--- Log opened Mon Jan 21 00:00:30 2008
00:01 < pabs3> ulrik: yeah, upstream libavcodec changed their structures again, in older ffmpeg it works
00:01 < pabs3> iirc there is a bug in the tracker about this
00:16 < dooglus> http://synfig.org/forums/viewtopic.php?t=5 maybe?
00:17 < ulrik> dooglus: sorry i was away looking at "Kikis delivery service" by Hayato Miyaki
00:18 < ulrik> dooglus: yes I see them, strange I didn't found them before
00:18 < ulrik> pabs: ok i see, so that's why
00:18 < dooglus> ulrik: I didn't change anything, other than deleting a few 2007 logs from the 2008 folder, for months
00:19 < pabs3> dooglus: ah, yes
00:20 < ulrik> dooglus: I had you're logs bookmarked inside the 2007 folder...that's why I didn't found the 2008 years :)
00:24 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has quit []
00:26 < dooglus> ulrik: right :)
00:28 < ulrik> I installed synfig with ffmpeg deactivated and now I activated it and now i'm rendering to .avi using ffmpeg.... :O 
00:32 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has joined #synfig
00:32 < MangoFusion> kiki's devlivery service? ah yes i remember that :)
00:33 < ulrik> mangofusion: don't tell me how it ends, i'm not finished yet, it's a nice little story
00:38 < MangoFusion> haha
00:39 < ulrik> :) 
00:54 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has quit []
01:06 < pxegeek>  l
01:07  * pxegeek loved "howl's moving castle" and "Spirited away", but hasn't seen Kiki yet
01:08 < ulrik> pabs: how do i write my ./configure without libavcodec and the --prefix=/opt/local --exec-prefix=/opt/local, in what order?
01:08 < ulrik> hi pxegeek!
01:08 < ulrik> nice work on the demo reel! :) 
01:08 < pxegeek> hi
01:08 < pxegeek> thanks
01:08 < pabs3> ulrik: don't think the order matters. I'd put it at the end
01:08 < pxegeek> (not really here just procrastinating from work)
01:09 < ulrik> pabs: ok, i'll try that, thanks
01:10 < ulrik> pxegeek: yes i loved those miayko movies too, kiki's del.... is a bit different though but it looks like i will love that one too
01:15 < dooglus> the only one I saw was about some kind of animals with large testicles
01:16 < ulrik> dooglus:  :D 
01:16 < dooglus> http://en.wikipedia.org/wiki/Pom_Poko
01:16 < ulrik> I want to see that too....
01:17 < dooglus> "It's worth noting to Western viewers that prominent testicles are an integral part of the Tanuki folklore, and they are shown and referred to throughout the movie, and also used frequently in their shapeshifting"
01:18 < ulrik> lol
01:18 < ulrik> what is "pouches"?
01:21 < dooglus> like a kangaroo has
01:21 < dooglus> to hold its young
01:21 < dooglus> like a pocket
01:21 < ulrik> ahhh, in swedish it's "pung"
01:22 < pabs3> genete: saxon is a java thing: http://www.saxonica.com/ http://sourceforge.net/projects/saxon/
01:22 < genete> pabs3: so I should run it as a normal java app?
01:22 < dooglus> pabs3: did you get to understand the 'straight' blend in the end?
01:22 < pabs3> dooglus: nope
01:22 < pabs3> genete: I guess
01:24 < dooglus> pabs3: blending layer X with amount A onto layer Y will give 100A percent of layer X and 100(1-A) percent of layer Y
01:24 < dooglus> pabs3: so if amount is 1.0, the result should just be layer X.  layer Y should be completely invisible.
01:25 < pabs3> and so it is (for the area that the layer affects)
01:26 < dooglus> pabs3: if you want the "for the area that the layer affects" then that's what "Composite" is for.
01:26 < dooglus> Composite takes transparency into effect.  Straight doesn't - it's a straight blend of the two layers.
01:28 < genete> night all ............
01:28 -!- genete [n=Genete@84.122.45.30.dyn.user.ono.com] has quit ["Abandonando"]
01:44 < ulrik> night all!
02:09 < dooglus> night ulrik
02:17 < CIA-38> synfig: dooglus * r1400 /synfig-core/trunk/src/synfig/layer_pastecanvas.cpp: Don't avoid not using negated macro names, because they can be confusing. ie. use "#ifdef SYNFIG_CLIP_PASTECANVAS" rather than "#ifndef SYNFIG_NO_CLIP".
02:17 < CIA-38> synfig: dooglus * r1401 /synfig-core/trunk/src/synfig/canvas.cpp: Tidying. No functional change.
02:17 < CIA-38> synfig: dooglus * r1402 /synfig-studio/trunk/src/gtkmm/dock_canvasspecific.cpp: Stop the 'canvas_view_changed' messages.
02:17 < CIA-38> synfig: dooglus * r1403 /synfig-studio/trunk/src/gtkmm/dock_canvasspecific.cpp: Stop 3 more messages.
02:17 < CIA-38> synfig: dooglus * r1404 /synfig-studio/trunk/src/gtkmm/canvasview.h: Remove more debugging prints.
02:17 < CIA-38> synfig: dooglus * r1405 /synfig-core/trunk/src/synfig/canvas.cpp: Give an example .sif file to explain why the 'SYNFIG_OPTIMIZE_PASTE_CANVAS' code is #ifdef'ed out.
02:17 < CIA-38> synfig: dooglus * r1406 /synfig-core/trunk/src/synfig/layer_pastecanvas.cpp: Rearranged a little. The functionality shouldn't have changed.
02:17 < CIA-38> synfig: dooglus * r1407 /synfig-core/trunk/src/synfig/renddesc.h: Fixed a typo.
02:17 < CIA-38> synfig: dooglus * r1408 /synfig-core/trunk/src/synfig/context.cpp: Tidying.
02:17 < CIA-38> synfig: dooglus * r1409 /synfig-core/trunk/src/synfig/layer_pastecanvas.cpp: Add a comment giving a .sif file that shows why 'SYNFIG_CLIP_PASTECANVAS' is needed.
02:17  * pabs3 detects a git svn dcommit
02:18 < dooglus> heh
02:18 < dooglus> 10 commits, no changes.
02:19 < dooglus> pabs3: see pxegeek's example blend here: http://synfig.org/Blend_Method#Straight - the left hand layer is transparent at the bottom, yet still covers the penguin when blended straight
02:20 < pabs3> what kind of layer is the left one?
02:20 < dooglus> I guess it's a gradient
02:22 < dooglus> from (0,0,0,1) at the top to (1,1,1,1) in the middle to (1,1,1,0) at the bottom
02:22 < pabs3> I imagine it covers (not the alpha channel, but the layer itself) the whole canvas rather than a region then?
02:22 < dooglus> where I'm using (R,G,B,A)
02:22 < dooglus> all layers cover the whole canvas though, conceptually
02:23 < pabs3> hmmm
02:25 < AkhIL> alpha over useful for masking like rotoscoping of image
02:32 < CIA-38> synfig: dooglus * r1410 /synfig-core/trunk/src/synfig/target_tile.cpp: Don't display the frame efficiency.
02:47 < CIA-38> synfig: dooglus * r1411 /synfig-core/trunk/src/synfig/canvas.h: Document the canvas versions.
03:08 -!- pabspabspabs [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
03:11 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Nick collision from services.]
03:11 -!- pabspabspabs is now known as pabs3
03:18 -!- pabspabspabs [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
03:18 -!- photomuse [n=photomus@ics156-82.icsincorporated.com] has joined #synfig
03:22 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Nick collision from services.]
03:22 -!- pabspabspabs is now known as pabs3
04:01 -!- pabspabspabs [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
04:12 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Nick collision from services.]
04:13 -!- pabspabspabs is now known as pabs3
05:22 -!- pabspabspabs [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
05:26 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Nick collision from services.]
05:27 -!- pabspabspabs is now known as pabs3
06:32 -!- pabspabspabs [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
06:52 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Read error: 110 (Connection timed out)]
06:57 -!- pabspabspabs is now known as pabs3
07:40 -!- pxegeek [n=chatzill@c-71-59-140-184.hsd1.or.comcast.net] has quit [Read error: 113 (No route to host)]
07:50 -!- pabspabspabs [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
08:13 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Nick collision from services.]
08:13 -!- pabspabspabs is now known as pabs3
08:17 -!- pabspabspabs [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
08:23 -!- Netsplit calvino.freenode.net <-> irc.freenode.net quits: CIA-38
08:23 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Nick collision from services.]
08:23 -!- pabspabspabs is now known as pabs3
08:40 -!- TMM [n=hp@ip565b35da.direct-adsl.nl] has quit [Read error: 110 (Connection timed out)]
08:44 -!- CIA-33 [n=CIA@208.69.182.149] has joined #synfig
09:05 -!- CIA-33 [n=CIA@208.69.182.149] has quit [Excess Flood]
09:11 -!- CIA-46 [n=CIA@208.69.182.149] has joined #synfig
09:31 -!- ZanQdo [n=Daniel@201.201.2.22] has joined #synfig
10:51 -!- ulrik [n=ulrikbod@213-65-95-57-no42.tbcn.telia.com] has quit ["bye! http://www.musikboden.se"]
12:30 -!- ZanQdo [n=Daniel@201.201.2.22] has quit ["Adios"]
13:06 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has joined #synfig
13:27 -!- Netsplit calvino.freenode.net <-> irc.freenode.net quits: MangoFusion, photomuse, pabs3, dyloxin
13:27 -!- Netsplit calvino.freenode.net <-> irc.freenode.net quits: KiBi, @ChanServ, htodd, AkhIL, rore
13:55 -!- ChanServ [ChanServ@services.] has joined #synfig
13:55 -!- ServerMode/#synfig [+o ChanServ] by irc.freenode.net
13:55 -!- htodd [n=htodd@i8u.org] has joined #synfig
13:55 -!- KiBi [n=kibi@kibi.dyndns.org] has joined #synfig
13:56 -!- AkhIL [n=AkhIL@90.188.221.229] has joined #synfig
13:56 -!- rore [n=rore__@d77-218-34-5.cust.tele2.fr] has joined #synfig
13:57 -!- pabspabspabs [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
13:57 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
13:57 -!- photomuse [n=photomus@ics156-82.icsincorporated.com] has joined #synfig
13:57 -!- dyloxin [n=dyloxin@c-71-198-19-209.hsd1.ca.comcast.net] has joined #synfig
14:08 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Read error: 110 (Connection timed out)]
14:15 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has joined #synfig
14:23 -!- Netsplit calvino.freenode.net <-> irc.freenode.net quits: KiBi, @ChanServ, pabspabspabs, photomuse, htodd, dyloxin, AkhIL, rore
14:24 -!- Netsplit over, joins: ChanServ
14:24 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
14:24 -!- Netsplit over, joins: dyloxin, photomuse, pabspabspabs, rore, AkhIL, KiBi, htodd
14:24 -!- ServerMode/#synfig [+o ChanServ] by irc.freenode.net
14:24 -!- Netsplit calvino.freenode.net <-> irc.freenode.net quits: KiBi
14:25 -!- Netsplit calvino.freenode.net <-> irc.freenode.net quits: htodd, AkhIL, rore
14:25 -!- Netsplit over, joins: htodd, KiBi
14:25 -!- Netsplit over, joins: AkhIL, rore
14:42 -!- pabspabspabs [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Read error: 110 (Connection timed out)]
15:30 -!- pabspabspabs [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
15:30 < dooglus> layer_shape.cpp:1284 says "//Ok, we're inside... bummmm ba bum buM..." :)  wtf?
15:30 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Nick collision from services.]
15:30 -!- pabspabspabs is now known as pabs3
15:30 < dooglus> pabs3: can you confirm this strange behaviour?
15:31 < dooglus> load examples/japan.sifz into studio, select the top layer (circle), edit the 'blend method' to anything - the parameter dialog immediately reflects the change, then edit it back to 'straight', and it appears not to change back at first
15:31 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has quit []
15:33 < pabs3> dooglus: straight -> alpha over -> straight was fine here
15:34 < pabs3> subtract too
15:34 < dooglus> ok
15:34 < dooglus> strange
15:37 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has joined #synfig
15:42 < dooglus> pabs3: I'm seeing that it is possible to set the blend method to 'straight', but that it takes an extra click after selecting the value from the menu before it updates on screen
15:43 < pabs3> one click on the menu item and it updates fine for me
15:43 < dooglus> pabs3: in fact, I'm seeing that whenever I try to set the blend-method to the 'same' as the current 'falloff' selection, it needs an extra click
15:44 < dooglus> falloff=cosine is '2', as is blend-method=brighten
15:44 < dooglus> so if I set falloff to cosine, I can't set blend to brighten immediately
15:44  * pabs3 tries that
15:44 < dooglus> maybe a difference in window managers is causing me to need the extra click
15:46 < pabs3> definitely not seeing the cosine/brighten thing
15:46 < pabs3> I'm using metacity, with the mist theme
15:46 < dooglus> ok
15:46 < dooglus> focus follows mouse?
15:46 < pabs3> nope
15:46 < dooglus> click to type then?
15:46 < pabs3> yeah
15:46 < dooglus> ok
15:47 < dooglus> I'll try that
15:47 < pabs3> that is probably it if you are using focus follows
15:48 < dooglus> I'm using xfce4 too, and xfwm
15:48 < pabs3> ok, I'm in gnome
15:49 < dooglus> even with click-to-focus I see the same - and even in studio 0.61.06
15:49 < pabs3> which gtkmm?
15:49 < dooglus> and I see "synfig(25279) [15:41:23] error: on_editing_done(): Called twice!" in the console after the 2nd click
15:49 -!- AkhIL [n=AkhIL@90.188.221.229] has quit [Read error: 110 (Connection timed out)]
15:49 < dooglus> ii  libgtkmm-2.4-1c2a                   1:2.12.0-0ubuntu1                   C++ wrappers for GTK+ 2.4 (shared libraries)
15:49 < dooglus> ii  libgtkmm-2.4-dev                    1:2.12.0-0ubuntu1                   C++ wrappers for GTK+ 2.4 (development files)
15:50 < pabs3> I don
15:50 < pabs3> 't see that message, this isn't a debug build though
15:50 < dooglus> good point - I'll try my -O build
15:51 < dooglus> exactly the same :)
15:51 < pabs3> I'm thinking I should setup automatic daily sid/hardy .deb builds with full debugging info
15:52 < dooglus> which gtkmm are you using?
15:52 < dooglus> I guess this will be a studio bug, but I don't know why you're not seeing it
15:53 < pabs3> gtkmm 2.12.3-2 gtk 2.12.5-1
15:53  * dooglus boots up the poor old sid laptop
15:53 < pabs3> you get your other laptop fixed?
15:54 < dooglus> the one that broke is the new one, and yes, I sent it away and it came back with a new hdd
15:54 < dooglus> the new one does the same 'tick every 5 seconds' thing that the old one did just before failing
15:54 < dooglus> but I reboot now as soon as it starts happening, and it's been OK.
15:54 < dooglus> it only does it once a month or so
15:55 < pabs3> :(
15:55 < dooglus> I thought it might be ext3 syncing to disk, but I'd had it happen in windows too, and even once at the grub menu
15:57 < dooglus> I've made a bunch of changes for the 'straight' blend method, and a new canvas version.  old .sif files will have existing 'straight' blends changed to 'composite', since 'straight' used to act like composite
15:57 < dooglus> that should allow existing sif files to keep working as intended
16:00 < MangoFusion> need to get one of those solid state disks
16:01 < MangoFusion> no moving parts to break ;)
16:01 < MangoFusion> neat
16:01 < pabs3> how big do they go these days?
16:03 < MangoFusion> last i heard it was something like 64gn
16:04 < MangoFusion> *gb
16:04 < MangoFusion> though still too expensive
16:04 < pabs3> aye
16:07 < dooglus> they still wear out though, right?  like flash memory does?
16:08 < dooglus> I read somewhere that each bit on a flash drive has a lifetime of about 1 million write cycles
16:08  * pabs3 wonders about that too
16:09 < dooglus> if so, you'd have to use a filesystem that spread the writes out pretty evenly.
16:09 < dooglus> I'm guessing the current ones write certain parts of the disk over and over, and other parts hardly at all
16:10 < MangoFusion> hard disks wear out too, y'know
16:10 < dooglus> true enough
16:10 < pabs3> yeah, there is JFFS and similar to do write leveling though dooglus 
16:10 < dooglus> oh, I didn't hear about that
16:13 < pabs3> http://en.wikipedia.org/wiki/JFFS2
16:13 < pabs3> there have been some articles about these flash-oriented filesystems on LWN a while ago
16:21 < dooglus> interesting
16:21 < dooglus> the old laptop has built core already - it's not as slow as I remember it being
16:22 -!- pabspabspabs [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
16:23 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Nick collision from services.]
16:23 -!- pabspabspabs is now known as pabs3
16:27  * pabs3 zzz
16:29 -!- genete [n=Genete@84.122.45.30.dyn.user.ono.com] has joined #synfig
16:29 < genete> hi 
16:31 < dooglus> hi genete 
16:31 < genete> hi dooglus 
16:32 < dooglus> genete: can you try reproducing the bug I told pabs3 about earlier please?
16:32 < dooglus> (at 15:31)
16:32 < genete> the japan one?
16:32 < dooglus> yes
16:32 < genete> It works fine for me.
16:32 < dooglus> ha
16:32 < dooglus> ok
16:32 < genete> I have a blend method question:
16:33 < genete> layer A have a determined blend method 
16:33 < genete> if I encapsulate layer A (single one layer) and make the paste canvas have the same method should I have same result?
16:34 < genete> (assume that layer A method is Composite when encapsulated
16:34 < dooglus> no
16:34 < dooglus> oh, sorry
16:35 < dooglus> A has blend method x, right?
16:35 < genete> yes
16:35 < dooglus> encapsulate A, set blend(A)=composite and blend(encap)=x
16:35 < genete> yes
16:35 < dooglus> funny you should ask
16:35 < dooglus> I wrote code to do exactly that, yesterday
16:36 < genete> I ask because my current version doesn't 
16:36 < dooglus> you forgot to set the amount
16:36 < genete> and changes from one type of layer to other
16:36 < dooglus> you need to set amount(encap) = amount(A); amount(A)=1.0
16:36 < genete> amount = 1 any situation
16:36 < dooglus> ok
16:37 < genete> for example text layer acts the same, circle layer not
16:37 < dooglus> with which blend method?
16:37 < genete> straight onto
16:38 -!- AkhIL [n=AkhIL@90.188.221.229] has joined #synfig
16:39 < dooglus> ok, thanks
16:39 < dooglus> that's very related to a bug I'm fixing at the moment
16:39 < dooglus> any other layers?
16:39 < genete> yes 
16:39 < dooglus> I mean - any other blend methods?
16:39 < genete> I'm testing... them layers vs methods
16:40 < dooglus> ok
16:40 < dooglus> I have some changes which work quite well, but they break other stuff
16:40 < genete> :/
16:40 < dooglus> I don't know whether to commit them so you can test, or leave it for a while until it's working more fully
16:41 < genete> what were the changes?
16:42 < dooglus> I'm fixing the 'straight' layer
16:42 < dooglus> 'straight' is supposed do a blend ignoring transparency
16:43 < dooglus> so at amount=1, a straight blend will completely cover all lower layers
16:43 < dooglus> even the transparent parts will block out the stuff below
16:44 < dooglus> like in the wiki example of the penguin for straight blend - the transparent part of the gradient still hides the penguin
16:44 < genete> gradieent + tux example is fine 
16:44 < genete> but not other layers
16:45 < dooglus> oh?
16:45 < dooglus> I mean - what do you mean?
16:46 < genete> I mean that gradient (straight) works fine but other layer (circle or others) doesn't
16:47 < genete> for example a feathered circle should show similar behavior as a radial gradient but it doesn't
16:47 < genete> a radial gradient form solid (center) to transparent (out)
16:51 < dooglus> did you notice that circles don't have a 'type of feather' menu, but regions do?
16:52 < genete> they have but different fucntions
16:52 < genete> ah
16:52 < genete> regions
16:52 < genete> eh?
16:52 < genete> regions have a feather method and circles a fall off type
16:52 < dooglus> ok
16:52 < genete> what's the difference?
16:53 < genete> I see none
16:53 < dooglus> I don't know.  maybe nothing :)
16:53 < dooglus> I'm looking at your radial gradient example
16:53 < dooglus> checking whether it's fixed here now
16:53 < genete> for get it
16:53 < genete> I'm messing the methods
16:54 < genete> anyway try it
16:54 < genete> .... to mubch brink buring lungh...
16:56 < dooglus> it looks fixed to me
16:56 < dooglus> to much to drink during lunch?
16:57 < genete> "too much drink during lunch"
16:57 < genete> ;)
16:59 < dooglus> ha, thanks for the grammar lesson :)
16:59 < genete> ;)
17:00 < genete> I'm reproducing the same examples than pxegeek did but with other layers and exncapsulated vs non encapsulated
17:07 < dooglus> I don't know if that's worth doing without my new changes
17:08 < dooglus> because I've changed quite a lot how that works
17:08 < dooglus> for Straight, anyway
17:08 < genete> I think it is worth becasue I can test your changes
17:08 < dooglus> and will have to do StraightOnto too, by the look of it
17:08 < dooglus> ok, right.  that would be useful, yes
17:08 < genete> I'm making an automatic file
17:08 < dooglus> we could use a good simple test suite - a collection of .sif files we can test synfig against after any change
17:09 < dooglus> to make sure it still loads and renders them properly
17:09 < dooglus> even one test per file is ok
17:09 < dooglus> that's probably better than using multiple frames
17:09 < dooglus> makes it easier to automate the tests I think
17:27 < genete> dooglus: where is defined the blend methods numbering?
17:34 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has quit []
17:34 < dooglus> genete: color.h
17:34 < dooglus> genete: can you VNC to me so I can try that param thing?
17:34 < dooglus> x11vnc -connect 89.176.28.156:55505
17:34 < genete> yes
17:35 < dooglus> genete: I just built studio on debian sid and see the same there...
17:39 < genete> what are you looking for?
17:39 < dooglus> genete: if I set the falloff to 'cosine', I can't set blend method to 'lighten'
17:39 < dooglus> on either of my machines here
17:39 < dooglus> but you can...
17:40 < dooglus> thanks for letting me test
17:40 < genete> ok it is fine
17:40 < dooglus> I just didn't know if you had understood what I was seeing :)
17:40 < genete> aaah
17:41 < genete> dooglus: did you know that there are some methods not used?
17:41 < dooglus> blend methods?  yes
17:41 < genete> but if you export the param you can set them ...
17:43 < dooglus> just alpha_{brighten,darken} isn't it?
17:44 < genete> yes
17:44 < dooglus> the code says they're deprecated, but not why
17:44 < genete> also I have a crash if set it to 23
17:45 < dooglus> cool
17:45 < dooglus> new bugs :)
17:47 < dooglus> so what should it do?
17:47 < genete> I don't know I couldn't render
17:47 < dooglus> it's an illegal value
17:48 < dooglus> BLEND_END=22
17:48 < dooglus> 22 is illegal, too
17:48 < genete> ok
17:49 < genete> should it be cyclic?
17:49 < genete> 22=0 =
17:49 < genete> ?
17:50 < dooglus> no
17:50 < dooglus> it's possible that a version of synfig in the future defines blend methods 22 and 23
17:50 < dooglus> we want synfig to complain and refuse to load the .sif file I guess
17:51 < genete> yes it is true
17:53 < dooglus> this is weird - I just switched to GNOME and still see that problem with 'cosine + brighten'
17:56 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has joined #synfig
18:03 < genete> http://www.darthfurby.com/genete/synfig/blendtest.sifz
18:03 < genete> dooglus:  can you render that file? I cannot render frame 21 it stops at frame 20 (png sequence)
18:04 < genete> I need to set render end time to 23f (it crashes) to render the last frame. What the hell???
18:04 < dooglus> I'll try
18:05 < dooglus> I just checked in a fix for your method=23 bug
18:05 < genete> ok nice
18:05 < dooglus> it just uses Composite instead, and spits out a warning
18:05 < dooglus> (warnings just go to the terminal at the moment, but once that's improved, it'll be more obvious)
18:06 < dooglus> it's stopping after making blendtest.0020.png
18:06 < genete> yes, why?
18:06 < genete> it is set from 0f to 21f
18:07 < dooglus> it never renders the last frame?
18:07 -!- Zelgadis [n=zelgadis@87.103.170.242] has joined #synfig
18:07 < genete> on?y of you set it to 23f (then crashes)
18:07 < dooglus> I just made a simple 4fps sif from 0s to 1s
18:08 < dooglus> it rendered 4 frames (0, 1, 2, 3)
18:08 < dooglus> it never renders the last one
18:09 < genete> ah so I should set the end time to 24f and render from 0f to 21f?
18:09 < genete> weird
18:09 < dooglus> when does it crash?
18:09 < genete> at 23f
18:10 < genete> when render ot place the time cursor
18:10 < MangoFusion> it's one of those for (i=0; i<frame_count; i++) render_frame(i); technicalities again ;)
18:11 < dooglus> I'm not seeing a crash
18:12 < genete> I'm discovering more bugs with other blend methods....
18:12 < dooglus> can you tell me how to see it?
18:12 < genete> the crash?
18:12 < dooglus> yes
18:12 < genete> Render from 0f to 23f it crashes at 23f
18:14 < dooglus> I typed "synfig blend_test.sif"
18:14 < dooglus> it rendered all 24 frames without a crash
18:14 < genete> I did it with studio
18:14 -!- Zelgadis [n=zelgadis@87.103.170.242] has quit ["Leaving."]
18:15 < dooglus> start time 0 end time 1?
18:15 < dooglus> or what?
18:15 < genete> start time = 0f end time = 23f
18:15 < genete> png sequence
18:15 < dooglus> that won't try to render frame 0023 though
18:15 < dooglus> will it?
18:15 < genete> maybe the buggy one is 22
18:15 < genete> http://i85.photobucket.com/albums/k74/Genete/synfig/noise-gradient0012.png
18:16 < dooglus> I think it would help to have a white background
18:16 < dooglus> I view the rendered images on a black background, and can't see the text
18:17 < genete> that is a single noise gradient with Behind method (encapsulated left; non encapsulated right)
18:17 < CIA-46> synfig: dooglus * r1412 /synfig-core/trunk/src/synfig/layer_pastecanvas.cpp: Don't use the optimization of using STRAIGHT blends instead of COMPOSITE when SYNFIG_CLIP_PASTECANVAS isn't defined, because it causes incorrect rendering in, for example, the 'examples/pirates.sifz' file.
18:17 < CIA-46> synfig: dooglus * r1413 /synfig-core/trunk/src/synfig/canvas.cpp: Declare the paste_canvas pointing in the block that uses it to limit its scope.
18:17 < CIA-46> synfig: dooglus * r1414 /synfig-core/trunk/src/synfig/layer_composite.cpp: Don't accept illegal values for 'blend_method'. Use 'Composite' instead.
18:17 < genete> http://s85.photobucket.com/albums/k74/Genete/synfig/?action=view&current=noise-gradient0012.png
18:17 < genete> try this
18:18 < dooglus> hmm
18:18 < dooglus> sif?
18:18 < genete> that file!
18:18 < genete> it has all kind of layers inside
18:18 < genete> just expand it! :)
18:19 < genete> you can check on the layers in pairs and see all the effects in one show
18:20 < dooglus> is there some way to turn them on and off from the command line?
18:20 < genete> aha! that's what I would ask you!
18:21 < genete> remember the "render this layer" option? 
18:22 < dooglus> no
18:22 < dooglus> :)
18:23 < dooglus> but for automated testing, it's no good to have to toggle layers on and off with the gui
18:23 < genete> you can save one file per type of layer
18:23 < dooglus> it's possible to add a new command line switch to do it, but how do we refer to each layer?
18:23 < dooglus> they don't have unique names
18:24 < dooglus> it would be good to use a single file, to make it easier to change things
18:24 < dooglus> (like how the 'lock' icons are 4 icons in 1 file maybe?)
18:24 < dooglus> (that's done using canvases)
18:25 < genete> (one off topic thing but not want to forget). I would like to have a special layer that render a time stamp as a text layer
18:26 < genete> you can create a fingle sifz file to do all the job but it is time consuming...
18:29 < genete> http://s85.photobucket.com/albums/k74/Genete/synfig/?action=view&current=checkboard0012.png
18:30 < genete> Behind have problems ...
18:31 < genete> http://s85.photobucket.com/albums/k74/Genete/synfig/?action=view&current=checkboard0021.png
18:32 < genete> I need to go.... see you!
18:32 < genete> ttyl
18:32 -!- genete [n=Genete@84.122.45.30.dyn.user.ono.com] has quit ["Abandonando"]
18:32 < MangoFusion> seems to be no end to these bugs
18:32 < MangoFusion> :(
19:03 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has quit []
19:09 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has joined #synfig
19:36 -!- TMM [n=hp@c5147518c.cable.wanadoo.nl] has joined #synfig
20:37 -!- xerakko [n=Miguel@debian/developer/xerakko] has joined #synfig
20:56 < dooglus> MangoFusion: astonishing, isn't it
20:57 < dooglus> MangoFusion: I'm guessing lots of these bugs are the same thing really, but it has been going on for quite a while now
20:58 < MangoFusion> i suppose
22:07 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has quit []
22:25 < dooglus> genete: the crash you were seeing earlier was because you used a 'linear' valuenode to step the blend method from 0 up to 23.  I couldn't see the crash because I had already fixed the code to check that blend_method < BLEND_END before using the value.
22:37 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has joined #synfig
22:40 -!- genete [n=Genete@84.122.52.3.dyn.user.ono.com] has joined #synfig
22:43 < genete> hi (again)
23:00 < dooglus> hi again :)
23:00 < dooglus> do you have a list of the problems you found earlier?
23:00 < genete> no yet
23:00 < dooglus> ok
23:01 < genete> I've just plugged the computer
23:01 < dooglus> I'm getting happier with my fixes
23:01 < dooglus> nearly ready to commit
23:01 < genete> nice!
23:01 < dooglus> didn't fix anything for straight-onto, but 'straight' is looking good
23:01 < genete> I'm testing dup layer speed outside the main window
23:01 < dooglus> ok
23:01 < genete> I think it is the last issue with duplicate.
23:07 < dooglus> ok
23:07 < dooglus> I was just looking at the alpha_brighten and alpha_darken
23:07 < dooglus> trying to see what they do
23:07 < genete> they are nice ones. 
23:07 < dooglus> "alpha brighten" makes things more transparent
23:08 < dooglus> "alpha darken" makes things less transparent
23:08 < dooglus> right?
23:08 < genete> it seems to do that
23:08 < dooglus> I think there might be - surprise surprise! - a bug in one of them
23:09 < dooglus> looking at the code, for a_brighten, the Amount is multiplied by the context alpha, not the layer alpha
23:10 < dooglus> I'll leave that for now, until I maybe understand what it's trying to do better
23:11 < genete> I think that what should be fized are the current available ones.
23:11 < genete> s/fized/fixed
23:13 -!- pabspabspabs [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
23:13 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Nick collision from services.]
23:13 -!- pabspabspabs is now known as pabs3
23:14 < dooglus> yeah
23:14 < dooglus> straight-onto is the only one I see problems with no
23:14 < dooglus> w
23:14 < genete> yeah! :)
23:15 < genete> are you using my sample file?
23:16 < dooglus> yes
23:16 < dooglus> only checked with 'circle' so far though
23:17 < dooglus> do some problems show up with other layers?
23:17 < genete> Behind and checkboard
23:18 < dooglus> the two checkerboards maybe interfere with each other?
23:19 < genete> no
23:19 < genete> sure
23:19 < genete> it is "behind" not a "color type" blend method
23:19 < genete> and they are at separeted canvases
23:20 < dooglus> separate, but composed on top of each other
23:20 < dooglus> each checkerboard goes being its own candy
23:21 < genete> yes
23:21 < dooglus> but the front checkerboard goes in front of the back candy
23:21 < genete> each one is behind each candy
23:21 < genete> right. dooglus = 1 genete =0
23:21 < dooglus> ok, let's start again.  what's the problem you see with behind and checkerb...   yay :)
23:33 < genete> render a layer outside the main window is only 1.2 times faster
23:33 < genete> doing the same with normal layers is 28 times faster.
23:34 < genete> I'm not sure if dup layer renders the layer if it is outside of the window render area though
23:41 < dooglus> how many dups?
23:41 < genete> 1000
23:41 < genete> feathred circles
23:41 < genete> random pos
23:42 < dooglus> the dup layer has to set the Index parameter to 1000 different values to see if any of the feathered circles are inside the main window
23:42 < genete> it is really a problem...
23:43 < genete> maybe the layer itself should decide if render or not depending of the current position once the index is updated?
23:43 < dooglus> I think really the Context code does it
23:44 < genete> look: 4000 circles from 4000 layers takes 3.5 seconds to render 12 frames if they are visible and 0.33 seconds if not visible
23:45 < genete> numbers are aproximate
23:45 < genete> and 1000 circles from a dup layer took 14 secs if non visible and 17 if visible
23:45 < genete> what do you think?
23:46 < genete> using synfig -b option
23:51 < genete> dooglus: have you fixed Straight for all kind of layers?
23:57 -!- ulrik [n=ulrikbod@213-65-95-57-no42.tbcn.telia.com] has joined #synfig
--- Log closed Tue Jan 22 00:00:30 2008