#jogamp @ irc.freenode.net - 20131130 05:05:55 (UTC)
20131130 05:05:55 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20131129050555.html
20131130 05:05:55 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20131130050555.html
20131130 07:56:18 * kermyt (~kermyt@anon) Quit (Ping timeout: 252 seconds)
20131130 08:00:43 * kermyt (~kermyt@anon) has joined #jogamp
20131130 13:33:38 * trigger (~trigger@anon) has joined #jogamp
20131130 14:43:42 * trigger (~trigger@anon) Quit (Read error: Operation timed out)
20131130 15:27:30 * trigger (~trigger@anon) has joined #jogamp
20131130 15:57:45 * trigger (~trigger@anon) Quit (Remote host closed the connection)
20131130 17:59:39 * kermyt (~kermyt@anon) Quit (Ping timeout: 252 seconds)
20131130 18:02:10 * kermyt (~kermyt@anon) has joined #jogamp
20131130 18:54:11 * xranby_ac100 (~xranby@anon) has joined #jogamp
20131130 20:28:37 * hharrison (~chatzilla@anon) has joined #jogamp
20131130 20:30:19 <hharrison> sigh, kernel bugs
20131130 20:31:17 <hharrison> Anybody have good recommendations for laptops with well-supported open drivers and OpenGL 3.2 capable?
20131130 21:13:28 * kermyt (~kermyt@anon) Quit (Ping timeout: 260 seconds)
20131130 21:21:51 * kermyt (~kermyt@anon) has joined #jogamp
20131130 22:08:15 * sgothel (~sgothel@anon) has joined #jogamp
20131130 22:08:15 * sgothel (~sgothel@anon) Quit (Changing host)
20131130 22:08:15 * sgothel (~sgothel@anon) has joined #jogamp
20131130 22:08:15 * ChanServ sets mode +v sgothel
20131130 22:29:00 * void256 (~void@anon) has joined #jogamp
20131130 23:29:23 <hharrison> sgothel: how's it going, found a patch I had laying around that cuts ~10% of the GL impl classes
20131130 23:30:10 <hharrison> the size of the .class files anyways
20131130 23:30:21 <sgothel> Hi Harvey, good thx - just returned from family trip .. now trying to migrate an SSD .. (trim trim trim)
20131130 23:30:34 <sgothel> sounds great, let me merge ..
20131130 23:31:18 <hharrison> Sure, I'll send the compatible one first, can get a slightly larger savings, but have to modify ProcAddressEmitter in gluegen a bit more
20131130 23:31:51 <hharrison> make it work a bitmore like BuildComposablePipeline and have an entrypoint to emit helper methods at the end of the impl class
20131130 23:31:57 <sgothel> bug 884 .. so Wade is doing it now .. thought yours was working .. hmm, not that you guys duplicate too much work .. however, me grateful when this process got replicated
20131130 23:32:20 <hharrison> Nope, mine isn't working properly opnce I finally got to testing
20131130 23:32:48 <sgothel> oh .. I see .. guess I need a fresh brain for review then .. (gluegen changes) .. but sure, sounds nice
20131130 23:32:56 <hharrison> So his work isn't really duplicative
20131130 23:33:04 * void256_ (~void@anon) has joined #jogamp
20131130 23:34:22 <sgothel> my dreamsetup .. that sdd + slow-hdd w/ raid 1, slow hdd as 'write-mostly' .. (works w/ mdadm), now I need to see how trim works .. if at all w/ such setup.
20131130 23:34:36 * void256 (~void@anon) Quit (Ping timeout: 252 seconds)
20131130 23:34:38 <sgothel> *ssd*
20131130 23:34:43 * void256_ is now known as void256
20131130 23:36:42 <hharrison> sgothel: just fixinf commit message, will push out shortly (gluegen patch)
20131130 23:36:53 <sgothel> great - thx!
20131130 23:37:24 <hharrison> Then I'll have a few questions about how the emitters work, hopefully can finish this up this weekend
20131130 23:39:00 <sgothel> Qs now ? in a few minutes ? or tomorrow ?
20131130 23:39:50 <hharrison> I'll collect my thoughts and maybe open a feature request...tomorrow
20131130 23:40:22 <hharrison> I'm sending a pull for the small patch that works for all cases so far
20131130 23:40:29 <sgothel> in UTC ? I am UTC+1
20131130 23:40:48 <sgothel> nice .. we will ofc diff the results
20131130 23:40:56 <hharrison> The save another 2% patch can wait until I understand the emitter a bit more
20131130 23:41:03 <hharrison> basically:
20131130 23:41:06 <sgothel> sounds safe
20131130 23:41:11 <hharrison> if (__addr == 0)
20131130 23:41:52 <hharrison> throw new Exception("Method \"" + "methodname" + "\" not available");
20131130 23:41:56 <hharrison> becomes:
20131130 23:42:12 <hharrison> throw new GLException(String.format("Method \"%s\" not available", "glAccum"));
20131130 23:42:46 <hharrison> so all the format strings get collapsed, and only the gl method strings remain (but are shared with other error messges too)
20131130 23:42:59 <hharrison> as opposed to a bunch of redundant error strings
20131130 23:43:24 <sgothel> ah .. same pattern as before .. for slow path .. great
20131130 23:43:29 <hharrison> yep
20131130 23:43:50 <hharrison> also will share with the other classes that have just the gl method name as a string
20131130 23:43:59 <hharrison> so effects on the heap are even greater
20131130 23:44:10 <hharrison> ~10k off the GL3esimpl class
20131130 23:46:01 <sgothel> sweet
20131130 23:49:01 <hharrison> So together with me previous patch to the debug classes, a much higher amount of shared constant strings on the heap
20131130 23:49:18 <hharrison> and the heap is way smaller not having all the large error messges in the permgen
20131130 23:49:22 <sgothel> precious heap ..
20131130 23:49:35 <hharrison> OK, patch pushed to my public git repo
20131130 23:49:45 <hharrison> need coordinates?
20131130 23:49:56 <sgothel> great, yup - when I walked through similar cases .. I applied your pattern as well
20131130 23:49:59 <sgothel> nope .. have it
20131130 23:50:05 <hharrison> master branch
20131130 23:50:26 <hharrison> This is the compatibe version as the 'unsupported exception' type can be changed by each user
20131130 23:50:50 <hharrison> But they all must accept a single String argument, so my patch changes nothing in that respect
20131130 23:51:23 <hharrison> What I really want is a way to emit a single error(String methodName) method that throws the exception
20131130 23:51:40 <hharrison> Not open-code the throw new exception bit
20131130 23:51:58 <hharrison> with a unique error method defined for each impl class
20131130 23:52:10 <hharrison> Need to step out for 10 minutes
20131201 00:16:47 <hharrison> back
20131201 00:47:19 * hharrison (~chatzilla@anon) Quit (Ping timeout: 246 seconds)
20131201 03:27:02 * void256 (~void@anon) Quit (Remote host closed the connection)
20131201 05:05:55 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20131201050555.html