--- Log opened Tue Jan 22 00:00:30 2008
00:04 < genete> I see problems with Straight on Text, Star, Region, Rectangle, Polygon, Outline and circle
00:06 < genete> I see problems with Straight Onto on Text, Star, Region, Rectangle, Polygon, Outline and circle too
00:07 < genete> Rectangle Striaght Onto is very similar encapsulated and not but not equal. See reanctgle method 14 too
00:07 < genete> Threre is a thin line different
00:08 -!- pabspabspabs [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
00:24 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Read error: 110 (Connection timed out)]
00:25 -!- pabspabspabs is now known as pabs3
00:33 < genete> dooglus: from some time ago this link doesn't work: http://kibi.dyndns.org:8083/~dooglus/gitweb.pl?p=synfig;a=summary
00:35 < KiBi> genete: arg.
00:35 < KiBi> genete: I'll power it up again as soon as I'm able to.
00:36 < genete> KiBi: that interface is interesting to see what does dooglus changed in each revision
00:39 < KiBi> genete: I know why it's used for ;-)
00:39 < KiBi> genete: I've to figure out whether it's possible to start a virtual machine remotely with virtualbox, and I'm not sure it's possible.
00:41 < KiBi> *what
00:45 -!- genete [n=Genete@84.122.52.3.dyn.user.ono.com] has left #synfig ["Abandonando"]
00:45 -!- genete [n=Genete@84.122.52.3.dyn.user.ono.com] has joined #synfig
00:47 -!- Yoyobuae [n=quete@201.224.135.156] has joined #synfig
00:47 < Yoyobuae> hi
00:48 < KiBi> genete: http://kibi.dyndns.org:8083/~dooglus/gitweb.pl?p=synfig is back.
00:48 < genete> KiBi: thanks!
00:48 < genete> hi Yoyobuae 
00:48 < KiBi> Sorry for the downtime.
00:48 < genete> NP
00:48 < KiBi> genete: Don't hesitate to poke me here if it goes down again.
00:48 < genete> sure :)
00:48 < KiBi> Looks like the machine got stuck at some point, dunno why.
00:49 < genete> maybe it needs some periodically care...
00:50 < genete> \_/?<
00:50 < KiBi> I thought of migrating it to the host machine anyway, but I had some troubles with... I don't remember.
00:50 < Yoyobuae> yay, duck!!
00:50 < KiBi> \_o<
00:50 < KiBi> >o
00:50 < KiBi> o<
00:50 < KiBi> And ducklings!
00:50 < pabs3> :)
00:53 < Yoyobuae> reading the logs it seems i opened a pandora's box with the Straight layer XD
00:53 < Yoyobuae> s/layer/blend method/
00:54 < genete> it was there from some time ago but not seriously taken...
00:57 < Yoyobuae> btw, how much time does synfig take to build?
00:57 < Yoyobuae> from scratch that is
00:58 < genete> it depends on your machine. Here 20-30 mins for core and 30-45 for studio I think
00:58 < genete> core less time I believe.
00:58 < genete> cd ..
00:59 < genete> lol
00:59 < Yoyobuae> wrong prompt?
00:59 < genete> yagh
00:59 < genete> just finished studio build ;)
01:00 < Yoyobuae> im half-way thru the "Building on Windows" instructions
01:01 < genete> dooglus: I continue seeing the Straight blend errors after update to 1414. Is that correct?
01:04 < genete> And it seems you have broken Striaght Onto over some layers that worked well previously.
01:06 < Yoyobuae> i haven't seen any commit that fixes Straight yet
01:06 < Yoyobuae> the only one i have seen related to Straight is:
01:06 < Yoyobuae> 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.
01:06 < genete> oh sorry Straight onto still fine
01:07 < Yoyobuae> nvm then =)
01:07 < dooglus> Yoyobuae: you've been keeping me busy
01:09 < Yoyobuae> yeah, been reading logs =)
01:09 < dooglus> genete: I didn't commit my changes yet.  what's broken newly?
01:09 < genete> nothing yet
01:09 < genete> :)
01:09 < dooglus> ok
01:10 < dooglus> Yoyobuae: I have a bunch of local changes that I'll commit as one, once I'm happy with them
01:10 < Yoyobuae> good =)
01:11 < Yoyobuae> btw, which OpenEXR version should I download? openexr-1.4.0a or the latest?
01:11 < dooglus> or maybe I'll just commit them and let genete test :)
01:11 < dooglus> I don't know, sorry.
01:11 < genete> dooglus: defenitively Striagth stills the same...
01:11 < Yoyobuae> which do you use? or you dont need it?
01:11 < dooglus> I don't use it, no
01:11 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has quit ["Not here"]
01:11 < Yoyobuae> oh i c
01:11 < dooglus> genete: that's a good thing
01:12 < dooglus> http://devrnd.net/img/MSN.png
01:12 < genete> Straight I mean
01:13 < genete> I see more "love" icons than others...
01:14 -!- Supersayoyin [n=quete@201.224.135.156] has joined #synfig
01:14 -!- Yoyobuae [n=quete@201.224.135.156] has quit [Read error: 104 (Connection reset by peer)]
01:14 -!- Supersayoyin [n=quete@201.224.135.156] has quit [Client Quit]
01:14 -!- Yoyobuae [n=quete@201.224.135.156] has joined #synfig
01:15 < genete> well, bed time here... dooglus good job :)
01:15 -!- genete [n=Genete@84.122.52.3.dyn.user.ono.com] has quit ["Abandonando"]
01:15 < pabs3> Yoyobuae: I'd use the one that your distro has (least effort), unless yr on windows or something
01:16 < Yoyobuae> Windows here =)
01:16 < Yoyobuae> ill download the one listed on the wiki with the build instructions
01:17 < pabs3> I'd use the latest, we need people to test with all different versions
01:17 < Yoyobuae> ill test the other one later then =)
01:24 -!- Yoyobuae [n=quete@201.224.135.156] has quit [Read error: 104 (Connection reset by peer)]
02:04 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Read error: 104 (Connection reset by peer)]
02:07 -!- Yoyobuae [n=quete@201.224.135.156] has joined #synfig
02:08 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
02:10 < Yoyobuae> i get an error when i try to build ETL: "autoreconf: `configure.ac' or `configure.in' is required"
02:16 -!- Yoyobuae [n=quete@201.224.135.156] has quit [Read error: 104 (Connection reset by peer)]
02:18 -!- pabspabspabs [i=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has joined #synfig
02:18 -!- pabs3 [n=pabs@d122-104-115-97.per9.wa.optusnet.com.au] has quit [Nick collision from services.]
02:18 -!- pabspabspabs is now known as pabs3
02:18 -!- Yoyobuae [n=quete@201.224.135.156] has joined #synfig
02:19 < Yoyobuae> if anyone reads just answer, ill catch response on logs
02:19 < Yoyobuae>  im building on Windows XP, following the http://www.synfig.org/Windows_build_instructions#_src_directory instructions
02:20 < pabs3> Yoyobuae: sounds like you have an incomplete checkout, configure.ac should be present in the checkout
02:25 -!- Yoyobuae [n=quete@201.224.135.156] has quit [Read error: 104 (Connection reset by peer)]
02:27 -!- Yoyobuae [n=quete@201.224.135.156] has joined #synfig
02:27 < Yoyobuae> i keep disconnecting for no reason, =/
02:28 < Yoyobuae> anyway im trying to find configure.ac manually first
02:29 < pabs3> KiBi: git can emulate a CVS server, have you heard of it being able to emulate an SVN server?
02:29  * pabs3 thinking about the eventual synfig git switch
02:30 < Yoyobuae> ohhh, found mistake while doing: $ svn co http://svn.voria.com/code/ETL/trunk/ ETL
02:31 < KiBi> pabs3: Nope, but I didn't search this way either.
02:31 < Yoyobuae> i incorrectly typed the last "ETL" without any space before it
02:31 < KiBi> pabs3: I think the git-svn maintainer is saying that it won't be added that much features, since it's mostly a hack and it for smoothing transition purpose only.
02:31  * pabs3 will investigate
02:32 < pabs3> Yoyobuae: that sounds very likely to be the problem
02:32 < Yoyobuae> i think maybe the wiki needs a note about those spaces
02:32 < Yoyobuae> on: http://www.synfig.org/Windows_build_instructions#_src_directory
02:33 < Yoyobuae> i really didnt see it the first time thru XD
02:33 < pabs3> KiBi: ok. I'm thinking about all the windows folks, since git needs a unixy environment 
02:34 < pabs3> Yoyobuae: feel free to add a note to that page, it is editable by all :)
02:34 < KiBi> pabs3: Is that still true?
02:34 < Yoyobuae> ok =)
02:36 < pabs3> KiBi: $ for f in /usr/bin/git* ; do head -n1 $f | grep ^# ; done | sort | uniq -c
02:36 < pabs3>      32 #!/bin/sh
02:36 < pabs3>       1 #!/bin/sh -e
02:36 < pabs3>       3 #!/usr/bin/perl
02:36 < pabs3>       5 #!/usr/bin/perl -w
02:36 < pabs3> so yeah, still needs at least sh and perl
02:37 < KiBi> pabs3: I meant that it might be that some kind of GUI may have been crafted to support basic functionalities.
02:38 < KiBi> People are crazy, don't forget ;)
02:38 < pabs3> true, I should look into that too
02:45 < pabs3> aha http://code.google.com/p/msysgit/
02:47 < pabs3> pxegeek/Yoyobuae: if you feel like you have time and hair to spare, please try out msysgit
02:49 < KiBi> No screenshot?!
02:50 < KiBi> How not to participate
02:50 < KiBi> So you do not want to help? Then there is nothing to see here, please move along. 
02:50 < KiBi> aha
02:51 < Yoyobuae> ok, will try to install later
02:52 < Yoyobuae> ETL built, YAY.  thanks for help paps3 =)
02:55 -!- Yoyobuae [n=quete@201.224.135.156] has quit [Read error: 104 (Connection reset by peer)]
02:59 -!- pxegeek [n=chatzill@c-71-59-140-184.hsd1.or.comcast.net] has joined #synfig
03:10 < pabs3> KiBi: I get the feeling it is command-line
03:12 < KiBi> ah, full moon today.
03:12 < KiBi> But cloudy, suck.
03:12  * KiBi gets back to writing manpages.
03:12 < pabs3> looks like there is a tortisecvs style thing: http://repo.or.cz/w/git-cheetah.git/
03:13 < KiBi> Reminds me that I only know this tool by its name.
03:14 < pxegeek> beautifully clear here tonight....
03:39 -!- factor [n=factor@ip68-14-160-70.ok.ok.cox.net] has joined #synfig
03:44 -!- Yoyobuae [n=quete@201.224.135.156] has joined #synfig
03:45 < Yoyobuae> pxegeek: how much does synfig take to build using MingW?
03:57 -!- Yoyobuae [n=quete@201.224.135.156] has quit [Read error: 104 (Connection reset by peer)]
04:00 < pxegeek> too long
04:00 < pxegeek> about 40 mins for core, and another 20 ish for studio
04:01 < pxegeek> Just a little over an hour for everything (P4, 3GHz)
04:02 < pxegeek> The windows build process scripting isn't the most intelligent - it does things twice - once for the install package, and again so that studio has something to build the icons with.
04:02 -!- Yoyobuae [n=quete@201.224.135.156] has joined #synfig
04:03 < Yoyobuae> oh man, this will take hours on my lowly 1.6GHz CPU XD
04:04 < pxegeek> PIII or Core?
04:04 < Yoyobuae> AMD Sempron XD
04:05 < pxegeek> All my machines are Intel (except for the kids which is a Via CPU)
04:05 < pxegeek> I tried compiling on a PIII ~500MHz, and it took over 24 hours.
04:05 < Yoyobuae> i usually go with AMD for the cheapness factor
04:06 < Yoyobuae> 24 hours ROFL
04:06 < Yoyobuae> luckily synfig-core just took 1 hour on mine, it just finished building =D
04:07 < pxegeek> :)
04:12 -!- Yoyobuae [n=quete@201.224.135.156] has quit [Read error: 104 (Connection reset by peer)]
04:16 < htodd> Semprini
04:25 < factor> I have both single core intel and a dual core X2 amd both around 3G
04:52 < pabs3> :( review-board isn't released yet
05:00 < KiBi> o<
05:00 < KiBi> >o
05:20 < pxegeek> >o
05:21 < factor> Ojo
05:30 < AkhIL> n8 ..zzZ
05:32 < AkhIL> I have feature request. Importing images one by one for cut-out animation is not cool. Will be nice to have multi-import thing
06:22 -!- Zelgadis [n=zelgadis@87.103.170.242] has joined #synfig
07:27 -!- pxegeek [n=chatzill@c-71-59-140-184.hsd1.or.comcast.net] has quit [Read error: 113 (No route to host)]
07:43 -!- ZanQdo [n=Daniel@201.201.2.22] has joined #synfig
07:57 -!- ulrik [n=ulrikbod@213-65-95-57-no42.tbcn.telia.com] has quit ["bye! http://www.musikboden.se"]
09:44 -!- ZanQdo [n=Daniel@201.201.2.22] has quit ["Adios"]
09:50 -!- xerakko [n=Miguel@debian/developer/xerakko] has quit ["Ex-Chat"]
09:59 < pabs3> dooglus: have you got a way to do the svn switch thing without doing a ctrl+c?
10:00 < pabs3> dooglus: I'm thinking about creating an automatic nightly svn->tarball cronjob, and later automatic nightly .debs (and even later .exes)
10:01 -!- TMM [n=hp@c5147518c.cable.wanadoo.nl] has quit [Read error: 110 (Connection timed out)]
10:04 < AkhIL> pabs3: I have automated synfig updater. Whats your problem with svn?
10:05 -!- Zelgadis [n=zelgadis@87.103.170.242] has quit [Remote closed the connection]
10:09 < pabs3> AkhIL: I want a way to produce the tarballs on http://synfig.org/Subversion without using ctrl+c
10:10 < pabs3> I guess I can just do it manually though
10:18 < pabs3> dooglus: did you answer the Fedora people?
10:18 < AkhIL> pabs3: i never used ctrl+c with svn. Lat me show my update script. It may help
10:19 < AkhIL> http://pastebin.com/m65e3b950
10:21 < pabs3> http://synfig.org/Subversion is a bit more advanced than your script - allows you to update ETL/synfig-core/synfig-studio using one svn up
10:22 < AkhIL> but doesn't require ctrl+c 
10:28 < pabs3> dooglus: added a comment to the bug: https://bugzilla.redhat.com/process_bug.cgi#c11
10:29 < pabs3> I mean https://bugzilla.redhat.com/show_bug.cgi?id=428567#c11
10:35 -!- mwiriadi [n=mwiriadi@203-59-35-30.dyn.iinet.net.au] has joined #synfig
10:35 < mwiriadi> ping dooglus 
10:36 -!- TMM [n=hp@c5147518c.cable.wanadoo.nl] has joined #synfig
10:36 < pabs3> hi mwiriadi, that was quick
10:37 < mwiriadi> lol
10:37 < mwiriadi> I'm in Australia so it's early evening
10:37 < mwiriadi> I've just posted to the bugzilla
10:37 < pabs3> same, perth
10:37 < mwiriadi> same
10:37 < mwiriadi> Mount Nasura nice side of Armadale
10:38 < pabs3> I'm in freo area :)
10:38 < mwiriadi> ahh nice
10:38 < mwiriadi> anywhoo
10:38 < mwiriadi> I'm the packager for Fedora
10:38 < mwiriadi> I have no issues in providing any info to you but that thread is really the crux of the issues
10:39 < pabs3> firstly ETL-config isn't really used, synfig/studio use pkgconfig instead, we could probably remove ETL-config in the next version
10:40 < mwiriadi> ok
10:40 < AkhIL> I have exported parameter value. How I can reuse this value?
10:41 < pabs3> mwiriadi: I don't know why the profile and config headers are needed, that is how it was when it got GPLed
10:42 < mwiriadi> Fair enough
10:42 < pabs3> mwiriadi: would renaming a few defines fix the other issues?
10:43 < mwiriadi> The issue is mainly obviously etl is required for synfig which is required for synfigstudio
10:43 < mwiriadi> I would say they would
10:43 < pabs3> ETL is required for studio too :)
10:43 < mwiriadi> It refers mainly to autoconf and gcc4.3 which is entering Fedora 9
10:43 < mwiriadi> Yep
10:44 < pabs3> um, is 4.3 released yet?
10:44  * pabs3 would have thought it was too bleeding-edge to be used by default in a release distro
10:45 < mwiriadi> I've seen worse pabs3 
10:45 < mwiriadi> We at NM in alpha stage in Fedora 8
10:45 < pabs3> ok :)
10:45 < mwiriadi> The issue is mainly we will have lots of bugs
10:46 < mwiriadi> and while they will be filed upstream I would probably need you to jump in on the bugzilla to have a looksie
10:49 < pabs3> ok
10:53 < pabs3> right now, I should get ready to go out, but patches for the gcc4.3 issues and so on are welcome
10:54 < mwiriadi> No worries any that I come up with well be sent through
10:57 < pabs3> mwiriadi: btw, in svn we switched from the crazy bootstrap scripts to autoreconf 
10:57 < AkhIL> hey. visible togle checkbox in "Groups" dialogure doesn't work
10:58 < pabs3> mwiriadi: dunno if that is relevant to you
10:58 < mwiriadi> It will be when the next version is released
10:58 < pabs3> AkhIL: yeah, groups dialog contains wombles
10:58 < mwiriadi> Not a major issue just changed the spec file and it's all good
10:58 < pabs3> mwiriadi: .spec files are also dropped in svn since they were so shitty
10:58 < mwiriadi> Thats fine
10:59 < mwiriadi> I can send you mine when it's done
10:59 < mwiriadi> Only issue is that it will work in Fedora
10:59 < mwiriadi> If another rpm based distro wants them they can have them and alter appropriately
10:59 < pabs3> I tend to think packaging for anything other than windows & macos belongs in the distro
10:59 < mwiriadi> Yep
11:00 < mwiriadi> So do I but I'm more than happy to help out
11:00 < mwiriadi> It is OSS after all :)
11:00 < pabs3> cool :)
11:01 < pabs3> mwiriadi: btw, there are some SVN/old packages for fedora already: http://synfig.org/Download#Download
11:01 < pabs3> but getting synfig in officially would be great
11:40 < dooglus> pabs3: once you've done the ctrl+c 'svn switch' thing, you can just use 'svn up' to keep it up to date.  or just download my latest tarball, extract it, and 'svn up' that.  I've done the switch so you don't have to...
11:44 < dooglus> pabs3: review-board is available from the svn server - it's in use by companies apparently, and seems to work pretty well.
11:45 < dooglus> AkhIL: http://synfig.org/ListImporter shows you how to import multiple frames at once
11:46 < AkhIL> Thanks. I found answor for my question. And I like synfig!
11:46 < dooglus> pabs3: I didn't answer the Fedora people.  I saw their questions as mostly related to build/packaging and though that was more your thing
11:46 < AkhIL> soon i'll release my animation with all source files
11:54 -!- TMM [n=hp@c5147518c.cable.wanadoo.nl] has quit [Read error: 110 (Connection timed out)]
11:57 < mwiriadi> dooglus, actually most of the issues are upstream related
11:57 < mwiriadi> upstream as in you guys :)
12:02 < mwiriadi> /usr/bin/ETL-config is not multilib ready
13:08 < dooglus> mwiriadi: ok, sure.  I was suggesting that this is more pabs3's area than mine
13:08 < dooglus> mwiriadi: pabs3 packages it for debian, and also has worked on the build scripts, so I see that whole area as "over there"
13:17 < pabs3> :)
13:25 < mwiriadi> ahh fair enough
13:32 < CIA-46> synfig: dooglus * r1415 /synfig-core/trunk/src/synfig/layer_composite.cpp: There's no need to use a new variable here.
13:32 < CIA-46> synfig: dooglus * r1416 /synfig-studio/trunk/src/gtkmm/state_normal.cpp: Don't show "STATE NORMAL:..." every time the normal tool receives a new event.
13:32 < CIA-46> synfig: dooglus * r1417 /synfig-core/trunk/src/synfig/ (target_scanline.cpp target_tile.cpp canvas.cpp canvas.h): Pass the current 'time' value when optimizing layers. We're going to need it in a following commit.
13:32 < CIA-46> synfig: dooglus * r1418 /synfig-core/trunk/src/synfig/color.cpp: Number the blend functions in the array.
13:32 < CIA-46> synfig: dooglus * r1419 /synfig-core/trunk/src/synfig/color.cpp: Make a note to investigate blendfunc_ALPHA_BRIGHTEN. It seems to be multiplying the wrong thing by 'amount'.
13:32 < CIA-46> synfig: dooglus * r1420 /synfig-core/trunk/src/synfig/loadcanvas.cpp: If a canvas in a .sif file doesn't specify its version, inherit the version of its parent.
13:32 < CIA-46> synfig: dooglus * r1421 /synfig-core/trunk/src/synfig/loadcanvas.cpp: Following the previous commit, there's no need to walk up the canvas tree to find the root when checking the canvas version.
13:38 < dooglus> pabs3: will you have time to look at the issues in the fedora bug report?  if not, I can have a go, but I expect you'll get through them better than I would
13:59 < pabs3> dooglus: I'll try, I think they are a bit over my head, will mainly rely on fedora folks for patches or their thoughts on how to fix the issues
14:42 < factor> out
14:42 -!- factor [n=factor@ip68-14-160-70.ok.ok.cox.net] has quit ["Ex-Chat"]
14:43 < dooglus> I'm checking in the 'straight blend' changes I've made.  I expect genete will be around soon and he'll want to test them no doubt.
14:43 < dooglus> it's quite possible they're not finished, but I'm sure he'll find whatever's missing.  they seem to be ok from the testing I've done.
14:47 < CIA-46> synfig: dooglus * r1422 /synfig-core/trunk/src/synfig/ (5 files): 
14:47 < CIA-46> synfig: First attempt at fixing the 'straight' blend method. It now applies to the
14:47 < CIA-46> synfig: whole layer, rather than just the pixels modified by the layer. Loading old
14:47 < CIA-46> synfig: canvases that use 'straight' will convert the blend method to 'composite' for
14:47 < CIA-46> synfig: all layers other than PasteCanvas, since the old 'straight' worked much like
14:47 < CIA-46> synfig: composite for them. The 'straight onto' blend method still needs work.
14:47 < CIA-46> synfig: dooglus * r1423 /synfig-core/trunk/src/synfig/canvas.h: Edit in the svn revision number for the canvas 0.3 change.
15:04 < dooglus> genete: here's a new 'blendtest' that cycles through all 22 blend types for each of 15 different layers: http://dooglus.rincevent.net/random/blendtest.sifz
15:04 < dooglus> genete: I also separated the two halves of the screen so they don't overwrite each other
15:06 -!- photomuse [n=photomus@ics156-82.icsincorporated.com] has quit ["Konversation terminated!"]
15:24 < dooglus> checking for frames which show differences: blends (14 and 21) for layers (7 through 14) all fail
15:24 < dooglus> that's 'alpha brighten' and 'straight onto'
15:37 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has joined #synfig
15:56 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has quit []
16:16 -!- genete [n=Genete@84.122.32.42.dyn.user.ono.com] has joined #synfig
16:16 < genete> dooglus: you waiting me? :)
16:17  * genete just updating synfig
16:18 < dooglus> genete: hi :)  you've got some re-testing to do...
16:18 < genete> :)
16:18 < genete> hi dooglus 
16:18 < genete> what's your opinion on Dup layer issue of render ouside the main window?
16:18 < dooglus> I didn't touch 'straight onto' yet - that probably has all the same problems that 'straight' used to.
16:19 < genete> OK I'll take notice on that
16:19 < dooglus> genete: I've not looked at it yet.  I'll let you know when I do.  Did you open a bug for it?  I guess now that the layer is mostly stable you can do, rather than use the wiki
16:20 < genete> I'm not sure if really is a bug or not. I have to do more bench marcking because I feel that Duplicate Layer is so much more slow that manual duplicate the same layer
16:25 < dooglus> genete: don't worry about it - I'll look at it next, once I've got the blending sorted out.
16:27 < genete> ok. Anyway I'll do my bench marck tests for that
16:28 < dooglus> genete: check out my new blendtest.sifz - I made it 22fps, so the time (in seconds and frames) corresponds to the layer and blend method
16:28 < genete> ok I'm finishing building
16:29 < dooglus> genete: you can jump through time one seconds at a time to jump to the next layer, or one frame at a time for the next blend method.  you know the shortcuts?
16:29 < dooglus> control-. and control-shift-.
16:29 < genete> I don't remember now ...
16:29 < genete> thanks
16:29 < dooglus> and control-, and control-sgift-, to go back
16:30 < genete> but dooglus, if have you rendered that file and Straight is fixed, what is supposed I should find?
16:31 < dooglus> genete: you're supposed to know which combinations were broken before I guess
16:31 < dooglus> and check that they're ok now?
16:31 < genete> ok, nice then
16:33 < dooglus> genete: I guess more importantly, check whether the new changes break any of your existing sif files
16:33 < dooglus> did you use 'straight' at all?
16:33 < genete> I don't remember if in ony of my old test I used Straight
16:34 < genete> I guess ther should be one
16:34 < dooglus> I have a command-line for checking:
16:34 < dooglus> for i in *.sifz; do echo $(zgrep -A 1 blend_method $i | grep '="1"' | wc -l) $i; done | sort -n
16:34 < dooglus> it lists all the .sifz files with a count of how many 'straight' blends each has
16:35 < genete> ok I'll check out
16:35 < dooglus> replace the first 'z' with a '*' to include .sif files too
16:37 < genete> I have some of them... clouds, fire, sea-star, straight-faliure  (lol)
16:37 < dooglus> ha
16:38 < dooglus> and examples/*.sifz - they all (but one) use straight
16:38 < dooglus> I think 'straight' used to be the default, instead of composite
16:38 < genete> cloud2 have 7 straight methods!
16:38 < dooglus> when you load it, most of them will be converted to composite.  (all of them except for the pastecanvas layers)
16:38 < genete> hey! cloud3 have 132!!!
16:39 < dooglus> think why you used straight, if you really needed it, if composite works just as well, whether it's now broken with the new change.
16:39 < genete> ok
16:39 < genete> dooglus: I have a little problem, I have to go at 5:00 and studio still compiling....
16:40 < dooglus> studio doesn't really need to be rebuilt
16:40 < dooglus> you can run it now and it'll work just the same
16:40 < dooglus> it's just that I bumped the version number in canvas.h and all the studio files use that header
16:40 < dooglus> but I didn't change anything in it that will affect studio
16:41 < genete> ok
16:41 -!- Zelgadis [n=zelgadis@87.103.170.242] has joined #synfig
16:42 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has joined #synfig
16:43 < genete> blendtest.sifz:<string>:3553: warning: Undefined value in <string>
16:43 < dooglus> my one?
16:44 < genete> that's the firts message I obtained from your file when run in command mode
16:44 < dooglus> ok
16:44 < genete> now itis rendering normally
16:44 < dooglus> that's the layer name
16:44 < dooglus> at 15s there's no layer, so I set the layer name to be blank
16:44 < dooglus> it's warning that the string is empty...  not a problem
16:45 < genete> how did you to cycle the layers?
16:49 < dooglus> genete: timeloop convert
16:50 < genete> have to look at it when studio finish
16:50 < dooglus> you should be able to run studio while it's building
16:50  * genete is playing the difference game now
16:51 < dooglus> oh, you rendered the hundreds of images to files and are viewing them?
16:51 < genete> yes
16:51 < genete> text layer + Straight is not fine, right?
16:52 < dooglus> oh?
16:52 < dooglus> what are the numbers?
16:52 < genete> oh sorry
16:52 < genete> it was the vertical division you made to separate the samples
16:52 < genete> it is fine
16:53 < dooglus> ok
16:53 < dooglus> I just added new encapslations, and blended each half 'Onto' a rectangle covering half the window
16:53 < genete> Star + Alpha Brigthen is the weirdest thing I've never seen...
16:54 < dooglus> alpha brighten uses whichever color is the most/least transparent?
16:55 < dooglus> it will either chose the green of the star or the blue of the candy, based on the transparency of each.  it doesn't blend the colors - it choses one of the other
16:55 < genete> rectangle + Straight onto seems ok but really they don't blend the same. Zoom it and you'll see a thin line different there.
16:55 < genete> frame 241
16:56 < dooglus> the frame should have 2 numbers in it.  the layer number is the seconds; the blend number is the frames
16:56 < dooglus> rectangle (10) + straight (1) is "10s 1f"
16:57 < genete> 10-rectangle - 21 straight onto
16:58 < genete> reviewed 330 images....
16:58 < genete> Strasight seems ok there. This evening will test my old files
16:58 < genete> now have to go.
16:58 < genete> ttyl
16:59 -!- genete [n=Genete@84.122.32.42.dyn.user.ono.com] has quit ["Abandonando"]
17:20 < dooglus> genete: I think "10s 21f" is very wrong
17:23 < dooglus> genete: "straight onto" with almost no overlap should result in an almost blank result - see http://synfig.org/Blend_Method#Straight_Onto - see how the tux's feet are hidden by the transparent gradient?
17:36 -!- Zelgadis [n=zelgadis@87.103.170.242] has quit ["Leaving."]
18:01 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has quit []
18:11 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has joined #synfig
19:00 -!- Yoyobuae [n=quete@201.224.135.156] has joined #synfig
19:01 < Yoyobuae> hi
19:02 < Yoyobuae> I see Straight layer is fixed =)
19:05 < Yoyobuae> will synfig build take less time when a rebuild it?  (built both core and studio yesterday, svn1414)
19:06 < Yoyobuae> s/when a/when I/
19:09 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has quit []
19:23 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has joined #synfig
19:46 -!- TMM [n=hp@ip565b35da.direct-adsl.nl] has joined #synfig
20:27 < dooglus> Yoyobuae: I have things set up so it takes about 15 seconds to rebuild, mostly.  but if canvas.h gets changed, then that's a 15 minute rebuild anyway.  everything depends on that.
20:28 < dooglus> Yoyobuae: some header files are used all over the place.  the 'fix' for 'straght' changed one of them, so the rebuild will take a long time
20:43 < Yoyobuae> i think i will need change the build scripts i use a bit
20:43 < Yoyobuae> the ones that the Windows build suggest always build everything from scratch, i think
20:44 < Yoyobuae> i did "svn up" and ran the script before, it took the same time as the first build
20:45 < Yoyobuae> i ran the script today, i mean
20:51 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has quit []
20:59 < Yoyobuae> http://members.lycos.co.uk/yoyobuae/Selective_Duplicate.sifz
21:00 < Yoyobuae> try changing "Hide_Up_To" and "Show_Up_To" values on the Child "panel" =)
21:01 < Yoyobuae> you can control which of the duplicate circles are shown, and which are "hidden" (semitransparent actually, but easily changed to completely hidden)
21:15 -!- genete [n=Genete@84.122.49.220.dyn.user.ono.com] has joined #synfig
21:20 < genete> dooglus: yeah, Straight Onto is very bad for most of the layers
21:22 < Yoyobuae> hi genete, see log right before you logged in =)
21:22 < genete> I'm loading your file...
21:23 < genete> but first dooglus request ;)
21:23 < Yoyobuae> ok
21:23 < Yoyobuae> my file is quite simple really =)
21:23 < genete> I'm looking at it
21:25 < genete> hey! how did you unexport the index. It should be ilegal!. I know how to do it but did not want to ask dooglus to fix it.
21:26 < genete> oh well it is allowed =)
21:26 < genete> did not remember it or understodd it wrongly
21:26 < Yoyobuae> if i remember correctly, disconnecting Duplicate index is ilegal
21:26 < genete> but not unexporting
21:27 < Yoyobuae> yup, you can re-export again when needed
21:28 < Yoyobuae> the right-click menu on Index ONLY allows Export, Unexport and Rename Exported Name
21:29 < genete> Yoyobuae: so your file is working ok or still have problems. I don't understand what what to do
21:30 < Yoyobuae> did you change the two exported values on the child window?
21:30 < genete> yes
21:31 < Yoyobuae> thats all it's meant to do, show/hide some of the duplicated circles
21:32 < Yoyobuae> i just left the "hidden" circles semitransparent for the example
21:32 < genete> I understand. I think you should change the min value (now as 0.5) to 0.0 to work properly. 
21:32 < Yoyobuae> yup
21:32 < Yoyobuae> the interesting part is that with Amount=0.0 the layers don't render at all
21:32 < genete> but for working properly you have to set Hideupto = -1 and Showupto = 11 to show the 12 circles
21:33 < genete> yeah very interesting
21:33 < genete> can I use that for my particles ?
21:33 < Yoyobuae> yeah, use it however you can =)
21:34 < genete> :)
21:36 < Yoyobuae> also, this trick is also useful with the new fixed Straight blend method
21:36 < Yoyobuae> you can show/hide the layers underneath some specific layer
21:36 < genete> how?
21:37 < Yoyobuae> ill make a new example =)
21:37 < genete> ok
21:40 < Yoyobuae> hmm, the new "About" window background image is waaay to dark right now
21:41 < genete> due to the straight fix?
21:42 < Yoyobuae> http://members.lycos.co.uk/yoyobuae/synfig_about.png
21:42 < Yoyobuae> it might be =D
21:42 < genete> that's wrong
21:42 < genete> I have a nice one
21:44 < Yoyobuae> weird =/
21:44 < Yoyobuae> did you rebuild studio?
21:45 < genete> I'm on the latest
21:46 < genete> this is the old one: http://i85.photobucket.com/albums/k74/Genete/synfig/about2.png
21:47 < genete> this is from today http://i85.photobucket.com/albums/k74/Genete/synfig/about-1.png
21:47 < genete> look the dates
21:47 < genete> and the versions
21:53 < Yoyobuae> lol, my build is messed up then XD
21:53 < genete> tha can cause that strange render
21:54 < genete> sometimes I make a manual update (no script) and forget to install. Then when run studio usually crash.
21:54 < dooglus> genete: I don't want you to be able to unexport the index either.  how do you do it?
21:55 < genete> I went to the child list and right click unexport. I was surprised too.
21:56 < Yoyobuae> but why not unexport?
21:57 < genete> Yoyobuae: you should always know what index is the parameter linked to when more than one dup layer exists
22:00 < dooglus> genete: about_dialog.sif does render too dark now
22:00 < genete> but I think that Rename exported valuenode implies unexport the old name
22:00 < genete> not mine
22:00 < dooglus> genete: your build script probably didn't rebuild it - try opening it in studio - isn't it dark?
22:01 < Yoyobuae> genete: But lets say you download a template with 10 Duplicate layers
22:01 < Yoyobuae> you don't know, nor want to know about the inner workings of that template
22:01 < Yoyobuae> yet Index 1 thru Index 10 appear on the child window
22:02 < dooglus> Yoyobuae: I guess you want to be able to unexport the index children, but still have them linked where needed
22:02 < genete> Yoyobuae: it is reasonable
22:03 < genete> dooglus: yes it is darken
22:05 < dooglus> genete: any idea why?
22:05 < genete> I'm searching if there is a stright there
22:05 < genete> straight
22:06 < dooglus> genete: there are lots of them
22:07 < dooglus> 16, apparently
22:08 < dooglus> genete: none of them are pastecanvas layers, so they'll all have been converted to composite now
22:09 < genete> phone...
22:11 < genete> ouch!
22:11 < genete> how to know what where the straight ones?
22:12 -!- photomuse [n=photomus@aluru-1.ece.iastate.edu] has joined #synfig
22:12 < genete> Frame->Gradient060. If change to straight onto it looks very similar to the old one
22:13 < Yoyobuae> oh, maybe straight onto was also affected, indirectly
22:13 < genete> I think no
22:14 < genete> I think that straight onto is doing the same than the old straight becasue the layer below covers all the canvas
22:16 < genete> the logo.sifz file also renders badly.
22:19 < photomuse> Are the directions to download the code on the wiki up to date?
22:22 < genete> photomuse: I think so
22:23 < genete> Yoyobuae: I cannot make visible the "ALPHA" yellow black bar, canyou?
22:31 < Yoyobuae> inside "Alpha Test" encap, turn off "Spherize"
22:32 < genete> yes I found it, but is spherize also broken?
22:37 < genete> oh dear! spherize is broken!
22:38 < genete> :(
22:38  * Yoyobuae wonders what other layer just broke XD
22:45 < genete> Mmm revert one step takes a lot of time due to the last modification touched a .h file...
22:48 < dooglus> genete: there's no real canvas.h change.  touch it to a future date to avoid rebuilds
22:48 < dooglus> genete: how is spherize broken?
22:48 < genete> dooglus: anyway I'm reverting to 1421 to see if you broke spherize there
22:49 < genete> it just doesn't render what is below
22:49 < genete> no matter what values have its parameters
22:51 < dooglus> oh, I see.
22:51 < dooglus> I blame you, really, for not including 'spherize' in your blendtest.sif :)
22:52 < genete> Spherize is not a geometry layer. It is a distortion one
22:52 < dooglus> yes
22:52 < dooglus> you test suite was unfortunately incomplete :(
22:52 < dooglus> haha
22:52 < genete> BTW there is a situation where Striaght should act like straight and not as composite
22:53 < genete> I understand that what you have done is emulate that every straight layer is composited as it is were encapsulated isn't it?
22:55 < genete> I understand "straight" as: "ok render this layer and none of the layers behind"
22:55 < genete> do you know where Straight is useful?
22:55 < genete> It is useful for noise distort
22:55 < genete> noise distort is the only one distort layer that have blend method I think
22:58 -!- MangoFusion [n=jamesu@host81-132-118-101.range81-132.btcentralplus.com] has joined #synfig
22:58 < genete> so if you only want to render the noised distorted image you have to set the noise distort layer to straight
22:58 < dooglus> straight with amount=1 is "just render this layer, and nothing under"
22:59 < genete> so Straight is affected by amount. I didn't know it.
22:59 < dooglus> straight with lower amounts is "render amount*100% of this layer and (1-amount)*100% of the layers underneath
22:59 < dooglus> make a new document, draw 2 circles, make the top one 'straight', and play with the amount
23:00 < dooglus> at amount=0, you only see the bottom one, at amount=1, you only see the top, in between, you see states in between
23:01 < genete> so what did you do at r1422? (in a few words)
23:02 < dooglus> heh
23:02 < genete> I'm not blaming you, I really whant to understand the changes
23:02 < dooglus> few words: "changed 'straight' so that every pixel in the layer has an effect on the result, even the transparent ones"
23:03 < genete> ah so that was the reason that it worked with "infinite" layers
23:03 < genete> they havent bounding box
23:03 < dooglus> a 'straight' blend is a blend that doesn't take transparency into account.  if you blend 100% of (r:1, g:0, b:0, a:0) onto (r:0, g:1, b:0, a:1), you get the first one.
23:04 < dooglus> 100% of A onto B == A
23:04 < dooglus> even if A is completely transparent
23:04 < dooglus> understand?
23:04 < dooglus> infinite layers always worked over the whole plane, even the transparent pixels, yes.
23:04 < dooglus> finite layers were lazy, thinking that the pixels outside their area of effect didn't do anything.
23:05 < genete> but they do or should do.
23:05 < dooglus> but for 'straight', *every* pixel does something, even the transparent ones
23:05 < dooglus> right.  they should, and do now.
23:05 < genete> Straight Onto have same "lazy" behavior
23:05 < dooglus> and, so should the transparent pixels in a 'straight onto' - but they don't yet.
23:05 < dooglus> you're a little ahead of me on each line :)
23:06 < dooglus> but that's good.  you agree with me...
23:06 < dooglus> the example you saw earlier - the rectangle + straight onto 
23:06 < dooglus> they were quite similar?
23:06 < genete> you're some minutes earlier than me. Praghe is eastern than Cadiz
23:06 < dooglus> but they were both very wrong - the candy should be hidden completely by the transparent pixels around the rectangle
23:06 < dooglus> yes, but I've been drinking wine since noon
23:07 < dooglus> so I must be a little behind...
23:07 < genete> yes the rectangle were both bad
23:07 < genete> xD
23:07 < dooglus> right.
23:08 < genete> I'll render the logo again and see what happen
23:09 < genete> weeeeee. svn r1421 fix spherize again!
23:09 < genete> and of course all the wrong straight ones.
23:10 < dooglus> ok
23:10 < dooglus> I'll look at that next
23:10 < dooglus> it'll be tomorrow
23:10 < dooglus> I drank too much today to do any more coding
23:10 < genete> yes, better.... when wine effect has gone....
23:13 < dooglus> the only problem I can see is there's another 5 bottles in the fridge...
23:13 < dooglus> :)
23:18 < genete> :)
23:25 < dooglus> I'd rather get 'straight' working properly before looking at 'straight onto' (and alpha-whatever, the other one that's broken)
23:25 < dooglus> (#14, whatever that is)
23:26 < genete> yes but take care on the distortions layers... ;)
23:27 < dooglus> do you want to extend the blending sif to include them?
23:27 < genete> I'll open yours and see how did you do the layer loop 
23:28 < dooglus> it was pretty straight forward - it's a timeloop convert of a linear
23:28 < genete> ok
23:31 < genete> why did you need a switch on the blend method?
23:31 < dooglus> oh, because switching 'amount' to 0 isn't enough
23:31 < dooglus> oh, maybe I'm wrong, I don't know
23:32 < dooglus> I was thinking that a straight blend, even with amount 0, was still going to interfere with things.
23:32 < dooglus> maybe I didn't need it.
23:32 < dooglus> I just wanted to be sure that everything was a simple 'composite' other than the layer of interest
23:33 < genete> I see all the "regular" layers to have amount=0 except plant
23:33 < dooglus> depends on the time
23:33 < genete> ah I see 
23:34 < dooglus> plant is the first one
23:34 < genete> depends on the time
23:34 < genete> lol
23:34 < dooglus> it has amount 1 from 0s to 1s-1f
23:35 < genete> ok so new layers have to set that amount manualy right?
23:35 < dooglus> I did it using constant waypoints
23:36 < genete> ok
23:36 < genete> ah, it is a 22fps file.
23:37 < genete> I understand
23:37 < dooglus> you'll need to change the framerate
23:38 < dooglus> and then the blend method text string waypoints will be out of step
23:38 < dooglus> I edited the .sifz file by hand - instead of "time=0.0466666s", I put "time=1f", "time=2f", etc. for each
23:39 < genete> 22 is the number of methods why need to change the frame rate?
23:41 < dooglus> ha, I'm sorry.
23:41 < dooglus> that was the wine talking.
23:41 < genete> eh!
23:41 < dooglus> I'll sign off.  I'm no use to anyone here...
23:42 < genete> :)
23:42 < dooglus> if you invent a new blend method, I meant.   um...  ahem...
23:43 < dooglus> it would be good to get that blend test sifz to be quite complete, and then add it to svn as part of a proper test suite
23:44 < genete> but Spherize doens't have blennd method...
23:44 < genete> when include it?
23:45 < genete> and where
23:45 < dooglus> I guess that's a separate test
23:45 < dooglus> "blend.sif" and "distort.sif"?
23:45 < dooglus> distort.sif for the various distort types
23:46 < genete> but noise distort has blend method too. It should be in both tests
23:46 < dooglus> we can check "correct" png files in, and the makefiles can run synfig on the .sif files and compare the output with the "known good" output.
23:46 < dooglus> if there are differences, it tells us, and we can check whether the differences are good or not.  if so, we check in the improved "known good" png files
23:47 < genete> it takes a little to render... should makefile decide when render the tests? or for every svn up?
23:47 < dooglus> yes, there's no problem with having something tested in two different places
23:47 < dooglus> it renders when you type "make test" generally
23:47 < genete> ok fine
23:48 < dooglus> so for you, just "svn up" then "make", you won't see it
23:48 < genete> Bevel and Shade should go at blend.sifz also.
23:48 < dooglus> but maybe pabs can have the synfig.org machine do a "svn up", "make test" each day and email me about failures, or get it to send messages to this chat room, or some such; like CIA does about checkins
23:49 < genete> but the errors should be compared by human eye
23:49 < dooglus> automated tests are a good thing, because people generally won't be testing everything regularly.  I didn't check "spherize" at all before checking in these "straght" fixes, for instance.
23:50 < genete> or can have the png file a msdsum when correct?
23:50 < dooglus> for the most part, you would expect the resulting .png files to be identical.  if they're not, then a human needs to get involved.
23:50 < dooglus> right
23:50 < genete> if the msdsum is possible we can aviod human review
23:51  * genete believes more everyday that programmers are the most lazy people in the world...
23:51 -!- photomuse [n=photomus@aluru-1.ece.iastate.edu] has quit ["Konversation terminated!"]
23:51 < dooglus> you know, I add a new layer, like "duplicate".  it's a relatively big change, I wouldn't expect it to break anything, but it's still good to have a daily check - do the 300 various blend method images still look the same?  yes?  good!
23:51 < Yoyobuae> is it posible for 2 PNGs to have the same exact colors on all pixels, but the files be different?
23:52 < dooglus> genete: some changes may actually improve the image quality - removing a slight artifact, or something, in which case the md5sum check fails, human needs to check the new images are better, and check in the new images instead.
23:52 < Yoyobuae> or will the test compare pixel data itself?
23:52 < dooglus> Yoyobuae: I think they may be a timestamp or some such in a .png file - I'm not sure.  if .png doesn't do it, I'm sure some other format will though.
23:52 < dooglus> Yoyobuae: the easiest way is just to md5sum the files.
23:53 < Yoyobuae> imagemagick has a compare.exe, dunno how it works though
23:53 < genete> dooglus: ok but anyway md5sum is the first automatic tester. If it fails then enter genete's eyes.... or imagemagick compare  or whatever
23:54 < dooglus> looks promising:
23:54 < dooglus> $ synfig pirates.sifz
23:54 < dooglus> $ mv pirates.png pirates1.png
23:54 < dooglus> $ synfig pirates.sifz
23:54 < dooglus> $ mv pirates.png pirates2.png
23:54 < dooglus> $ md5sum pirates?.png
23:54 < dooglus> 475fed9d531749a23c3ba3109886a78f  pirates1.png
23:54 < dooglus> 475fed9d531749a23c3ba3109886a78f  pirates2.png
23:55 < dooglus> can you try the same command?
23:55 < dooglus> do you get the same md5sum for pirates.png?
23:56 < genete> 56f9359f92ca9780aba02d760cee9ca7  pirates.png
23:56 < genete> :(
23:56 < dooglus> Yoyobuae: it's possible for 2 pngs to have the exact same pixels, but a different "creator" string - there are a few "metadata" fields in a png which don't affect the image itself
23:56 < dooglus> genete: render it again.  do you get the same next time?
23:57 < genete> 56f9359f92ca9780aba02d760cee9ca7  pirates.png
23:57 < dooglus> genete: I guess you have a different version of the png library
23:57 < genete> command?
23:58 < dooglus> $ identify pirates2.png 
23:58 < dooglus> pirates2.png PNG 648x526 648x526+0+0 DirectClass 8-bit 275.221kb 
23:58 < genete> pirates.png PNG 648x526 648x526+0+0 DirectClass 275kb 
23:58 < dooglus> $ dpkg -l 'libpng*-dev' | grep ^ii
23:58 < dooglus> ii  libpng12-dev   1.2.15~beta5-2 PNG library - development
23:58 < genete> ii  libpng12-dev   1.2.8rel-5.1ubuntu0.2 PNG library - development
23:59 < dooglus> not surprising, you have a different version of ubuntu than me
23:59 < genete> yagh!
23:59 < dooglus> identify -verbose pirates2.png
23:59 < dooglus> shows lots of stuff
--- Log closed Wed Jan 23 00:00:16 2008