--- Log opened Wed Mar 21 00:00:02 2007
00:11 -!- pabs3 [n=pabs@dsl-220-235-163-152.wa.westnet.com.au] has joined #synfig
00:21 -!- xerakko [n=xerakko@debian/developer/xerakko] has quit [Read error: 110 (Connection timed out)]
00:23 < dooglus> I love it when a plan comes together
00:27  * pabs3 files a bug on the internet for forcing me to use a light colour theme on my desktop
00:28 < dooglus> because all the web pages have light backgrounds?
00:29 < pabs3> yep
00:30 < pabs3> wots this plan of yours?
00:31 < dooglus> I've been staring at the lineargradient/gradient/color code for a long time, and it's beginning to come together in my mind
00:31 < dooglus> there are at least 2 bugs in it that cause the weird looking gradients
00:32 < dooglus> but my current fix looks a bit like this:
00:32 < dooglus> Color pool(gradient(1.0-((1.0-(dist-supersample*0.5))*0.5),1.0-(dist-supersample*0.5)).premult_alpha()*(1.0-(dist-supersample*0.5))/supersample);
00:32 < dooglus> which could probably be a little clearer...
00:33 < pabs3> ah, hmm, ok
00:35 < dooglus> a pixel on the screen can take its colour from 2 different ends of a looped gradient
00:35 < dooglus> the existing code assumes it takes 50% of its colour from each end of the gradient, I think
00:43 < dooglus> is x*0.5 the same as x/2 if x is a float?
00:43 < dooglus> is there any reason to prefer *0.5 over /2?
00:43 < dooglus> it makes ugly code uglier, but other than that...
00:45 < pabs3> not really. unless maybe divides are slower on some CPU
00:46 < pabs3> I expect compilers will be able to optimise that though
00:48 < igli> divides are evil
00:49 < dooglus> I would expect *0.5 and /2 to generate the same code, but maybe I'm wrong
00:58 < igli> only if compiler is smart
01:03 < dooglus> ok
01:03 < dooglus> I'll leave the 0.5 there then
01:06 < igli> man floats are weird anyway, mantissa and exponent bit fields
01:06 < dooglus> sure
01:06 < igli> you can't just shift em
01:07 < dooglus> well...
01:07 < dooglus> if you're doing a /2 you can just decrement the exponent, can't you?
01:07 < igli> man don't ask me i'm in bash mode atm :)
01:07 < dooglus> heh
01:08 < igli> long time since i looked at binary floats
01:08 < dooglus> I think the value of a float is mantissa*2^exponent, isn't it?
01:08 < dooglus> like a number in scientific notation (3*10^6) but using base 2 instead of 10
01:09 < igli> yeah should be
01:09 < igli> you're right i think :)
01:10 < igli> that'd be the fastest way if the maths holds :)
01:15 < dooglus> I've not looked at how floating point numbers are stored since writing some Z80 assembly code on my ZX81...
01:15 < dooglus> I guess things may be done slightly differently now
01:21 < igli> heh BBC-B and Amstrad, but i remember the zx80 and 81; man those PSUs
01:28 < dooglus> power supply units?
01:30 < igli> heh i just remember that if we nudged it it would reset
01:31 < igli> something on the back of it. i was only 7 ffs ;)
01:31 < dooglus> ok, the 16K RAM pack.
01:32 < igli> nah before that, when it had 1k
01:32 < dooglus> the only thing that would cause mine to crash regularly without the RAM pack plugged in would be if I used 'FAST mode'
01:33 < igli> zx80?
01:33 < dooglus> (that's where the CPU doesn't bother updating the screen if it's busy)
01:33 < igli> heh
01:33 < dooglus> I never had a zx80 - that *only* had fast mode...
01:33 < igli> always wanted a C64
01:33 < igli> had hardware sprites
01:33 < igli> all my CPU time went on graphics
01:34 < igli> hated it
01:34 < igli> lol
01:34 < igli> them were the dayz
01:36 < dooglus> I was always envious of my friends with C64s, too
01:36 < igli> innit
01:36 < dooglus> they had everything - sounds, graphics, the lot!
01:36 < igli> yeah SID chip
01:36 < igli> Llamasoft!
01:37 < dooglus> Yak's still coding games isn't he?
01:37 < igli> Minter? dunno
01:37 < dooglus> http://www.llamasoft.co.uk/
01:37 < igli> woah wicked
01:37 < igli> checking
01:38 < dooglus> "Llamasoft has pioneered the Light Synthesiser genre and continues to push the boundaries of this exciting technology today.
01:38 < dooglus> that'll be 'trip-a-tron' then...
01:38 < igli> was just thinking of psychedelia
01:38 < igli> thanks for the link ;)
01:55 < dooglus> #llamasoft on irc.blitzed.org too, apparently
01:58 < dooglus> I've nearly fixed the gradient bug: http://dooglus.rincevent.net/synfig/gradient2.html
01:58 < dooglus> but there's something still not quite right there
01:58 < dooglus> (see the last picture)
02:02 < pabs3> go dooglus!
02:21 -!- pxegeek [n=Miranda@c-71-59-140-184.hsd1.or.comcast.net] has joined #synfig
02:30 -!- pxegeek [n=Miranda@c-71-59-140-184.hsd1.or.comcast.net] has quit [Remote closed the connection]
02:32 -!- teak [n=teak@24.244.163.157] has joined #synfig
02:43 -!- pxegeek [n=Miranda@c-71-59-140-184.hsd1.or.comcast.net] has joined #synfig
02:48 < pxegeek> So how come it all get's quiet when I join?  ;)
02:49 < pabs3> dooglus probably got distracted by code :)
02:49 < pxegeek> It's just taken me 20 mins to read up what happened since I was last online
02:50  * pxegeek was a ZX81 -> BBC Micro user
02:52 < pxegeek> dooglus - you can still get zx81's if you want to relive your childhood - http://www.zx81kit.com/
02:54 < dooglus> heh
02:54 < dooglus> I don't think so.
02:55 < dooglus> did you see I made the hex colour box you were asking for?
02:55 < dooglus> anyway - goodnight.  it's late.
02:58 -!- teak [n=teak@24.244.163.157] has quit ["Leaving"]
03:29 -!- tonsofpcs [n=tonsofpc@ool-435385da.dyn.optonline.net] has quit [Remote closed the connection]
03:29 -!- tonsofpcs [n=tonsofpc@ool-435385da.dyn.optonline.net] has joined #synfig
03:31 -!- tokyo [n=tokyo@c-67-175-134-199.hsd1.il.comcast.net] has quit [Read error: 104 (Connection reset by peer)]
03:35 -!- tokyo [n=tokyo@c-67-175-134-199.hsd1.il.comcast.net] has joined #synfig
03:41 -!- pxegeek [n=Miranda@c-71-59-140-184.hsd1.or.comcast.net] has quit [Read error: 110 (Connection timed out)]
04:48 -!- pxegeek [n=Miranda@71.59.140.184] has joined #synfig
04:49 < pxegeek> Thankyou dooglus!  I'll look forward to an updated windows build! 
04:52 < pabs3> dooglus: btw, what do you think about that problem building on win32 that atrus mentioned?
06:26 -!- pxegeek [n=Miranda@71.59.140.184] has quit []
06:44 -!- crazy_bus [n=philip@203.51.37.184] has joined #synfig
07:18 -!- igli [n=igli@82.152.221.75] has quit [Read error: 104 (Connection reset by peer)]
08:42 < dooglus> pabs3: which problem is that?  I remember he had a problem, but didn't he fix or work around it?
08:44 < pabs3> see his latest message on the devel list - your suggestion of undefining DELETE worked, but he asked if that is the correct solution
08:46 < dooglus> ok
08:46 < dooglus> I just went through the lists and hid all the spam (on the sf.net archive)
08:47 < pabs3> oh, cool. I didn't know that could be done
08:47 < dooglus> you have to click the red 'no entry' icon in the top right while viewing spam
08:47 < dooglus> "6 messages have been excluded from this view by a project administrator."
08:48 < pabs3> ah
08:50 < dooglus> I can't find his message
08:51 < dooglus> it's not as if it's a busy list
08:51 < dooglus> is it here? http://sourceforge.net/mailarchive/forum.php?forum_id=46804
08:53 < pabs3> lookn
08:54 < dooglus> it's ok
08:54 < dooglus> I see it in my inbox...
08:54 < dooglus> sf.net is giving me '500 - Internal Server Error'
08:55 < pabs3> http://news.gmane.org/find-root.php?message_id=%3c46000010.000004.24434%40webmail14.yandex.ru%3e
09:20 -!- crazy_bus [n=philip@203.51.37.184] has quit [Remote closed the connection]
09:38 < dooglus> ok
09:38 < dooglus> I see the problem
09:39 < dooglus> gtkmm/preview.h is including <gtkmm/canvasview.h> before <ETL/clock>
09:39 < dooglus> switching those 2 include lines will fix the problem
09:39 < dooglus> ETL/clock indirectly defines DELETE
09:40 < dooglus> and canvasview indirectly fixes that definition, but only if DELETE has already been defined
09:42 < pabs3> hmm, ok
09:43 < pabs3> perhaps reply to atrus and ask if that also fixes it
09:43 < dooglus> I have done
09:44 < dooglus> with nasty diagram of how the #include stuff is happening
09:44 < dooglus> I've commited a change to preview.h which should fix it
09:45 < dooglus> a few days ago you mentioned how the Changelog was showing ugly paths when I commit to synfig and synfig-studio at the same time.
09:46 < dooglus> I wonder what we should do about that.  it looks to me like a bug in svn2cl
09:52 < pabs3> probably is a bug
09:52 < pabs3> hmm, although
09:53 < pabs3> I don't think there is any other correct behaviour
09:54 < dooglus> well, I think it should only mention files which are in or under the current directory
09:54 < dooglus> the other files will be mentioned in their own directory's changelog
09:56 < pabs3> hmm, ok
09:56 < dooglus> I looked at the svn2cl script, but it wasn't obvious how to fix it
09:58 < pabs3> btw, yr message to the list was held, shall I let it thru?
09:59 < pabs3> need to modify the .xsl file
10:01 < dooglus> no, delete it
10:01 < dooglus> I'll send it again from the proper address :)
10:01 < pabs3> ok
10:02 < dooglus> that's better.
10:02 < dooglus> I was wondering why it wasn't showing up on gmane
10:06 < dooglus> it might be worth linking to the gmane archives for the mailing lists from the home page
10:06 < dooglus> it seems the sf.net archives are out of date
10:06 < pabs3> yeah
10:10 < dooglus> is the bug tracker supposed to email the devel list with all new bugs?
10:10 < dooglus> I see some on http://news.gmane.org/gmane.comp.video.synfig.devel , but most aren't there
10:11 < pabs3> yes. I only recently fixed it so they would get thru. there were tons waiting in the admin, so I just discarded them not wanting to flood the list
10:12 < dooglus> ok
10:12 < dooglus> interesting that spam was getting through but mail from sf.net itself wasn't...
10:13 < pabs3> yeah, it wasn't addressed to the list, triggered the implicit destination test
10:15 < dooglus> I see
10:15 < dooglus> good that it's working again
10:39 -!- xerakko [n=xerakko@iescamarles.xtec.net] has joined #synfig
11:28 -!- xerakko [n=xerakko@debian/developer/xerakko] has quit [Read error: 104 (Connection reset by peer)]
11:35 -!- xerakko [n=xerakko@iescamarles.xtec.net] has joined #synfig
11:46 -!- xerakko [n=xerakko@debian/developer/xerakko] has quit [Read error: 104 (Connection reset by peer)]
12:08 -!- tokyo [n=tokyo@c-67-175-134-199.hsd1.il.comcast.net] has quit ["the house bug has bitten!"]
12:30 < dooglus> at last!  http://dooglus.rincevent.net/synfig/gradient2.html
12:31 < dooglus> it's good to be able to fix a bug and make the code smaller at the same time :)
12:35 < pabs3> yay!
12:40 < pabs3> ok, I'm starting a News page on the wiki
12:43 < pabs3> for now it just has "dooglus continues to kick the ass of tons of bugs", but I'll go thru the backlog of mail/irc/svn and add stuff
12:43  * pabs3 bbl, tv
12:53 < dooglus> :)
14:05 < dooglus> is there some way to ask "file > render" to use both CPU cores?  it currently only uses one of them
14:24 < dooglus> also, which rendering, synfigstudio seems to attempt to keep responding to new input, but it does so very slowly, even though one of the CPU cores is entirely free.
14:24 < dooglus> it would be good if the rendering could be in a separate thread (if it isn't already)
14:38 < pabs3> indeed
15:07 < pabs3> http://www.worm.bluesfear.com/index2.html
16:30 < pabs3> Bombe: hmm, your blog seems broken
16:36 < Bombe> Hmm, there's something wrong with the way it handles its own permalinks.
16:36 < Bombe> Thanks for bringing my attention to it, work is boring. :)
16:36 < dooglus> another bug?  http://dooglus.rincevent.net/synfig/gif-target.html
16:37 -!- wildhostile [n=chatzill@ALamentin-104-1-41-164.w80-8.abo.wanadoo.fr] has joined #synfig
16:37 < wildhostile> yep dooglus 
16:37 < dooglus> ?
16:38 < wildhostile> what's up?
16:38 < dooglus> nothing much
16:38 < dooglus> slowly driving myself mad with synfig debugging :)
16:38 < wildhostile> it works now
16:38 < dooglus> good stuff
16:38 < wildhostile> very good
16:39 < wildhostile> i have to see how it work now
16:39  * dooglus goes to update http://wiki.synfig.com/Build_instructions and sees that pabs beat him to it...
16:40 < pabs3> dooglus: how many colours in the png? iirc gif can only have 256
16:40 < wildhostile> wow
16:41 < wildhostile> very good pabs3 
16:41 < dooglus> ooh
16:41 < pabs3> dooglus: which bit? the pkg-config stuff? added that last nite
16:41 < wildhostile> i see you works hard
16:42 < wildhostile> work
16:42 < dooglus> synfig(30211): info: curr_palette.size()=254
16:42 < dooglus> that's for the gif
16:42 < dooglus> so it's a problem of palette optimisation
16:42 < Bombe> My blog's fixed, thanks pabs. :)
16:42 < dooglus> it's not chosing the closest available colour for some of those pixels
16:43 < pabs3> hmm, now I vaguely remember reading somewhere a trick to make gifs support more than 256 colours. seems like it was from planet gnome
16:43 < pabs3> Bombe: cool :)
16:44 < pabs3> dooglus: gimp says 1883 colours in the png
16:45 < dooglus> right
16:45 < dooglus> that'll probably be it then
16:46 < dooglus> it still looks like a bug though - see those 3 dark pixels in the lower right?
16:46 < pabs3> yeah
16:46 < dooglus> there's definitely a better choice for those
16:48 < dooglus> how did you get the gimp to tell you how many colours there are?
16:48 < pabs3> aha http://jimmac.musichall.cz/weblog.php/Oops_Dept/MightyGIF
16:48 < pabs3> layer -> color -> colourcube analysis
16:53 < dooglus> nice
16:54 < dooglus> that hi-colour gif page is interesting
16:57  * pabs3 wonders if it works in ie
16:59 < dooglus> I have ie6 for wine: http://dooglus.rincevent.net/random/hicolour-gif-in-ie6.png
17:01 < pabs3> seems to work then
17:01 < dooglus> I don't know enough about how wine works to know whether that means it really works in IE6, but I would think it does
17:01 < dooglus> it doesn't work in the 2 ways I most commonly view images: GNU Emacs and gqview
17:16 -!- pabs3 [n=pabs@dsl-220-235-163-152.wa.westnet.com.au] has quit ["zzz"]
18:32 -!- wildhostile [n=chatzill@ALamentin-104-1-41-164.w80-8.abo.wanadoo.fr] has quit ["ChatZilla 0.9.67+ [SeaMonkey 1.0.8/2007030119]"]
18:46 < dooglus> http://www.youtube.com/watch?v=u46eaeAfeqw
19:10 -!- omry__ [n=omry@l192-115-19-216.broadband.actcom.net.il] has joined #synfig
19:13 -!- xerakko [n=xerakko@98.Red-83-61-143.dynamicIP.rima-tde.net] has joined #synfig
19:18 -!- omry [n=omry@l192-115-19-250.broadband.actcom.net.il] has joined #synfig
19:26 -!- omry_ [n=omry@l192-117-123-181.broadband.actcom.net.il] has quit [Connection timed out]
19:31 < zloc> dooglus, did you make that youtube animation?
19:34 -!- omry__ [n=omry@l192-115-19-216.broadband.actcom.net.il] has quit [Read error: 110 (Connection timed out)]
19:37 -!- omry_ [n=omry@l192-114-45-195.broadband.actcom.net.il] has joined #synfig
19:53 -!- omry [n=omry@l192-115-19-250.broadband.actcom.net.il] has quit [Read error: 110 (Connection timed out)]
20:01 -!- omry__ [n=omry@l192-117-123-123.broadband.actcom.net.il] has joined #synfig
20:16 -!- omry [n=omry@l192-114-45-236.broadband.actcom.net.il] has joined #synfig
20:17 -!- omry_ [n=omry@l192-114-45-195.broadband.actcom.net.il] has quit [Read error: 110 (Connection timed out)]
20:22 -!- omry__ [n=omry@l192-117-123-123.broadband.actcom.net.il] has quit [Read error: 60 (Operation timed out)]
20:26 -!- igli [n=igli@82.152.221.75] has joined #synfig
20:54 -!- omry [n=omry@l192-114-45-236.broadband.actcom.net.il] has quit [Read error: 110 (Connection timed out)]
20:55 -!- omry [n=omry@l192-115-19-165.broadband.actcom.net.il] has joined #synfig
22:13 -!- omry [n=omry@l192-115-19-165.broadband.actcom.net.il] has quit [Read error: 110 (Connection timed out)]
22:14 -!- omry [n=omry@l85-130-137-160.broadband.actcom.net.il] has joined #synfig
22:37 < dooglus> zloc: which one?
22:37 < dooglus> zloc: oh, the one I pasted the URL for?  no way!
22:37 < dooglus> that looks like it would have taken so long to do
22:40 < dooglus> I made this one - it's about the limit of my artistic abilities :)
22:40 < dooglus> http://youtube.com/watch?v=djk4B_GiN-Q
22:52 -!- xerakko [n=xerakko@debian/developer/xerakko] has quit ["Me'n vaig"]
22:52 -!- xerakko [n=xerakko@98.Red-83-61-143.dynamicIP.rima-tde.net] has joined #synfig
23:28 -!- omry [n=omry@l85-130-137-160.broadband.actcom.net.il] has quit [Read error: 104 (Connection reset by peer)]
23:29 -!- omry [n=omry@l85-130-137-160.broadband.actcom.net.il] has joined #synfig
23:50 -!- igli [n=igli@82.152.221.75] has quit [Read error: 110 (Connection timed out)]
--- Log closed Thu Mar 22 00:00:02 2007