#jogamp @ irc.freenode.net - 20160428 05:06:19 (UTC)

20160428 05:06:19 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20160427050619.html
20160428 05:06:19 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20160428050619.html
20160428 06:14:19 * jvanek (jvanek@anon) has joined #jogamp
20160428 06:30:23 * elect (~GBarbieri@anon) has joined #jogamp
20160428 06:36:08 * elect (~GBarbieri@anon) Quit (Ping timeout: 276 seconds)
20160428 06:43:39 * elect (~GBarbieri@anon) has joined #jogamp
20160428 06:44:18 <elect> hi
20160428 07:07:25 * monsieur_max (~maxime@anon) has joined #jogamp
20160428 07:45:29 * Eclesia (~husky@anon) has joined #jogamp
20160428 07:45:56 <Eclesia> hi
20160428 07:51:54 <elect> hi
20160428 08:31:23 * bigpet (uid25664@anon) has joined #jogamp
20160428 14:48:43 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20160428 15:52:00 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20160428 16:08:10 * Eclesia (~husky@anon) has left #jogamp
20160428 16:22:59 * elect (~GBarbieri@anon) Quit (Ping timeout: 260 seconds)
20160428 16:42:06 * monsieur_max (~maxime@anon) has joined #jogamp
20160428 18:08:41 * Eclesia (~eclesia@anon) has joined #jogamp
20160428 19:25:38 * Eclesia *pom pom pom*
20160428 19:33:20 <rmk0> *BLAM*
20160428 19:44:55 <Eclesia> ^^
20160428 19:45:55 <Eclesia> rmk0: I had a deeper look at openexr, it's going to be painful (at least for me) to implement it. for compressions the docs just say : 'go see the c++ code'
20160428 19:46:04 <rmk0> erk
20160428 19:46:17 <rmk0> yeah... not good
20160428 19:46:30 <rmk0> compression algorithms are typically unreadable
20160428 19:46:58 <Eclesia> well we may not support all types. there are basic types. lossless,yuv422,zip
20160428 19:47:03 <rmk0> it's on my "nice to have but don't actually need" list
20160428 19:50:19 <Eclesia> I really dislike such formats, VP8 codec is the same. they code something that work, buzz about how it's amazing and hyper-efficient, and then they extrapolate a 'specification' with a small quote 'for more info check the code' or 'in case of difference, the code takes precedence'
20160428 19:50:46 <rmk0> yeah
20160428 19:51:07 <rmk0> they're often the same type of programmer that thinks that static types are there to get in the programmer's way
20160428 19:54:13 <Eclesia> I should make a page on my project, the 'Wall of shame' of formats.
20160428 19:57:29 <rmk0> hehe
20160428 19:58:12 <rmk0> https://www.cs.auckland.ac.nz/~pgut001/pubs/pfx.html
20160428 19:58:16 <rmk0> pkcs 12 format
20160428 20:09:38 <Eclesia> hahaha awesome XD
20160428 20:11:47 <Eclesia> I would add use ISO spec references for useless things. this way you make the devs rack a bit :)
20160428 20:12:25 <rmk0> consistently irritated by the private and expensive nature of ISO specifications
20160428 20:17:37 <Eclesia> and small sentences like 'for a complete reference, comply with nato regulation X.X' (which you can't find on the web since it's a first level restricted document), forcing the dev to contact them by mail, explaining why and so on. to finaly get the document which just have a tiny paragraph with a redirection to another spec. (truthful, I experimented that one)
20160428 20:18:06 <rmk0> hehe
20160428 20:21:55 <Eclesia> you can't start to imagine how horrible it is to write a decoder for some old buoy format written in the 80-90th with odd, obscure compressions to win a few bits, a mixed ascii/binary structure, with only spec a pdf of a scanned document where you can still see some handwriting in the margins ...
20160428 20:23:37 <rmk0> ugh
20160428 20:29:49 <bruce-> a bit of this: http://i.imgur.com/0h2nr4s.gifv
20160428 20:38:59 <Eclesia> at least you can take revenge on the one who kicked the ball.
20160428 20:46:34 <bruce-> hm bridj does not seem to understand anything about .framework bundles
20160428 20:54:14 <zubzub> framework bundles?
20160428 20:54:38 <bruce-> yeah, it is how some libraries are distributed on osx
20160428 20:54:45 <zubzub> bridj was a nice project
20160428 20:55:06 <zubzub> until it's scope exploded and it ended up with a bunch of partially supported features and the author gave up
20160428 20:55:58 <bruce-> yeah
20160428 20:56:12 <zubzub> so I wrote my own lib :p
20160428 20:57:27 <zubzub> it doesnt have cpp or objc support like bridj
20160428 20:57:38 <zubzub> but neither does bridj (as in usable support)
20160428 20:57:57 <zubzub> what it does have is arm hf
20160428 20:58:08 <zubzub> and speed :p
20160428 20:58:52 <bruce-> I don't know, from what I have seen thus far it actually looks like writing the wrappers by hand in JNI is less work
20160428 20:59:10 <bruce-> but that's because JNAerator is highly undocumented
20160428 20:59:49 <zubzub> jnaerator is awfull, you have to let it generate uncompiled source code
20160428 20:59:54 <zubzub> open it with your idea
20160428 20:59:58 <zubzub> *ide
20160428 21:00:05 <zubzub> delete all the cruft
20160428 21:00:10 <zubzub> fix all the compilation errors
20160428 21:00:17 <zubzub> and voila you have something that actually works
20160428 21:00:34 <zubzub> but often I'm only interested in specific parts of a native library
20160428 21:00:39 <zubzub> and I just map them as I go in Java
20160428 21:00:45 <bruce-> well, the code it generates actually compiles
20160428 21:00:50 <zubzub> lucky you
20160428 21:00:56 <bruce-> but
20160428 21:01:34 <bruce-> wrapping it manually would actually result in something that is easier to work with than C
20160428 21:01:38 <zubzub> some more shameless promotion: https://github.com/udevbe/jaccall
20160428 21:02:18 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20160428 21:02:50 <bruce-> you currently only support linux, that is the only OS I don't work with
20160428 21:03:16 <Eclesia> zubzub: be shameless +1
20160428 21:03:44 <zubzub> bruce-: :O
20160428 21:03:58 * zubzub puts nose up, walks away
20160428 21:04:22 <zubzub> linux is the only OS you should care about :p
20160428 21:04:28 <zubzub> unless you're interested in making money :p
20160428 21:05:13 <bruce-> there is that
20160428 21:05:22 <zubzub> the reason I don't support other OSes for now is simply lack of will
20160428 21:05:25 <bruce-> I also want to actually use my computer
20160428 21:05:36 <zubzub> it's technically perfectly doable
20160428 21:05:52 <zubzub> I just have to put it full with macro ifdef blocks
20160428 21:06:18 <zubzub> and make some changes to my cmake files
20160428 21:06:25 <zubzub> and add a jni cross compiler for it
20160428 21:07:51 <bruce-> I was so tired of fixing perpetually broken graphics and sound setups that I just had to give up on linux
20160428 21:08:22 <zubzub> how long ago was that?
20160428 21:08:36 <zubzub> and define broken? :)
20160428 21:08:43 <bruce-> 6 years
20160428 21:08:50 <bruce-> broken as in, not functioning
20160428 21:08:54 <zubzub> I admint ifyou want to go beyond stereo on linux, you're pretty screwed
20160428 21:09:00 <bruce-> because some upgrade of the OS broke everything
20160428 21:09:10 <zubzub> aside from that, everything simpley works out of the box 99% of the time
20160428 21:09:11 <bruce-> and audio was just broken by design back then
20160428 21:15:15 <zubzub> stereo sound these days on linux is a complete non issue
20160428 21:22:58 <bruce-> it was a hell, you had so many apis and when one program opened the sound device through API A, then it was impossible to open the sound device through API B
20160428 21:23:17 <bruce-> numerous attempts where made to have apis simulate each other
20160428 21:23:36 <bruce-> were
20160428 21:23:38 <zubzub> those days are long gone
20160428 21:25:44 <Eclesia> good night all ++
20160428 21:25:55 * Eclesia (~eclesia@anon) Quit (Quit: Leaving.)
20160428 23:12:45 * bigpet (uid25664@anon) Quit (Quit: Connection closed for inactivity)
20160429 05:06:20 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20160429050620.html