--- Log opened Mon Jan 14 00:00:24 2008
--- Day changed Mon Jan 14 2008
00:00 < genete> any usage of blending there?
00:00 < dooglus> notice how the navigator shows 2 flowers, but the workarea shows 4
00:00 < dooglus> no, just composite
00:00 < genete> ah yes
00:00 < genete> you have to make hard duplications to see the bug
00:01 < dooglus> that was the idea
00:01 < dooglus> I disabled the curvegradient because it made it too slow to render
00:01 < dooglus> ha
00:02 < genete> dooglus: btw, when duplicate a layer, what renders first the "from" or the "To" index?
00:02 < dooglus> I think the whole time I was trying to reproduce the bug yesterday, I had 'single thread' enabled!
00:02 < dooglus> I just realised!!
00:02 < genete> xD
00:02 < dooglus> I turned if off once I'd given up, and didn't even think about it
00:02 < dooglus> genete: from
00:02 < dooglus> genete: that's why I called it 'from' - it's where the index starts from
00:03 < dooglus> and to is "to where it goes" :)
00:03 < genete> I still wondering that some layers should not be duplicated in some way
00:03 < dooglus> genete: it will go backwards if necessary.  the sign of 'step' is ignored
00:03 < genete> yes I've noticed it 
00:04 < dooglus> I documented it.  I think it's ok that way.
00:04 < genete> but can you reverse that flowers?
00:04 < genete> I mean the small petals before and the big after?
00:05 < genete> I always get a mess and have the sam figure...
00:05 < dooglus> I draw the big ones first
00:05 < dooglus> 'cos I'm using 'composite'
00:05 < dooglus> so I have to draw the background before the foreground
00:05 -!- omry [n=omry@bzq-84-108-20-56.cablep.bezeqint.net] has joined #synfig
00:05 < genete> so you did from big to small
00:05 < dooglus> yes
00:06 < genete> ok I have to play more with the indexes...
00:06 < genete> coming back to the two days ago conversation...
00:06 < dooglus> look at the 'Duplicate Layers' layer:  From (biggest layer zoom) To (smallest layer zoom)
00:06 < genete> I would like that some layers were not duplicated.
00:07 < genete> we were talkig about how to tell that to Dup Layer.
00:07 < dooglus> right
00:07 < genete> It would allow some layers interfere with the Z Depth of the duplicated.
00:08 < genete> because they are a the same canvas scope level
00:08 < dooglus> the dup layer could just have a parameter for how deep to work
00:08 < genete> I think It can be done now but it is a waste of time to render N times a layer (what can be big) for nothing
00:09 < genete> explain it more...
00:09 < dooglus> new parameter in "Duplicate" layer, which is an integer, and say how many of the layers under it it duplicates
00:10 < dooglus> so if you have Dup then circle then rectangle, and this new 'depth?' parameter is '1', it will dup only the circle, rendering the rectangle (and anything under the rectangle) only once
00:10 < dooglus> but if the Depth param was 2 or more, it would dup both the circle and the rect
00:11 < genete> I would let the layer itself if it will be duplicated or not without dependence of its relative Z position to the other layers under Dup.
00:11 < genete> decide*
00:11 < dooglus> don't forget about nesting though
00:11 < genete> I mean you would want the non duplicated layer to be the first one
00:11 < dooglus> I don't understand
00:12 < dooglus> can you give an example?
00:12 < genete> Dup/Circle/Rectangle
00:12 < genete> and the circle not duplicated and the rectange yes
00:12 < dooglus> just move the dup down one slot
00:13 < genete> "Something at the Circle layer that stop Dup to duplicate it
00:13 < genete> but I want the rectangles to be behind
00:13 < dooglus> circle/dup/rect
00:13 < genete> Dup/Rectangle/Circle/Rectangel
00:13 < genete> and the circle not duplicated
00:13 < dooglus> ok
00:13 < genete> a "sandwich"
00:14 < dooglus> suppose all blend is 'composite' and each dup makes 2 copies
00:14 < dooglus> where will the circle be?
00:14 < dooglus> dup/recta/circle/rectb
00:14 < dooglus> renders: recta1 recta2 circle rectb1 rectb2
00:14 < dooglus> in which order?
00:14 < genete> in that order
00:15 < genete> that's why I asked about the Z order of duplciate
00:16 < genete> what's the order for this?:
00:16 < genete> dup/circle/rectangle/region
00:16 < genete> if all are normal layers?
00:16 < genete> and dupliction is 2 times?
00:17 < dooglus> currently: c1,rt1,rg1, c2,rt2,rg2
00:17 < genete> and why not c1, c2, rt1, rt2, rg1, rg2?
00:17 < dooglus> because rendering is depth first?
00:18 < genete> It is something that confused me from the firt day
00:18 < genete> how do Dup know when go deep or not in nested layers?
00:19 < dooglus> all dup does is sets its index to each value in turn and renders the stuff under it, then composes the results
00:19 < dooglus> dup never asks the rectangle or region to render
00:19 < dooglus> dup just asks the circle to render
00:19 < dooglus> the circle renders the rectangle
00:19 < dooglus> and the rectangle renders the region
00:19 < dooglus> it's recursive
00:20 < genete> ouch!
00:20 < dooglus> so the dup can't render just the circle
00:20 < genete> so Dup layer only throw a thread and it is spreaded by  the layers!
00:20 < dooglus> the dup tells the circle to render, which really means 'circle plus everything under the circle'
00:20 < genete> in a recursive way!
00:21 < dooglus> it's not a thread - just a function call, but yes.
00:21 < genete> interesting
00:21 < genete> ok it is a way of speak
00:21 < genete> for each index call the stack to render"
00:21 < genete> isn't it?
00:23 < dooglus> the code in layer_duplicate.cpp is basically:
00:23 < dooglus>         do
00:23 < dooglus>         {
00:23 < dooglus>                 // render stuff under us
00:23 < dooglus>                 // compose it
00:23 < dooglus>         } while (duplicate_param->step());
00:24 < genete> fine
00:24 < dooglus> KiBi: http://kibi.dyndns.org:8083/~dooglus/gitweb.pl?p=synfig;a=shortlog isn't working?
00:24 < dooglus> "The connection to the server was reset while the page was loading."
00:25 < dooglus> genete: that makes nested dup work without extra code.  "stuff under us" might include another "dup" layer, but the parent dup layer needn't know.  it's just "context"
00:25 < dooglus> just layers that know how to render themselves
00:26 < dooglus> genete: the idea of stopping the dup doesn't really fit with how synfig is designed
00:26 < genete> So my intentions are void of content
00:26 < dooglus> just like a blur layer can't selectively blur only some layers
00:26 < dooglus> a twirl layer can't selectively twirl, etc.
00:27 < genete> because not possible to make a "sandwich"
00:27 < dooglus> your intention is good - you want to be able to put something 'under' the duplicated layers
00:27 < dooglus> the way we generally do that is with encapsulation layers
00:27 < genete> under, in the middle, avobe...
00:28 < genete> yes but I want to pass through the duplicated layers like in my first smoke example...
00:28 < dooglus> are you thinking of something like your text moving through the smoke?
00:28 < dooglus> ha
00:28 < dooglus> yes
00:28 < dooglus> so:  (encap: dup smoke-particle) text (encap: dup smoke-particle)
00:28 < dooglus> two dups, each in encaps
00:28 < dooglus> text isn't dup'ed
00:29 < genete> yes you told that the first time.
00:29 < dooglus> it's a workaround I guess
00:29 < genete> yeah
00:29 < dooglus> you want a more natural way to do it
00:29 < genete> but don't want to press you more with that
00:29 < dooglus> now you know more about how synfig works internally
00:30 < genete> yes
00:30 < dooglus> maybe you can have an idea about how it can be done
00:30 < genete> no for the moment
00:30 < dooglus> I'll think about it
00:30 < dooglus> I'm going to sleep now :)
00:30 < dooglus> night!
00:31 < genete> hey! don't take as a challenge! if not possible, not possible and that's all
00:31 < dooglus> still can't believe I had single-thread on all that time!  silly me.
00:31 < genete> sleep well 
00:31 < genete> bed time for me too.
00:31 < dooglus> I found a funny picture
00:31 < dooglus> just a second
00:31 < genete> ok
00:32 < dooglus> http://dooglus.rincevent.net/random/RubikChamp.jpg
00:32 < dooglus> me, as a kid
00:32 < dooglus> :)
00:32 < genete> wow!
00:32 < genete> it is you!!!!1
00:33 < timonator> you had ridiculous glasses :)
00:33 < dooglus> yup :)
00:33 < genete> "rubick cube champioship..."
00:33 < dooglus> free "national health service" glasses
00:33 < genete> it is nice recognize the face from people when was young
00:34 < genete> more hair now... 
00:35 < dooglus> oh yeah!
00:35 < genete> rest well dooglus, you deserve it!
00:35 < genete> night all!
00:36 < pxegeek> 'night genete
00:36 < genete> :)
00:36  * genete works tomorrow after 15 holidays ... 
00:37 < genete> :[
00:37 < genete> bye!
00:37 -!- genete [n=Genete@84.122.44.154.dyn.user.ono.com] has quit ["Abandonando"]
00:37 -!- TMM [n=hp@ip565b35da.direct-adsl.nl] has quit ["Ex-Chat"]
00:38  * timonator should prepare a presentation
00:38 < timonator> gotta get out of bed in ~6 hours :/
00:41 < pxegeek> go!
00:41 < pabs3> morning all
00:41 < pxegeek> Hi ya pabs!
00:41 < pabs3> nite timonator, thanks for the patch :)
00:41 < pabs3> hey pxegeek 
00:41  * pxegeek wonders whose bright idea it was to do a demoreel anyway....
00:42 < pabs3> how is that going btw?
00:42 < pxegeek> mutter, mutter, mutter
00:42 < pxegeek> rendering everything separately and using a video editor to splice it together. 
00:43 < pxegeek> rendering takes ages.
00:43 < pxegeek> got most of the examples done - now working on the contirbuted works
00:44 < pabs3> cool
00:47 < CIA-37> synfig: pabs * r1350 /synfig-studio/trunk/AUTHORS: Add Timo Paulssen (timonator) to the contributors.
00:47 < timonator> woah!
00:47 < timonator> that's just insace
00:47 < timonator> insane
00:48 < pabs3> pxegeek is in there, his patch fixed a typo :)
00:48 < pabs3> it is good to credit everyone !
00:48 < timonator> i can has make more patches tomorrow
00:48 < timonator> maybe
00:48 < pabs3> cool
00:48 < timonator> i hope
00:50 < pabs3> ooh, there is an svg2sif xslt converter in the forums!
00:50 < pabs3> http://synfig.org/forums/viewtopic.php?t=30
00:52 < pabs3> dooglus: would you mind adding a wiki page about the sif format and how it has evolved with the changes you made?
00:54 < pxegeek> Oh oh - there goes the demoreel....
00:54 < pxegeek> ;)
00:54  * pxegeek gets distracted by xlst files
00:54 < pabs3> it would also be interesting to write a script that automatically converts the source code into a description of the file format
00:55 < pxegeek> sounds like a job for doxygen?
00:57 < timonator> i don't think doxygen could handle that
00:57 < pabs3> hmm. possibly, but I think it only operates on comments
00:58 < MangoFusion> neat converter
00:58 < MangoFusion> always thought one could make a swf converter (though obviously not using XSLT)
00:59 < MangoFusion> will be interesting to see what happens in the future i guess
00:59 < MangoFusion> :)
01:01 < pxegeek> interesting - he's trying to map to radial composite, rather than just composite 
01:02 < pxegeek> I wonder if he realizes synfig can just take x& y for tangent arguments instead of radius and theta...
01:05 < pabs3> is that in the .sif?
01:17 < CIA-37> synfig: dooglus * r1351 /synfig-studio/trunk/src/synfigapp/actions/valuedescconnect.cpp: Improve the History dialog's "Connect" entry.
01:20 < pxegeek> just looking at the code in the xlst file.  
01:22 < pxegeek> I can't get  it working though - I've been using msxml rather than saxon.  And my test svg is not inkscape generated
01:22 < pxegeek> but apart from that it looks promising!
01:27 < pabs3> doesn't seem to work with xsltproc because it uses XSLT version 2
01:32 < CIA-37> synfig: dooglus * r1352 /synfig-studio/trunk/src/synfigapp/actions/valuenoderemove.cpp: Mention the exported name of the ValueNode in the History dialog when unexporting it.
01:32 < CIA-37> synfig: dooglus * r1353 /synfig-studio/trunk/src/synfigapp/actions/valuenoderename.cpp: Show the old and new names of the ValueNode in the History dialog when renaming it.
01:47 < pxegeek> Well I got it to cough up a valid sif file
01:47 < pxegeek> unfortunately the layers that come with it are blank.
02:02 -!- MangoFusion [n=jamesu@host81-156-80-146.range81-156.btcentralplus.com] has quit ["ttg"]
02:10 -!- AkhIL [n=AkhIL@90.188.192.229] has quit [Read error: 110 (Connection timed out)]
02:40 < htodd> whee
02:41 < htodd> I feel like I'm coming down with something so now I'm going to see if I can update synfig in fink to teh svn version
02:49 < pxegeek> There's a lot of it about...
02:50 < pxegeek> All of my family have had stomach flu followed by a nasty cough.
02:50 < pxegeek> My daughter just relapsed....
02:50 -!- jdc18 [n=jdc@200.107.46.12] has left #synfig []
02:51 < timonator> aaw, that's really bad
02:51 < pxegeek> (especially when you're the one holding the garbage can because it was the only thing within reach...)
02:52 < pxegeek> Yes, not much fun all around.  She's gone to bed early.  Hopefully she'll feel better when she wakes up.
02:56 < htodd> that's too bad
02:56 < htodd> there's someone at work that's coughing up a lung lately
02:57 < htodd> I'm not about to go over to introduce myself
03:00 < pxegeek> :)
03:08 -!- ZanQdo [n=Daniel@201.201.2.22] has joined #synfig
03:30 < htodd> hmm, 1353 synfigstudio not building for me
03:38 -!- ZanQdo [n=Daniel@201.201.2.22] has quit [Read error: 104 (Connection reset by peer)]
03:38 -!- ZanQdo [n=Daniel@201.201.2.22] has joined #synfig
03:55 -!- ZanQdo [n=Daniel@201.201.2.22] has quit ["Adios"]
05:00 -!- AkhIL [n=AkhIL@90.188.192.229] has joined #synfig
05:09 < pxegeek> and now my son just threw up.  It's going to be a long night.
05:25 < pxegeek> Ha!
05:25 < pxegeek> http://synfig.org/forums/viewtopic.php?p=108#108
05:26 < pxegeek> Looks like a good point to call it a night.  
05:48 < htodd> :(
06:31 -!- ZanQdo [n=Daniel@201.201.2.22] has joined #synfig
07:26 -!- pxegeek [n=chatzill@c-71-59-140-184.hsd1.or.comcast.net] has quit [Read error: 113 (No route to host)]
08:02 -!- ZanQdo [n=Daniel@201.201.2.22] has quit ["Adios"]
08:07 < dooglus> pabs3: re. .sif and "how it has evolved with the changes you made" - it hasn't, really.  I've added a bunch of new valuenodes, and alternative names for layers, but maintained backwards compatibility.  so generating a .sif that loads in old versions should be fine for new versions too.
08:09 < pabs3> yeah, thats what I mean - new parameters, new layers and so on
08:11 < dooglus> pabs3: there's nothing new in the .sif format, really.  the new valuenodes are documented with the others, the new layers too.
08:23 < dooglus> pabs3: I may be misunderstanding you.  what is the purpose of having another wiki page listing the valuenodes, layers, and parameters?
08:24 < pabs3> it doesn't have to be another page, just some information about SVN revisions/version numbers where parameters were added, layers were added and other changes were made
08:25 < pabs3> the purpose would be to have a record of when those changes occurred
08:32 < dooglus> ok
08:33 < pabs3> and also to have a record of which changes occurred
08:33 -!- Zelgadis [n=zelgadis@87.103.170.86] has joined #synfig
08:39 < dooglus> like this? http://synfig.org/Convert#Compatibility
08:40 < pabs3> cool, thanks
08:45 < dooglus> updated to include the revision number at release of 0.61.[67] so it's easy to see which valuenodes were in which release
08:55 < Zelgadis> pabs3: what do you yhink about splitting "Dialogs" to "Dialogs" and "Panels"?
08:58 < pabs3> I don't have an opinion either way really
08:59 < Zelgadis> pabs3: I think it's strange to name "New File Dialog" and "Layers Dialog" as the same class -- "Dialog"
09:00 < Zelgadis> They're completely different things
09:02 < pabs3> ok, sounds reasonable. I don't feel strongly enough about it to do the page renames and editing myself though
09:03 < Zelgadis> I feel strongly ;)
09:04 < Zelgadis> Then it's ok?
09:04 < pabs3> I think so, dunno about everyone else
09:05 < pabs3> can't see why anyone might object though
09:05 < pabs3> just be sure to update all the links to the renamed pages
09:06 < Zelgadis> OK, i'll wait some time and then do the work
09:13 < dooglus> Zelgadis: it's a good idea
09:14 < dooglus> there are dockable dialogs, and non-dockable
09:14 < dooglus> the dockable ones stay on screen, whereas the non-dockable are shown as needed
09:14  * Zelgadis glad
09:15 < dooglus> Zelgadis: I don't know what term you would use to distinguish the two classes though
09:15 < Zelgadis> Panels
09:22 < Zelgadis> dooglus: dockable dialogs = "Panels", non-dockable dialogs = "Dialogs"
09:50 -!- MangoFusion [n=jamesu@host81-156-80-146.range81-156.btcentralplus.com] has joined #synfig
10:02 < CIA-37> synfig: dooglus * r1354 /synfig-core/trunk/src/synfig/ (valuenode_subtract.cpp valuenode_add.cpp): Extend the Add and Subtract ValueNodes to work with gradients.
10:05 -!- Zelgadis [n=zelgadis@87.103.170.86] has quit ["Leaving."]
10:24 -!- xerakko [n=xerakko@lpri.gva.es] has joined #synfig
10:31 -!- MangoFusion [n=jamesu@host81-156-80-146.range81-156.btcentralplus.com] has quit []
10:47 < CIA-37> synfig: dooglus * r1355 /synfig-core/trunk/src/synfig/valuenode.cpp: Check that a ValueNode is compatible with the given type before attempting to create it.
10:47 < CIA-37> synfig: dooglus * r1356 /synfig-core/trunk/src/synfig/loadcanvas.cpp: 
10:47 < CIA-37> synfig: Refer the user to 'http://synfig.org/Convert#Compatibility' if we try to load a
10:47 < CIA-37> synfig: .sif file that uses a not-yet-implemented ValueNode type, or that tries using a
10:47 < CIA-37> synfig: ValueNode with a type which it doesn't yet support. Both of these are usually a
10:47 < CIA-37> synfig: sign that the .sif file was created with a newer version of synfig than the user
10:47 < CIA-37> synfig: has.
11:18 < CIA-37> synfig: dooglus * r1357 /synfig-core/trunk/src/synfig/valuenode.cpp: Revision 1355 broke the Duplicate layer. Oops!
11:32 < CIA-37> synfig: dooglus * r1358 /synfig-core/trunk/src/synfig/ (layer_duplicate.cpp layer_duplicate.h): Use a mutex to prevent the two render threads (rendering the workarea and the navigator dialog thumbnail) trampling on each other while rendering the Duplicate layer.
11:55 -!- xerakko [n=xerakko@lpri.gva.es] has quit [Remote closed the connection]
11:55 -!- xerakko [n=xerakko@lpri.gva.es] has joined #synfig
12:14 -!- MangoFusion [n=jamesu@host81-156-80-146.range81-156.btcentralplus.com] has joined #synfig
12:52 -!- pabs3 [n=pabs@121.220.225.161] has quit [Remote closed the connection]
12:56 -!- pabs3 [n=pabs@121.220.225.161] has joined #synfig
13:16 -!- AkhIL [n=AkhIL@90.188.192.229] has quit [Read error: 104 (Connection reset by peer)]
13:17 -!- AkhIL [n=AkhIL@90.188.192.229] has joined #synfig
14:11 -!- xerakko [n=xerakko@lpri.gva.es] has quit ["ChatZilla 0.9.79 [Firefox 2.0.0.11/2007120410]"]
14:54 -!- Zelgadis [n=zelgadis@87.103.170.86] has joined #synfig
15:19 -!- Netsplit calvino.freenode.net <-> irc.freenode.net quits: dyloxin, omry
15:20 -!- Netsplit over, joins: omry, dyloxin
15:35 -!- TMM_ [n=hp@ip565b35da.direct-adsl.nl] has joined #synfig
15:38 -!- TMM_ is now known as TMM
15:56 -!- xerakko [n=Miguel@debian/developer/xerakko] has joined #synfig
16:15 -!- genete [n=Genete@84.122.44.154.dyn.user.ono.com] has joined #synfig
16:21 < genete> dooglus: if you continue committing at that rate you'll reach revision 2000 on March 20th!!! Wow!
16:22 < genete> a 10 commits per day rate!!!
16:35 < dooglus> genete: hi :)
16:35 < genete> hi dooglus! ;)
16:35 < dooglus> genete: I just saw the bug that pxegeek had a long time ago
16:35 < dooglus> where the tiles are 'staggered' diagonally across the workarea
16:35 < genete> which?
16:35 < genete> ah
16:36 < genete> I saw it only one time
16:36 < genete> repeteable?
16:38  * genete testing the workarea and main window "fighting" for Dup index recently fixed...
16:39 < dooglus> genete: quite repeatable
16:39 < genete> got a sifz file?
16:39 < dooglus> genete: it happens when you zoom such that the canvas width is a multiple of half the tile width, but not of the whole tile width
16:39 < dooglus> genete: ie. to a multiple of 64 but not 128
16:40 < dooglus> genete: to make it repeatable, I made the tiles by 2x2 instead of 128x128
16:40 < dooglus> then it happens 50% of the time
16:40 < genete> ah
16:40 < dooglus> rather than <1% of the time
16:40 < genete> so for a normal user like me, qhat's the zoom value to have that strange tiles?
16:41 < dooglus> just drag the zoom slider in the nav. dialog very slowly to the right
16:41 < dooglus> it has to be zoomed enough to be using tile rendering
16:41 < dooglus> and you have a 1 in 128 chance of it going wrong :)
16:41 < dooglus> (at each zoom)
16:41 < dooglus> here's an example: http://dooglus.rincevent.net/synfig/skew.png
16:42 < dooglus> and another: http://dooglus.rincevent.net/synfig/skew2.png
16:42 < genete> lol
16:43 < genete> so it depends also on how wide is your navigation window
16:44 < dooglus> ah, maybe
16:44 < dooglus> yes, it does :)
16:44 -!- Zelgadis [n=zelgadis@87.103.170.86] has left #synfig []
16:44 < genete> because when I slide the zoom slider it changes its values depending on the navigator total width
16:44 < dooglus> right...  you can move the slider pixel-by-pixel
16:45 < dooglus> and the amount it changes depends on the length of the slider :)
16:45 < dooglus> I didn't find a fix yet
16:45 < dooglus> turning off 'view>low res' always fixes it
16:45 < dooglus> (low res halves the resolution and is the source of the bug)
16:46 < dooglus> workarea.cpp is full of delightful code like this:
16:46 < dooglus> const int width_in_tiles(w/tile_w+(w%tile_w?1:0));
16:46 < dooglus> ie. add on 1 to the number of tiles per row if the width isn't divisible by the tile width
16:46 < MangoFusion> haha
16:47 < dooglus> in some parts of the code, this is checked before halving everything, in other places it's done after
16:47 < MangoFusion> when one starts adding in workaround code in the wrong places, things tend to go wrong very badly
16:47 < dooglus> so we get different ideas of whether the tiles exactly fit the width
16:48 < dooglus> that conditional 'add another tile' thing happens in at least 5 places that I've seen
16:48 < dooglus> like: if(rend_desc().get_w()%get_tile_w()!=0)tw++;
16:48 < dooglus> also in workarea.cpp
16:48 < dooglus> that exact text appears 4 times alone
16:49 < genete> BTW can be a check box added to the Toogles entries in View menu? (like the ones for the ducks visibility) I would really help...
16:49 < genete> s/I/It
16:50 < genete> it seems you have fixed the last bug from the Dup layer! ("index fighting")
16:54 < dooglus> genete: try it.  is it ok?
16:54 < genete> yeah 
16:54 < dooglus> genete: I don't know if it causes any new problems, but I hope not
16:55 < genete> I cannot reproduce it now
16:55 < dooglus> genete: there's another outstanding issue with the dup layer - just one now
16:55 < dooglus> the 'bounding box' thing
16:55 < genete> what's that?
16:55 < dooglus> that's 'just' an optimization though
16:55 < dooglus> to stop synfig recalculating the dup'ed layers for each tile if they're not IN the tile
16:56 < genete> does the motion blur do the same?
16:56 < dooglus> I didn't check
16:56 < dooglus> I did mean to
16:56 < dooglus> but forgot
16:56 < dooglus> came up against this 'stagger' bug instead
16:57 < genete> how to calculate if a layer is to be renderd in the current tile?
16:57 < dooglus> most layers know how big their effect is
16:57 < genete> even the blurred and the feathered ones?
16:58 < dooglus> a circle knows it doesn't affect anything outside a box of size 2r, centred on its centre, unless it's feathered or inverted, for example
16:58 < dooglus> the translate layer knows it moves the effects, and so on
16:58 < dooglus> so synfig can avoid rendering stuff that won't be seen
16:59 < genete> for example an exported canvas that is not linked to any canvas pointer?
16:59 < dooglus> if it's not linked, nothing in it will be rendered
16:59 < dooglus> the rendering is recursive, remember?
16:59 < dooglus> so there's nothing to call any of its layers' render function
17:00 < genete> ok
17:00 < genete> the speed improvements for dup layer can be used for motion blur too if successfully work
17:01 < dooglus> right
17:02 < genete> dooglus: The current particle design is:
17:02 < genete> Dup/Translate/Time Loop/Particle layer
17:02 < genete> The old one was:
17:03 < genete> Several Inline Canvases where the Origin has the magic conversions 
17:03 < genete> and the timeoffset the needed random offset for each particle
17:04 < genete> now with the improvement of dup layer affecting to paste canvases I wonder if it would be good to make a new particles version using dup but with old style.
17:04 < dooglus> I guess you can avoid timeloop and translate - use a pastecanvas instead?
17:04 < dooglus> I guess the less layers the better
17:04 < dooglus> in terms of speed
17:04 < genete> translate+Time loop would be slower than sme things for a paste canvas?
17:05 < dooglus> then again, maybe the overhead of having an extra inline canvas makes it slower.  I don't know
17:06 < genete> I think the time loop layer would be slower than the off set parameter for the paste canvas layer
17:06 < genete> time offset I mean
17:07 < genete> becasue I'm using the time loop as a Time Offset 
17:07 < genete> I think that for the moment I'll let it as it is by now
17:07 < genete> and continue refining the behavior
17:08 < genete> and finnaly adopt the best setup for speed improvement
17:08 < genete> I'll work a little witht some examples for the wiki.
17:18 < genete> dooglus: what's the difference between click on JMP at the history dialog and uncheck the history entry?
17:19 < genete> Ah I see
17:19 < genete> JMP is like undo or redo
17:19 < genete> check  or uncheck enable/disable that entry in the history so it cnanot be redone or undone if unchecked.
17:19 < genete> cool
17:20 < genete> amazimg improvements on history dialog...!
17:27 < dooglus> genete: JMP takes you to that point
17:27 < dooglus> undo-ing or redo-ing what it needs to, to get you there
17:27 < dooglus> unchecking an entry undoes *just that one thing*
17:28 < genete> you cnanot re-do an action if unchecked
17:28 < dooglus> ok
17:28 < genete> I think uncheck is like disable for redo or undo stack
17:29 < dooglus> what actions are the most important ones still missing proper variable descriptions now?
17:29 < dooglus> right.  it's like for layers
17:31 < genete> I don't have a long and profuse history log at hand to reply your question.
17:31 < dooglus> ok
17:31 < genete> let me look at it more continuously meanwhile working and let you know,
17:32 < genete> I have also a lot of translating work pending...
17:32 -!- TMM [n=hp@ip565b35da.direct-adsl.nl] has quit ["Ex-Chat"]
17:35 < dooglus> sure, ok
17:35 < dooglus> it seems more useful for most of the operations I do now
17:35 < dooglus> but I guess you do different things
17:36 < genete> oh I was distracted... what did you mean?
17:47 < dooglus> I mean it seems to be OK if you convert, export, connect
17:47 < dooglus> that's what I mostly have been playing with
17:47 < genete> ah ok
18:26 -!- genete [n=Genete@84.122.44.154.dyn.user.ono.com] has quit ["Abandonando"]
19:02 < CIA-37> synfig: dooglus * r1359 /synfig-core/trunk/src/synfig/target_tile.h: Define the TILE size here, instead of having it in several locations.
19:02 < CIA-37> synfig: dooglus * r1360 /synfig-core/trunk/src/synfig/target_tile.cpp: Use the TARGET_SIZE defined in target_tile.h.
19:02 < CIA-37> synfig: dooglus * r1361 /synfig-studio/trunk/src/gtkmm/workarea.cpp: Use the TILE_SIZE defined in target_tile.h.
19:02 < CIA-37> synfig: dooglus * r1362 /synfig-studio/trunk/src/gtkmm/ (workarea.cpp renderer_canvas.cpp): 
19:02 < CIA-37> synfig: Fix 1824437: studio occasionally renders the tiles in the wrong places. This
19:02 < CIA-37> synfig: was happening when the canvas was one pixel greater than an integer number of
19:02 < CIA-37> synfig: tiles wide, and low-resolution rendering (the default) was enabled.
19:32 < CIA-37> synfig: dooglus * r1363 /synfig-studio/trunk/src/synfigapp/actions/layerraise.cpp: Get the 'Raise Layer' action to list the layer(s) that were raised in the History dialog.
19:32 < CIA-37> synfig: dooglus * r1364 /synfig-studio/trunk/src/synfigapp/actions/layerlower.cpp: Get the 'Lower Layer' action to list the layer(s) that were lowered in the History dialog.
20:48 -!- MangoFusion [n=jamesu@host81-156-80-146.range81-156.btcentralplus.com] has quit []
21:02 -!- MangoFusion [n=jamesu@host81-156-80-146.range81-156.btcentralplus.com] has joined #synfig
21:03 -!- xerakko [n=Miguel@debian/developer/xerakko] has quit [Read error: 104 (Connection reset by peer)]
21:14 -!- TMM [n=hp@ip565b35da.direct-adsl.nl] has joined #synfig
22:19 -!- genete [n=Genete@84.122.32.227.dyn.user.ono.com] has joined #synfig
22:36 < genete> hi
22:37 < genete> dooglus: converting a seed of a random type seems to loose the ability of generate new seeds when manual duplication
22:37 < genete> I  mean if the seed is converted or linked then the manual dupliucation randomize trick doesn't work, right?
22:42 < dooglus> hi
22:42 < dooglus> that's right
22:42 < genete> thanks for the confirmation.
22:42 < dooglus> if you've converted it, you've defined it - it's not random any more
22:43 < genete> I then need to convert again inside seed to a random to let it be different when manual duplicate
22:43 < dooglus> specifically:
22:43 < dooglus> modules/mod_noise/valuenode_random.cpp:
22:43 < dooglus> if(!link->is_exported() && link->get_name() == "constant") ... randomise_seed()
22:44 < genete> I want to have more than one particle layer with the same general parameters. Manual duplciation the particle stuff under the Dup file did not generate another random particle to be duplicate.
22:45 < genete> so I've converted the internal part of the seed to a random again  (not animated)
22:45 < genete> then it would produce fresh seeds when duplicate.
22:45 < genete> manually
22:46 < genete> later dup layer would produce new seeds for each particle layer for each duplication.
22:46 < dooglus> you converted the seed so it would be set by the duplicate layer?
22:46 < dooglus> right
22:46 < genete> yes
22:46 < dooglus> so the seed isn't random - it's defined by the 'index'
22:46 < genete> seed is a add type converted
22:46 < genete> yes
22:47 < dooglus> so you can use a new random as the RHS of the seed?
22:47 < genete> but I want to just duplicate the stuff below the Dup layer to have otre particle with other shape in the same particle system
22:47 < genete> yeah!
22:47 < genete> that's what I've just done.
22:47 < genete> it works
22:48 < dooglus> in the history dialog,
22:48 < dooglus> how do you want it shown if you edit that seed manually?
22:48 < genete> but, only the seed or any parameter?
22:48 < dooglus> it's quite deeply nested I guess
22:48 < dooglus> any deeply nested param
22:48 < dooglus> is it helpful to see the path in full?
22:49 < genete> Mmm no
22:49 < dooglus> position>composite:x>random:seed>23
22:49 < dooglus> ?
22:49 < dooglus> currently I think it just says "particle -> seed" but you don't know which seed
22:50 < genete> layer3>Origin>LHS>LHS>Seed>LHS>... ouch!
22:50 < genete> I think that layer is mandatory
22:50 < dooglus> you're just naming the params - you could name the convert Types too
22:50 < dooglus> sure.
22:50 < genete> layername I mean
22:50 < genete> then first param name
22:51 < dooglus> layer3>origin:Composite>Add:LHS>Add:LHS>Random:Seed>Add:LHS
22:52 < genete> itwoudl be useful for someone but really do you think that too much detail helps?
22:52 < dooglus> I don't know what's useful
22:52 < dooglus> does searching work?
22:52 < dooglus> if so, it could be more useful
22:53 < dooglus> (I know I modified something to do with a 'gradient rotate', but the final sub-param was a 'random seed'
22:53 < dooglus> I search for 'rotate' and find it in the History dialog
22:53 < genete> how do you serach in the history?
22:53 < dooglus> I guess you just type, like in any other dialog
22:54 < dooglus> maybe its not enabled, I don't kno
22:54 < dooglus> w
22:54 < dooglus> but it shouldn't be hard to do
22:54 < genete> it is enabled but you only can search by the 
22:54 < genete> beggining of the string
22:54 < genete> For instance:
22:55 < genete> Export "Inline Canvas"=>Local TIme as 't*'
22:55 < genete> and I type "t*" and is not found
22:55 < dooglus> right
22:55 < dooglus> so i see
22:55 < genete> I acn search for the beggining of the string
22:56 < dooglus> you can search the whole string if you're prepared to type it exactly...
22:56 < genete> yes :(
22:56 < dooglus> I guess that's easy to fix
22:56 < dooglus> see how it works in the layer dialog?
22:56 < dooglus> you can type anything from a layer name
22:56 < dooglus> not just the beginning
22:57 < genete> oh yeah
22:57 < genete> it is fine!
22:57 < genete> one more thing to the wiki!
22:57 < dooglus> the up and down arrows jump between matches
22:57 < genete> So then it could be useful to have a full path for the params modifications.
22:58 < genete> I think only 5% of users would design a complex convert type like particles...
22:58 < genete> that would produce that long string
22:58 < dooglus> even if they do, the long string just goes off the edge of the window
22:58 < dooglus> it doesn't really matter
22:58 < genete> yes
22:58 < genete> go then!
22:58 < dooglus> maybe have a checkbox in the dialog to select 'short actions'
22:59 < dooglus> did I sent the 'raise' and 'lower layer' updates?
22:59 < genete> yes you did
22:59 < dooglus> o
22:59 < dooglus> k
22:59 < genete> but I did not update yet ;)
23:00 < dooglus> I didn't fix any of the activepoint or waypoint actions at all
23:00 < dooglus> mostly because I don't use either of them
23:01 < genete> dooglus: one nasty thing I've noticed is that editing a Bline below a dup layer becomes very difficult
23:02 < genete> modify the ducks positions is really difficult because everytime you do it the Bline ducks runs all the dups of itself and the ducks go here and there if the dups are a t different positions (like in particles)
23:02 < CIA-37> synfig: dooglus * r1365 /synfig-core/trunk/src/synfig/valuenode_bline.cpp: Revert the incorrect change to this file made in r1326.
23:02 < genete> I have to disable rhe dup layer to edit the Bline properly.
23:03 < dooglus> genete: want to add it to the wiki 'duplicate layer' list of problems?
23:03 < dooglus> genete: which dialogs have search build in?
23:03 < dooglus> built
23:03 < genete> yes I'll do
23:03 < dooglus> the source files with it seem to be: keyframe children layergroup layer
23:04 < genete> Child dialog have it
23:04 < genete> meta data also 
23:05 < dooglus> keyframes have it, but only at start of string
23:06 < genete> Params dialog the same
23:07 < genete> (only the beggining)
23:07 < dooglus> what does it search for?
23:07 < dooglus> (in params)
23:07 < genete> the param name
23:07 < genete> "Tiempo de Enlace"
23:07 < dooglus> oh?
23:07 < dooglus> not for me
23:07 < genete> I type "Tiem"
23:07 < genete> and select the first
23:07 < genete> I think it is a gnome thing
23:08 < dooglus> maybe only works for translations?
23:08 < dooglus> someone has to tell GNOME which column to search in, surely?
23:08 < genete> no 
23:08 < genete> It search it if it is already expanded
23:08 < dooglus> can you search for types, or values?
23:08 < genete> param names
23:08 < genete> initial lettters
23:09 < genete> if it is contracted nothing found
23:09 < dooglus> aah, it only searches for sub-parameters
23:09 < dooglus> it won't search for layer parameters - only valuenode parameters
23:09 < dooglus> I didn't have any converts in my Circle layer
23:09 < dooglus> just regular Circle params
23:09 < genete> he he
23:10 < genete> I can borrow you some converts
23:10 < genete> ;)
23:10 < dooglus> I'm sure
23:10 < dooglus> you have pages of them :)
23:14 < genete> updated wiki for the edition of blines under dups
23:16 < dooglus> thanks
23:18 < genete> dooglus: I've come back to the original setup for particles (inline canvas resposible of position and time offset) instead of the TRanslate and Time Loop layers.
23:26 < dooglus> do you see any difference?
23:35 < genete> I've not done any benchmark yet but it is more easy to manage
23:50 < genete> now I can mix different types of particles wit hthe same parameters
23:59 < dooglus> I just got the history dialog search working on substrings
--- Log closed Tue Jan 15 00:00:30 2008