#jogamp @ irc.freenode.net - 20170125 05:06:08 (UTC)

20170125 05:06:08 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20170124050608.html
20170125 05:06:08 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20170125050608.html
20170125 06:59:58 * elect (~elect@anon) has joined #jogamp
20170125 07:08:35 * elect (~elect@anon) Quit (Ping timeout: 240 seconds)
20170125 08:09:21 * elect__ (~elect@anon) has joined #jogamp
20170125 08:14:03 <elect__> hi
20170125 08:15:45 <monsieur_max> hello
20170125 09:56:04 * caelum19 (~Caelum@anon) has joined #jogamp
20170125 11:09:21 * elect__ (~elect@anon) Quit (Read error: Connection reset by peer)
20170125 11:09:39 * elect (~elect@anon) has joined #jogamp
20170125 13:09:48 * elect (~elect@anon) Quit (Read error: Connection reset by peer)
20170125 13:09:59 * elect (~elect@anon) has joined #jogamp
20170125 13:48:54 * elect_ (~elect@anon) has joined #jogamp
20170125 15:57:38 * elect_ (~elect@anon) Quit (Ping timeout: 276 seconds)
20170125 17:32:34 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20170125 17:42:27 * caelum191 (~Caelum@anon) has joined #jogamp
20170125 17:43:22 * elect (~elect@anon) Quit (Ping timeout: 240 seconds)
20170125 17:45:15 * caelum19 (~Caelum@anon) Quit (Ping timeout: 260 seconds)
20170125 17:54:28 * elect_ (~elect@anon) has joined #jogamp
20170125 19:39:17 * elect_ (~elect@anon) Quit (Ping timeout: 276 seconds)
20170125 19:53:24 * monsieur_max (~maxime@anon) has joined #jogamp
20170125 20:33:36 * elect_ (~elect@anon) has joined #jogamp
20170125 20:35:25 * Eclesia (~eclesia@anon) has joined #jogamp
20170125 20:35:47 <Eclesia> ho
20170125 20:36:04 <zubzub> he
20170125 20:39:27 <Eclesia> anything new ?
20170125 20:54:39 <zubzub> not really
20170125 20:57:24 <Eclesia> zubzub: with jacall is it possible to execute native call. I mean if had a byte buffer with native cpu instructions would it be possible ?
20170125 21:06:35 <zubzub> Eclesia: depends, is it possible to do that in C?
20170125 21:07:03 <Eclesia> I think yes, we can find code samples such as : http://stackoverflow.com/questions/27883570/c-run-byte-array-in-memory
20170125 21:08:03 <zubzub> ok so you have an array, get the address of the array, cast it to a function pointer
20170125 21:08:07 <zubzub> should be doable I think
20170125 21:10:52 <Eclesia> a woking one in C : http://stackoverflow.com/questions/9960721/how-to-get-c-code-to-execute-hex-bytecode
20170125 21:11:08 <Eclesia> hmhmhm, that opens many possibilities for me ^^
20170125 21:11:42 <zubzub> you can easlly get the address of a bb in jaccall
20170125 21:11:50 <zubzub> then you need to declare a function pointer interface
20170125 21:12:12 <zubzub> then simply wrap the long (address) in the type of the function pointer
20170125 21:12:17 <zubzub> call it
20170125 21:12:19 <zubzub> profit!
20170125 21:12:41 <Eclesia> groopies!
20170125 21:12:47 <zubzub> https://github.com/udevbe/jaccall#function-pointer-api
20170125 21:14:03 <zubzub> btw that first C example
20170125 21:14:11 <zubzub> the get's the address of an array
20170125 21:14:17 <zubzub> -t
20170125 21:14:25 <zubzub> the address of the array is the array itself
20170125 21:14:33 <zubzub> #lackofCunderstanding
20170125 21:17:25 <Eclesia> I was just looking at x86/arm cpu instructions tutorials, and just wondering if it would be fun to generate bytecode and then execute it. and still remain in java.
20170125 21:18:18 * Eclesia is crazy
20170125 21:20:07 <zubzub> generate bytecode?
20170125 21:20:15 <zubzub> you mean machine code?
20170125 21:20:22 <Eclesia> yes
20170125 21:21:03 * Eclesia will make his own VM ... one day
20170125 21:21:36 <zubzub> afaik it should werk just fine with jaccall
20170125 21:21:39 <zubzub> *work
20170125 21:23:01 <zubzub> PointerMyFuncPointer.wrapFunc(JNI.unrwap(byteBuffer)).$(myArg);
20170125 21:23:04 <zubzub> that's it
20170125 21:24:13 <zubzub> (with MyFuncPointer as an @Functor annotated single method interface)
20170125 21:26:30 <zubzub> if you use Pointer<Byte> instead of a ByteBuffer it becomes simpler
20170125 21:27:11 <zubzub> instructionsPointer.castp(MyFuncPointer.class).$(myArg);
20170125 21:27:35 <Eclesia> zubzub: slow down, I'm still farrrrr from there :)
20170125 21:27:54 <zubzub> giving up already? :p
20170125 21:27:57 <Eclesia> it was only a theorical question
20170125 21:28:30 <zubzub> I think you should do it and become famous
20170125 21:28:44 <Eclesia> zubzub: nope, I have to finish my database implementation first, to store all instruction tables
20170125 21:28:55 <zubzub> then hurry up! :p
20170125 21:29:08 <Eclesia> rewriting everything takes some time
20170125 21:30:07 <Eclesia> and noone helps :'( ... like on jogl
20170125 21:31:54 <zubzub> if you want people to help you need to start from an existing community with basically the same interest, or create a community first before starting
20170125 21:32:20 <zubzub> else it'll remain a pet project, unless it becomes popular
20170125 21:32:27 <Eclesia> the 'public domain' community is really small
20170125 21:32:41 <zubzub> yup
20170125 21:32:55 <zubzub> all of my projects are one man efforts as well
20170125 21:33:33 <zubzub> but that's ok, I think in 2-3 years I might finally dogfood the dream I had about 10y ago
20170125 21:33:56 <zubzub> (my own window manager/compositor/display server)
20170125 21:34:26 <Eclesia> well, my isn't a one man effort, there are over 50 contributors, but I'm the only active developer
20170125 21:34:28 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20170125 21:34:30 <Eclesia> mine*
20170125 21:35:46 <Eclesia> even rmk0 is part of it :)
20170125 21:36:32 <zubzub> :)
20170125 21:36:50 <zubzub> problem with my projects is that they're too complex/exotic I think