#jogamp @ irc.freenode.net - 20140821 12:31:31 (UTC)
20140821 12:31:31 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20140821050530.html
20140821 12:31:31 -NickServ- This nickname is registered. Please choose a different nickname, or identify via /msg NickServ identify <password>.
20140821 12:31:37 * jogamp (~jogamp@anon) has joined #jogamp
20140821 12:31:37 * Topic is 'http://jogamp.org | Hacking 3D Graphics, Multimedia and Processing across Devices'
20140821 12:31:37 * Set by rmk0 on 20130116 23:58:04
20140821 12:31:50 -NickServ- You are now identified for jogamp.
20140821 12:39:42 <xranby> rmk0: do the jcanephora API enforce static typechecks and use inside GLSL shaders? example: can the jcanephora API enforce that the user provide the uniforms that the shader uses?
20140821 12:40:17 <xranby> this is things that regular -Djogl.debug.DebugGl do not catch at runtime
20140821 12:43:13 <rmk0> xranby: yeah
20140821 12:43:39 <rmk0> it's not exactly static, though
20140821 12:43:51 <rmk0> i'd have to be generating code to do that
20140821 12:44:23 <rmk0> it checks the types of uniforms at run-time, checks that you've assigned all uniforms (including uniforms that have been optimized out)
20140821 12:44:40 <rmk0> won't let you assign a uniform that never existed in the first place
20140821 12:44:44 <rmk0> same goes for attributes
20140821 12:45:18 <xranby> rmk0: great so basically, YES :) i can point jcanephora as a solution to people that wants a more typesafe way to do GLSL
20140821 12:45:59 <xranby> rmk0: i am trying to find a good way to help this user: https://twitter.com/jan_ekholm/status/501743168816422912
20140821 12:46:41 * rmk0 eyes it
20140821 12:47:26 <rmk0> yeah... it would have caught his error
20140821 12:47:44 <rmk0> http://mvn.io7m.com/io7m-jcanephora/documentation/p5s3.xhtml#st200_p5s3
20140821 12:48:54 <xranby> great i have posted this answer: https://twitter.com/xranby/status/502437023853338624
20140821 12:50:08 <rmk0> as that documentation page says, the worst part of GLSL is that there isn't an API (unless one's been added in the most recent GL versions) that lets you see the *declared* uniforms/attributes of a program
20140821 12:50:21 <rmk0> it'll only give you those that weren't optimized out
20140821 12:50:55 <rmk0> is one of the major reasons that Parasol exists
20140821 12:51:34 <xranby> http://io7m.com/software/parasol -> 404
20140821 12:51:45 <rmk0> hehe, yeah, just spotted that
20140821 12:51:58 <rmk0> will fix the docs, it moved to http://mvn.io7m.com/io7m-jparasol/
20140821 12:53:11 <xranby> typesafe and modular language to build GLSL.. neat
20140821 12:54:38 <rmk0> the type system isn't much more exciting than in GLSL, but the module system is better than a preprocessor
20140821 12:56:20 <rmk0> it does at least get rid of all the overloading that GLSL does
20140821 12:56:31 <rmk0> no idea why people consider overloading a feature... just causes endless problems
20140821 12:58:20 * xranby pokes jogamp
20140821 12:58:56 <rmk0> LWJGL overloads all of the opengl names, doesn't it?
20140821 12:59:00 <rmk0> i remember liking that JOGL didn't
20140821 12:59:29 <rmk0> like... they expose a single glUniform name that takes values of all of the different types as arguments
20140821 13:01:21 <rmk0> i don't blame jan eckholm for not wanting to use my stuff
20140821 13:01:30 <rmk0> once you have something that doesn't fall over at the slightest touch, you don't want to mess with it
20140821 13:01:37 * rmk0 kicks opengl
20140821 13:01:37 * rmk0 kicks glsl
20140821 13:03:54 <rmk0> they've also not reached 1.0.0 so there aren't stable API guarantees
20140821 13:04:17 <rmk0> the APIs for both aren't likely to change now, but... i can't promise that!
20140821 13:04:35 <rmk0> have a ton of code that uses both heavily
20140821 13:04:53 <rmk0> have been documenting the renderer for the past couple of weeks in preparation to get something visible out of the door
20140821 13:06:12 * rmk0 sharpens quill
20140821 13:06:16 <xranby> rmk0: ^_^ i agree that OpenGL and GLSL can become more deterministic, at the same time i praise that OpenGL exist as a vendor neutral way to access the various hardware, and i am happy that work exist like JogAmp and your effort that show working implementations
20140821 13:06:38 <rmk0> yeah... would hate to be working with one proprietary API per platform
20140821 13:07:03 <rmk0> just seriously dislike the lack of type safety and the fact that GLSL is based on a language that was crap 20 years ago
20140821 13:07:30 <rmk0> amusingly, the people most likely to provide a type-safe API and a well designed language these days are microsoft!
20140821 13:07:34 <xranby> rmk0: have you looked at the new LLVM based immediate representations?
20140821 13:07:45 <rmk0> nope
20140821 13:07:51 <rmk0> where are those?
20140821 13:11:17 <xranby> rmk0: khronos unveils the SPIR intermediate language for general purpose GPU now used by OpenCL
20140821 13:11:55 * rmk0 eyes it
20140821 13:12:18 <rmk0> nice!
20140821 13:12:27 <rmk0> hehehe
20140821 13:12:39 <rmk0> main feature: "Is not C source code"
20140821 13:13:06 <rmk0> only a matter of time before this makes it into the opengl API, then
20140821 13:13:45 <xranby> yes that is my thinking as well.. they like to move all opencl features into opengl
20140821 13:13:58 <xranby> for some unknonw reason
20140821 13:14:41 <rmk0> not sure what's keeping mesa from supporting opencl
20140821 13:15:16 <rmk0> it seems like there's all this stuff being developed using llvm, out in the open, and yet i still can't access any sort of opencl functionality even with pretty bleeding edge drivers + kernel
20140821 13:15:34 <rmk0> am anxious to get started with it
20140821 13:20:47 <xranby> imho, opengl should not add features to the spec unless a free software implementation existed ;)
20140821 13:22:00 <rmk0> yep
20140821 15:06:31 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20140821 15:07:33 * jvanek (jvanek@anon) has joined #jogamp
20140821 15:08:02 * jvanek (jvanek@anon) Quit (Client Quit)
20140821 15:56:19 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20140821 16:03:11 * eclesia (~husky@anon) has left #jogamp
20140821 16:43:05 * monsieur_max (~maxime@anon) has joined #jogamp
20140821 20:40:52 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20140821 22:25:18 * xranby (~xranby@anon) Quit (Ping timeout: 260 seconds)
20140821 22:38:38 * xranby (~xranby@anon) has joined #jogamp
20140822 02:34:28 * rmk0 (~rmk0@anon) Quit (Ping timeout: 255 seconds)
20140822 02:39:53 * rmk0 (~rmk0@anon) has joined #jogamp
20140822 04:52:17 * rmk0 (~rmk0@anon) Quit (Ping timeout: 245 seconds)
20140822 04:53:02 * rmk0 (~rmk0@anon) has joined #jogamp
20140822 04:53:02 * rmk0 (~rmk0@anon) Quit (Changing host)
20140822 04:53:02 * rmk0 (~rmk0@anon) has joined #jogamp
20140822 06:01:57 * jvanek (jvanek@anon) has joined #jogamp
20140822 07:22:28 * monsieur_max (~maxime@anon) has joined #jogamp
20140822 12:19:25 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20140822 15:48:30 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20140822 16:43:29 * monsieur_max (~maxime@anon) has joined #jogamp
20140822 18:43:34 * rmk0_ (~rmk0@anon) has joined #jogamp
20140822 18:45:57 * rmk0 (~rmk0@anon) Quit (Ping timeout: 240 seconds)
20140822 19:19:57 * rmk0_ (~rmk0@anon) Quit (Quit: Segmentation fault)
20140822 19:20:23 * rmk0 (~rmk0@anon) has joined #jogamp
20140822 19:20:23 * rmk0 (~rmk0@anon) Quit (Changing host)
20140822 19:20:23 * rmk0 (~rmk0@anon) has joined #jogamp
20140822 21:37:58 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20140822 21:45:16 * zzuegg (~zzuegg@anon) Quit (Quit: Nettalk6 - www.ntalk.de)
20140823 05:05:31 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20140823050531.html