?
[17:17] <DBO> andyrock, can you restate that?
[17:17] <andyrock> jcastro, https://bugs.launchpad.net/unity/+bug/727903
[17:17] <andyrock> DBO, sure...
[17:18] <andyrock> DBO, from xdnd procol
[17:18] <andyrock> Step 5:
[17:18] <andyrock> The target sends a ClientMessage of type XdndStatus. (2) This tells the source whether or not it will accept the drop, and, if so, what action will be taken. It also includes a rectangle that means "don't send another XdndPosition message until the mouse moves out of here".
[17:19] <DBO> ah so here is the tricky part
[17:19] <DBO> so you have to think about the sender implementation
[17:19] <DBO> the way they implement basically goes like this:
[17:19] <DBO> 1) get mouse motion event
[17:19] <DBO> 2) figure out window under mouses new location
[17:20] <DBO> 3a) if window is the same as previous, check bounding area and maybe send new XdndPosition signal
[17:20] <DBO> 3b) if window is new window from previous, send XdndLeave to previous window, send XDndEnter to new window, send XdndPosition to new window
[17:21] <DBO> 4) awaite XdndStatus before sending next XdndPosition
[17:21] <DBO> 5) return to 1
[17:21] <DBO> now in our case what we do is move the window out from under Xdnd
[17:21] <DBO> however the implementor wont see this until the next motion event
[17:21] <DBO> (where it will then take path 3b
[17:22] <DBO> andyrock, ^
[17:23] <andyrock> DBO, i'm translating! :)
[17:24] <andyrock> DBO, i got it
[17:24] <DBO> andyrock, you see the inherent problem then?
[17:24] <DBO> we could solve the issue by changing how these implementations work a little
[17:24] <DBO> but that would require patching at least GTK and Qt
[17:25] <andyrock> DBO, what do you mean for `now in our case what we do is move the window out from under Xdnd`
[17:25] <andyrock> ?
[17:27] <DBO> so when compiz opens up the scale switcher thing
[17:27] <DBO> it pops up a full screen input window
[17:27] <DBO> which traps all Xdnd events
[17:27] <DBO> (which is how it figures out which window you hovered)
[17:29] <andyrock> DBO, i got it... but when i was talking about the `rectangle thing` i wanted just solve the timeout problem
[17:30] <andyrock> DBO, that is restarted also for 1px motion.... :)
[17:31] <DBO> andyrock, I am not sure I understand what you mean
[17:32] <andyrock> DBO, let me explain.... as xdnd specifications say: «It also includes a rectangle that means "don't send another XdndPosition message until the mouse moves out of here".»
[17:32] <andyrock> it also -> xdndstatus
[17:32] <DBO> right
[17:32] <andyrock> so if we include a "small" rectangle with the center in the current mouse position
[17:32] <DBO> the smallest rectangle you can send is 1x1
[17:33] <DBO> which most things do send
[17:34] <andyrock> we won't receive other xdndposition, unless the mouse movement is worthy of note
[17:34] <DBO> right
[17:35] <andyrock> if we don't receive xdndposition, the timer will not reset everytime...
[17:35] <DBO> yes but you also risk not detecting if the mouse is over the right window
[17:36] <DBO> I see what you are saying however
[17:36] <DBO> and if you are careful about your rectangles, you may have a very viable solution
[17:36] <DBO> in fact I rather like it the more I think of it
[17:36] <DBO> +1
[17:36] <DBO> actually I just thought of a problem
[17:37] <DBO> XDND says that you dont *have* to respect that rectangle
[17:37] <DBO> if an implementation chose
[17:37] <DBO> it could just send updates all the time, no matter what
[17:37] <andyrock> let me say: xdnd sucks :)
[17:38] <DBO> yes
[17:38] <DBO> yes it does
[18:19] <jaberwokey> Hi everyone!  I'm got a simple question about the C api for app indicators.  Given the following code, how would I set up a callback for when an item is clicked?  https://gist.github.com/1045233  Total GTK noob, sorry. :P
[19:09] <Daekdroom> Why every time I click and drag an item from a list within a qt app, the launcher won't hide anymore?
[19:54] <davidboy> After some poking around in DevHelp, I come up with the following: http://pastie.org/2117383
[19:54] <davidboy> Only one problem: It doesn't work. :D
[19:54] <davidboy> Anyone can tell me what I'm doing wrong?
[22:04] <coz_> hey all