--- Log opened Sat Feb 23 00:00:26 2008
00:01 < Yoyobuae> i wonder if on some CPU in the future there will be no separate cores, just a number of thread support hardware blocks with a number of CPU resources
00:02 < Yoyobuae> the resources being shared seamlessly =)
00:02 < dooglus> the color edit specifically sets the minimum R,G,B,A values, eg:
00:02 < dooglus> R_adjustment.set_lower(-10000000);
00:03 < Yoyobuae> talk about a wide margin to avoid clipping XD
00:05 < Yoyobuae> is that the reason behing HDRI? so that algorithms don't accidentally go off the range of colors?
00:05 < Yoyobuae> oh also to allow greater variety of colors i guess
00:05 < dooglus> http://en.wikipedia.org/wiki/High_dynamic_range_rendering shows some good examples
00:05 < dooglus> to avoid clipping, basically, it seems
00:06 < dooglus> the sun is so bright it still maxes out the display even after passing through a window that cuts of 50% of the light
00:06 < Yoyobuae> pretty examples, eyecandy. but tells nothing about whats really going on =/
00:06 < dooglus> so it needs to have been >200% bright to start with to get that effect
00:06 < Yoyobuae> yeah that's what i meant
00:07 < Yoyobuae> the circle on plasma_circle.sifz gets it's color toned down by feather
00:07 < cizra> Waaah! I've lost my ducks!
00:07 < cizra> Where did they go?
00:07 < Yoyobuae> but the higher values take longer to decay
00:08 < Yoyobuae> cizra: have you tried show/hide ducks on view menu 
00:09 < cizra> Ach! How good.
00:10 < dooglus> cizra: keyboard shortcuts alt-1 through alt-6
00:10 < cizra> That might explain why they disappeared.
00:10 < dooglus> yup :)
00:12 < Yoyobuae> maybe a new tutorial is in order: Keyboard Tutorial XD
00:15 < dooglus> there's the list of shortcuts
00:15 < dooglus> it's linked from the Help menu now
00:25 < Yoyobuae> oh, right forgot about that =)
00:29 < CIA-27> synfig: dooglus * r1800 /synfig-studio/trunk/src/gtkmm/widget_coloredit.cpp: Make the spinbuttons with the R,G,B,A color values in them a bit wider. The values can get quite large. Also, display one decimal place, rather than rounding to an integer.
--- Log closed Sat Feb 23 00:38:45 2008
--- Log opened Sat Feb 23 00:38:51 2008
00:38 -!- dooglus [n=dooglus@rincevent.net] has joined #synfig
00:38 -!- Irssi: #synfig: Total of 17 nicks [1 ops, 0 halfops, 0 voices, 16 normal]
00:39 -!- Irssi: Join to #synfig was synced in 37 secs
00:49 -!- MangoFusion [n=jamesu@host81-132-205-163.range81-132.btcentralplus.com] has quit ["Not here"]
00:50 < dooglus> ?
00:52 < cizra> You're the One.
00:58 < dooglus> you're mixing me up with keanu reeves?
00:58 < cizra> Who's that?
00:58 < dooglus> oh, ok, maybe you're not then
01:02  * rubikcube was sucked again into the synfig wiki
01:02 < pixelgeek> Jet Li?
01:02 < pixelgeek> http://www.imdb.com/title/tt0311363/
01:13 < dooglus> :)
01:19 < dooglus> Yoyobuae: I'm not sure about assuming that canvases won't have spaces in their exported ids
01:19 < dooglus> I know they can't at the moment, but it would be nicer if they could
01:25 -!- pabs3 [n=pabs@d122-105-74-150.per9.wa.optusnet.com.au] has quit [Read error: 110 (Connection timed out)]
01:33 < dooglus> pixelgeek: did you try building yet?
01:44 -!- Yoyobuae_ [n=Yoyobuae@201.224.135.156] has joined #synfig
01:44 -!- Yoyobuae_ is now known as Yoyobuae
01:45 < Yoyobuae> dooglus: dunno, canvases also have a name, which can be anything
01:45 < Yoyobuae> i don't like the idea of saving the information as a file line by line either though
01:46 < Yoyobuae> it would be better to change 'recent_files' format to something like 'settings' format
01:49 < Yoyobuae> or perhaps put everthing inside 'settings', why have more than one file?
01:54  * pixelgeek going home now to try the compile
01:55 -!- pixelgeek [i=86868801@gateway/web/ajax/mibbit.com/x-bc91d9ac2017676c] has left #synfig []
02:21 < dooglus> Yoyobuae: one other thing - if I load a canvas, open and close a bunch of subcanvases, move them around and resize them them quit and restart, the positions aren't remembered
02:22 < dooglus> I need to save the unmodified canvas to keep the positions
02:23 < Yoyobuae> i kinda intended to be like that
02:23 < Yoyobuae> what happens if you close an unsaved composition?
02:24 < Yoyobuae> i guess the problem is that synfig closes without asking anything XD
02:25 < dooglus> if it's unmodified and unsaved, it'll ask if you want to save
02:25 < dooglus> but if it's not modified, it just quits
02:26 < dooglus> I made a few changes, and am committing now
02:26 < dooglus> there's no need to pass the root canvas or the filename if we have the instance handle
02:26 < dooglus> we can find the rest from the instance
02:26 < dooglus> and one of the 
02:27 < dooglus> methods can be private - it's only used internally
02:27 < Yoyobuae> yeah sure
02:27 < dooglus> private:
02:27 < dooglus> 	static void add_recent_file(const std::string &filename, const std::string &window_size = std::string());
02:27 < dooglus> public:
02:27 < dooglus> 	static void add_recent_file(const etl::handle<Instance> instance);
02:27 -!- pixelgeek [n=chatzill@c-71-59-140-184.hsd1.or.comcast.net] has joined #synfig
02:27 < dooglus> just takes an instance, ok?
02:27 < Yoyobuae> its the same really
02:28 < Yoyobuae> i just though passing the readily available root canvas handle was easier =)
02:30 < Yoyobuae> dooglus: do you think it would be better to use a 'settings' like file format for 'recent_files'
02:30 < Yoyobuae> ?
02:30 < dooglus> I guess it would be best to use XML for everything really
02:30 < dooglus> rather than all this custom parsing
02:31 < Yoyobuae> true.
02:31 < CIA-27> synfig: dooglus * r1801 /synfig-studio/trunk/src/gtkmm/ (11 files): Conditionally compile out the 'single threaded' stuff. It doesn't work very well still, so it's better not to present it to the user yet.
02:31 < CIA-27> synfig: dooglus * r1802 /synfig-studio/trunk/src/gtkmm/ (instance.cpp app.h app.cpp): Checked in a patch from Gerald Young (Yoyobuae) to save and restore the state (with canvases are open, and their position and size) for each of the recent files.
02:31 < dooglus> but that's for the next release I guess
02:31 < Yoyobuae> the settings format is a bit more readable IMHO =)
02:33 < dooglus> yes, I guess so
02:33 < Yoyobuae> although a good XML editor does wonders too
02:33 < Yoyobuae> i use one for editing SIF files =)
02:39 < dooglus> which one?
02:39 < dooglus> and, did you update from svn and build recently?
02:39 -!- pabs3 [n=pabs@d122-105-74-150.per9.wa.optusnet.com.au] has joined #synfig
02:40 < Yoyobuae> uhmm, a Windows one =)
02:40 < Yoyobuae> last night
02:41 < pixelgeek> cia says 1802, svn says 1803
02:41 < pixelgeek> Any more commits incoming dooglus ?
02:47 < pixelgeek> compiling now
02:47 < CIA-27> synfig: dooglus * r1803 /synfig-studio/trunk/src/gtkmm/app.cpp: Add Gerald Young's copyright.
02:51 < dooglus> pixelgeek: nope
02:51 < dooglus> Yoyobuae: another comment: it always opens the root canvas, whether I had it open when I saved or not
02:52 < Yoyobuae> oh, until now i though the root canvas HAD to be open
02:54 < pixelgeek> Hey - I've got that combination on my luggage! - http://uk.youtube.com/watch?v=yZHCQOmCMwg
02:55 < Yoyobuae> when the studio::Instance is created it automatically opens the root canvas
03:00 < dooglus> Yoyobuae: 'tis true, but we can hide it I guess?
03:02 < Yoyobuae> yeah we could. or simply don't show it unless necesary =).
03:02 < Yoyobuae> but then it would be necesary to make sure that at least the root is shown, if other canvases are not
03:05 < dooglus> don't you always save at least one canvas?
03:05 < rubikcube> In a new animation, exported values don't show up in the children dialog until a save and open again...
03:07 < dooglus> oh really?
03:07 < dooglus> I've never seen that
03:07 < dooglus> what version are you running?
03:08 < rubikcube> 61.07
03:08 < dooglus> hmm
03:09 < dooglus> "new animation -> circle tool -> draw circle -> export radius" doesn't show a '>' mark next to ValueBase Nodes in the Children panel?
03:09 < Yoyobuae> dooglus: it is possible that recent_files_window_size simply doesn't exists, in which case all the entries are assumed to be blank.
03:10 < rubikcube> hmm, can't reproduce it now...
03:10 < rubikcube> it was not expandable before...
03:10 < dooglus> it does in 0.61.06 and in current SVN;  I don't have 0.61.07 built here
03:11 < dooglus> Yoyobuae: ok, so:
03:11 < dooglus> 	if (shown_non_root && !seen_root)
03:11 < dooglus> 		instance->find_canvas_view(instance->get_canvas())->hide();
03:11 < dooglus> or similar
03:13 < Yoyobuae> i guess, maybe at the end of set_recent_file_window_size(). but then the returns would need to be changed
03:13 < Yoyobuae> it's simple enough
03:14 < dooglus> I don't think it returns unless something goes wrong does it?
03:18 < Yoyobuae> yeah it mostly used to stop reading from the string with the canvas positions/size
03:18 < dooglus> ok
03:18 < dooglus> the first 'return' becomes a 'break' then
03:19 < dooglus> committed.
03:19 < Yoyobuae> no! XD
03:19 < Yoyobuae> if the string is empty we leave the root canvas open
03:19 < Yoyobuae> so returning is ok
03:21 < Yoyobuae> do you mean the first return INSIDE the while (yeah because break wouldn't make sense otherwise) XD
03:22 < Yoyobuae> i kinda got confused for a while
03:22 < dooglus> :)
03:22 < dooglus> the first in the loop, yes
03:23 < pabs3> should we change pxegeek in the code to pixelgeek?
03:24 < Yoyobuae> i think the other returns could also become breaks.
03:25 < Yoyobuae> if recent_files_window_size is messed with, the other returns might be used.  but it's still possible for part of the canvases to be shown
03:27 < pixelgeek> I'll answer to both :)
03:27 < pixelgeek> Haven't figured out an easy way to update my name on the wiki
03:29 < pabs3> dooglus: you know there are already ETL/_mutex_pthreads.h and ETL/_mutex_win32.h right?
03:30 < pixelgeek> talking of which....
03:30 < pixelgeek> In file included from C:/msys/1.0/home/Chris/synfig/temp/ETL/include/ETL/handle:32,
03:30 < pixelgeek>                  from ../../../src/synfig/value.h:39,
03:30 < pabs3> dooglus: you also forgot to add _mutex_simple to the Makefile.am so it doesn't install and synfig won't build
03:30 < pixelgeek>                  from ../../../src/synfig/value.cpp:33:
03:30 < pixelgeek> C:/msys/1.0/home/Chris/synfig/temp/ETL/include/ETL/_handle.h:41:29: _mutex_simple.h: No such file or directory
03:31 < pabs3> I recon revert _mutex_simple.h and use the existing ones
03:31 < Yoyobuae> pixelgeek: i thinkg ETL needs updating
03:31 < pabs3> pixelgeek: yeah, thats what I see too
03:31 < pixelgeek> I did.
03:31 < CIA-27> synfig: dooglus * r1804 /synfig-studio/trunk/src/gtkmm/app.cpp: Hide the root canvas if we have saved the position and size of other canvases and not of the root canvas.
03:31  * pabs3 wonders how dooglus built synfig against the new ETL mutex
03:32  * Yoyobuae thinks dooglus has his own private ETL
03:32 < pixelgeek> It's magic, I tells ya!
03:34 < dooglus> pabs3: aah...
03:35 < pabs3> pixelgeek: should we change pxegeek in the code/images to pixelgeek?
03:35 < dooglus> pabs3: I don't install ETL, because doing so touches all the timestamps
03:36 < pixelgeek> I'll answer to either, but yes, I suppose going forwards....
03:36 < dooglus> that's how it builds for me - it uses the local copies
03:36 < pabs3> ah, I see
03:36 < dooglus> is adding the 2 headers to Makefile.in enough?
03:36 < pabs3> pixelgeek: doing the change, will leave svn logs and NEWS file though
03:37 < dooglus> I'm going to bed now, have committed Makefile.in for ETL
03:37 < dooglus> night all
03:37 < pabs3> dooglus: I'd suggest just dropping _mutex_simple.h in favour of  ETL/_mutex_pthreads.h and ETL/_mutex_win32.h
03:37 < pabs3> dooglus: but it is Makefile.am not .in
03:38 < dooglus> pabs3: _mutex_pthreads does more than is necessary for refcounting
03:38 < dooglus> I wanted a simpler implementation, since it's called a lot
03:39 < Yoyobuae> pabs3: dooglus is so good he doesn't need automake ;)
03:40 < dooglus> ha
03:40 < pabs3> :)
03:40 < pabs3> ah, ok
03:40 < dooglus> _mutex_win32 isn't included from anywhere either
03:41 < dooglus> and quite a few ETL headers have obvious typos
03:41 < dooglus> I don't know if pthread mutexes work on windows or not - I guess the library must be available though, right?
03:42 < pabs3> we seem to have 3 ETL mutexes, 1 synfig mutex and studio uses the glib mutex :)
03:46 -!- rubikcube [n=kvirc@dslb-088-070-012-149.pools.arcor-ip.net] has quit ["KVIrc 3.2.6 Anomalies"]
03:47 < CIA-27> synfig: dooglus * r1805 /ETL/trunk/ETL/Makefile.am: Include 'mutex' and '_mutex_simple.h' in the list of ETL headers to install.
03:53 < Yoyobuae> MingW doesn't seem to include pthreads.h
03:54 < pixelgeek> C:/msys/1.0/home/Chris/synfig/temp/ETL/include/ETL/_mutex_simple.h:33:21: pthread.h: No such file or directory
03:54 < pixelgeek> yup
03:54 < pabs3> I gues this is why _mutex_win32.h exists
03:55 < Yoyobuae> i found this though http://sourceware.org/pthreads-win32/
04:00 < pabs3> hmm
04:00 < pabs3> synfig(6281) [11:59:41] error: -42 canvases not yet deleted!
04:48 -!- pabs3 [n=pabs@d122-105-74-150.per9.wa.optusnet.com.au] has quit [Remote closed the connection]
04:52 -!- pabs3 [n=pabs@d122-105-74-150.per9.wa.optusnet.com.au] has joined #synfig
05:16 < CIA-27> synfig: pabs * r1806 /synfig-studio/trunk/ (16 files in 3 dirs): Replace PXEGeek with pixelgeek
05:27 < pixelgeek> Ah - my 15 minutes of fame.....
05:28 < pixelgeek> I wonder if I can change my Slashdot ID....
05:36 < pixelgeek> apparently not
05:36 < pixelgeek> so much for my exclusive 6 digit ID....
06:53 -!- AkhIL [n=AkhIL@90.188.195.134] has quit [Read error: 104 (Connection reset by peer)]
07:05 -!- AkhIL [n=AkhIL@90.188.195.134] has joined #synfig
07:57 -!- pixelgeek [n=chatzill@c-71-59-140-184.hsd1.or.comcast.net] has quit [Read error: 113 (No route to host)]
08:09 -!- Yoyobuae [n=Yoyobuae@201.224.135.156] has quit [Read error: 110 (Connection timed out)]
08:37 -!- ZanQdo [n=Daniel@201.201.2.22] has joined #synfig
09:38 -!- factor [n=factor@ip68-14-160-70.ok.ok.cox.net] has quit [Read error: 104 (Connection reset by peer)]
09:53 -!- factor [n=factor@ip68-14-160-70.ok.ok.cox.net] has joined #synfig
10:14 -!- factor [n=factor@ip68-14-160-70.ok.ok.cox.net] has quit ["Ex-Chat"]
10:22 -!- factor [n=factor@ip68-14-160-70.ok.ok.cox.net] has joined #synfig
10:25 -!- factor [n=factor@ip68-14-160-70.ok.ok.cox.net] has quit [Remote closed the connection]
10:31 < AkhIL> .us/img149/6692/synfigonpuppykx1.jpg - synfig on puppylinux. I'm searching host to share my pet packages
10:31 < AkhIL> ups
10:31 < AkhIL> http://img149.imageshack.us/img149/6692/synfigonpuppykx1.jpg
10:34 < cizra> JPG?
10:34 < cizra> You use JPEG for screenshots? Why?
10:35 < AkhIL> I have slow internet
10:35 < cizra> PNG is better. It makes smaller images.
10:35 < cizra> Try it.
10:35 < cizra> While you're at it, tell me too by how much it's different.
10:35 < AkhIL> ok
10:35 < cizra> JPEG is for photos, PNG is for graphics like screneshots.
10:36 < cizra> screenshots*
10:37 < AkhIL> 84K     synfig.on_puppy_test.jpg
10:37 < AkhIL> 212K    synfig.on_puppy_test.png
10:38 < cizra> Strange.
10:38 < AkhIL> but yes. quality is much better
10:38 < cizra> I really thought it's the other way around.
10:38 < dooglus> jpg is lossy, png isn't
10:38 < cizra> IIRC PNG can be configured to be lossy asw ell?
10:38 < AkhIL> png good if you have a lot of solid colors without gradients
10:39 < cizra> Ok, I did a screenshot of this very window. IRSSI, basically.
10:39 < cizra> -rw-r--r-- 1 elmo elmo 233615 2008-02-23 11:39 screenshot.jpg
10:39 < cizra> -rw-r--r-- 1 elmo elmo  11572 2008-02-23 11:39 screenshot.png
10:39 < cizra> There you go, 20 times smaller.
10:40 < AkhIL> try jpeg with quality 50
10:40 < cizra> I don't know how. I used imagemagick's import command.
10:40 < AkhIL> import -quality 50 blah.jpg
10:40 < cizra> OK, trying again
10:41 < cizra> -rw-r--r-- 1 elmo elmo 93763 2008-02-23 11:41 scr1.jpg
10:41 < cizra> -rw-r--r-- 1 elmo elmo 47040 2008-02-23 11:41 scr1.png
10:41 < cizra> PNG still wins by a factor of 2.
10:41 < pabs3> what window manager does puppy use?
10:42 < AkhIL> jwm
10:42 < AkhIL> just MS windows 95 like
10:43 < cizra> So, my dear AkhIL, I don't know how you do your screenshots, but in my tests PNG wins consistently and by a wide margin.
10:44 < AkhIL> $ import synfig.on_puppy_test.png
10:44 < AkhIL> $ import -quality 50 synfig.on_puppy_test.jpg
10:44 < dooglus> AkhIL's screenshot has lots of colors.  irssi generally doesn't
10:45 < cizra> AkhIL: You're being unfair. Why didn't you add -quality 50 to the PNG screenshot?
10:46 < AkhIL> cizra: because of for png quality is commpression rate like for zip
10:46 < AkhIL> anyway png lossless
10:46 < cizra> OK
10:49 -!- factor [n=factor@ip68-14-160-70.ok.ok.cox.net] has joined #synfig
10:51 < pabs3> fosdem stream is a talk about movies with linux: http://streaming.linux-magazin.de:8080/vortrag.ogg
10:54 < dooglus> pabs3: what can I use to watch it?
10:54 < dooglus> mplayer shows a static single frame
10:54 < pabs3> vlc, totem work here 
10:55 < dooglus> http://dooglus.rincevent.net/random/vlc.png
10:55 < dooglus> it's moving, but stripy
10:55 < pabs3> weird, works fine here
10:55 < dooglus> $ totem http://streaming.linux-magazin.de:8080/vortrag.ogg
10:55 < dooglus> The program 'totem' received an X Window System error.
10:55 < dooglus> This probably reflects a bug in the program.
10:55 < pabs3> vlc 0.8.6c
10:56 < dooglus> VLC media player 0.8.6c Janus
10:58 < pabs3> i386?
10:58 < dooglus> yes
10:59 < pabs3> hmm, old theora lib?
11:00 < dooglus> 1.0~beta2-2~gutsy1
11:00 < AkhIL> my mplayer plays few seconds and falls
11:00 < pabs3> 1.0~beta2-3 here
11:01 < dooglus> oh, realplay can play it
11:02 < dooglus> but without sound
11:02 < AkhIL> hehe vlc for windows under wine plays it 8)
11:02 < dooglus> "aoss realplay [url]" gives me sound and video together
11:04 < AkhIL> will they have this video for download?
11:05 < pabs3> not sure
11:10 < pabs3> slides http://streaming.linux-magazin.de/events/fosdem08/slide.htm
11:28 -!- ZanQdo [n=Daniel@201.201.2.22] has quit [Read error: 104 (Connection reset by peer)]
11:41 -!- xerakko [n=Miguel@debian/developer/xerakko] has joined #synfig
12:16 < CIA-27> synfig: dooglus * r1807 /ETL/trunk/ETL/ (6 files): Use _mutex_win32.h on Windows.
12:45 -!- AkhIL [n=AkhIL@90.188.195.134] has left #synfig []
12:47 -!- AkhIL [n=AkhIL@90.188.195.134] has joined #synfig
13:38 -!- xerakko [n=Miguel@debian/developer/xerakko] has quit ["Ex-Chat"]
14:15 < CIA-27> synfig: dooglus * r1808 /synfig-studio/trunk/src/gtkmm/app.cpp: 
14:15 < CIA-27> synfig: Use resize() to resize windows. set_default_size() doesn't work on the root
14:15 < CIA-27> synfig: window, since it has already been presented, and set_size_request() prevents the
14:15 < CIA-27> synfig: user from reducing the window's size. Also, wait until the size and position
14:15 < CIA-27> synfig: are set before presenting the window.
14:45 < factor> out to linux meeting
14:45 -!- factor [n=factor@ip68-14-160-70.ok.ok.cox.net] has left #synfig ["Ex-Chat"]