#jogamp @ irc.freenode.net - 20130319 05:06:39 (UTC)
20130319 05:06:39 -CatOut- Previous @ http://jogamp.org/log/irc/jogamp_20130318050626.html
20130319 05:06:39 -CatOut- This channel is logged @ http://jogamp.org/log/irc/jogamp_20130319050639.html
20130319 06:24:35 * DemoscenePassiv (~Lutsche@anon) has joined #jogamp
20130319 06:31:16 * hharrison (~chatzilla@anon) Quit (Ping timeout: 248 seconds)
20130319 07:27:32 * DemoscenePassiv (~Lutsche@anon) Quit (Ping timeout: 246 seconds)
20130319 08:04:15 * [Mike] (~Mike]@anon) Quit ()
20130319 10:35:36 * odin_ (~Odin@anon) Quit (Ping timeout: 264 seconds)
20130319 10:39:25 * odin_ (~Odin@anon) has joined #jogamp
20130319 16:03:27 * David_ (c13766ea@anon) has joined #jogamp
20130319 16:03:38 <David_> Hello
20130319 16:03:43 <David_> Is anyone online?
20130319 16:13:48 <sgothel> yup .. good morning
20130319 16:16:33 <David_> :)
20130319 16:16:44 <David_> Good morning
20130319 16:17:33 <David_> I'm having an issue in gluegen with intptr_t pointers mapped to longs for a 32 bits compilation
20130319 16:18:06 <sgothel> which is the common denominator .. java code is same for all platforms
20130319 16:18:12 <David_> yes
20130319 16:18:32 <sgothel> pls check the GlueGen package API doc, I dropped a table there ..
20130319 16:18:41 <sgothel> so what is the problem then ?
20130319 16:20:02 <David_> but the problem is that the 64bits long arrays passed from java are mapped directly to intptr_t* so I'm getting null pointers because of the RAW bytes set to 0
20130319 16:20:13 <sgothel> arrays .. I see
20130319 16:20:29 <sgothel> for pointer arrays .. you need to map it to PointerType
20130319 16:20:34 <David_> I bypassed that by copying the long[] to int[] and making the call with that ...
20130319 16:20:43 <David_> wow, what's that?
20130319 16:21:03 <sgothel> this is usually done if using 'void *' AFAIK .. but let me see ..
20130319 16:22:13 <sgothel> that is the general overview .. http://jogamp.org/deployment/jogamp-next/javadoc/gluegen/javadoc/overview-summary.html
20130319 16:23:28 <sgothel> http://jogamp.org/deployment/jogamp-next/javadoc/gluegen/javadoc/com/jogamp/common/nio/PointerBuffer.html
20130319 16:24:46 <sgothel> Now checking where / how we use it, guess I need to drop it into that overview
20130319 16:25:02 * David_ (c13766ea@anon) Quit (Ping timeout: 245 seconds)
20130319 16:25:54 <sgothel> i.e. the PointerBuffer interprets the underlying memory area as either 32bit or 64bit wide, since it has to be of native size - guess that is what you desire. (And now he is gone already :)
20130319 16:27:18 * David_ (c13766ea@anon) has joined #jogamp
20130319 16:28:03 <David_> I'm sorry :P
20130319 16:28:14 <David_> I read what you posted in the logs
20130319 16:28:35 <David_> I will try that, and yes, it looks pretty much the solution I need !
20130319 16:30:09 <sgothel> size_t and intptr_t is always a PointerBuffer since size is arch dependent
20130319 16:31:27 <David_> but why am I getting them generated as LongBuffer ?
20130319 16:31:28 <sgothel> if: pointer or array annotation is being used, i.e. size_t *, size_t[]
20130319 16:34:53 <David_> I have something like this in my .h:
20130319 16:34:58 <David_> typedef intptr_t PlTerm;
20130319 16:35:04 <David_> int Pl_Query_Call(int func, int arity, PlTerm *arg_adr);
20130319 16:35:16 <David_> and the PlTerm *, which is an intptr_t * is mapped to a LongBuffer
20130319 16:36:09 <sgothel> Lets see how GlueGen's own unit test makes a PointerType ..
20130319 16:43:25 <sgothel> hmm .. any <type>* should be PointerBuffer .. checking
20130319 16:43:55 <sgothel> uh .. meaning pointer type arrays
20130319 16:51:14 <sgothel> confirmed ..
20130319 16:52:11 <David_> ok
20130319 16:54:03 <David_> so I have to add the long[] to int[] before each call to the native method? or is there any configuration that will help me with the 32 bits scenario?
20130319 16:55:01 <sgothel> how about typedef void * PlTerm; ?
20130319 16:57:04 <David_> mmm... but I think that would create weird code for functions such as: void Pl_Write_Simple(PlTerm term); but I guess I could live with it? :P
20130319 16:57:18 <sgothel> works here - only problem: you cannot to PiTerm ** lala
20130319 16:57:43 <sgothel> I consider this a bug in GlueGen .. and need to fix it I guess, i.e. consider intptr_t a 'void *'
20130319 16:58:27 * DemoscenePassiv (~Lutsche@anon) has joined #jogamp
20130319 16:58:49 <sgothel> which works for a single intptr_t .. (case included in JavaEmitter) but not 'intptr_t *' .. have to check why it isn't picked up correctly
20130319 17:03:51 <sgothel> I see , the bug is regarding the typedef itself :)
20130319 17:04:00 <sgothel> i.e. w/o typedef it works .. so have to fix it
20130319 17:04:00 <David_> oh
20130319 17:04:48 <David_> I got NativeSizeBuffer in the PlTerm* and Buffer for the PlTerm, I think is not that bad :)
20130319 17:05:33 <David_> Hopefully that will work in 32 and 64 bits versions
20130319 17:06:53 <David_> Thank you, I will work with this !
20130319 17:07:02 <sgothel> What is 'NativeSizeBuffer' ?
20130319 17:07:20 <David_> mmm... that's a gluegen class
20130319 17:07:33 <sgothel> not the GlueGen I use here ..
20130319 17:07:34 <David_> Hardware independent container for native size_t arrays.
20130319 17:07:38 <David_> crap !
20130319 17:08:01 <sgothel> JogAmp's GlueGen has no NativeSizeBuffer
20130319 17:08:34 <David_> MAybe Jogl ?
20130319 17:08:39 <sgothel> 'grep -RI NativeSizeBuffer src' is empty
20130319 17:09:01 <sgothel> nope .. and must be in GlueGen
20130319 17:09:16 <sgothel> which GlueGen do you use ?
20130319 17:09:19 <David_> now I'm not sure which version I'd downloaded
20130319 17:09:21 <sgothel> which repo ?
20130319 17:10:01 <David_> git://github.com/mbien/gluegen.git
20130319 17:10:23 <sgothel> ah well, thats mbien's branch incompatible w/ ours
20130319 17:10:32 <David_> CRAP !
20130319 17:10:39 <David_> which is yours?
20130319 17:10:55 <sgothel> indeed -> http://jogamp.org .. and see git repo link
20130319 17:11:18 <David_> oh ok
20130319 17:11:24 <David_> I'll give it a look !
20130319 17:11:27 <David_> thank you!
20130319 17:11:38 * David_ (c13766ea@anon) Quit (Quit: Page closed)
20130319 17:38:12 * [Mike] (~Mike]@anon) has joined #jogamp
20130319 20:34:59 -RichiH- [Global Notice] Hi all. PDPC, freenode’s parent organisation, has been dissolved. Details can be found at http://blog.freenode.net/ with a static copy at http://planet.freenode.net/
20130319 20:40:51 -mrmist- [Global Notice] As a P.S. to the last global, no, we're not dying or going away. If the blog is down, you can read details at https://plus.google.com/b/104326727082310562426/104326727082310562426/posts/CMW4Gst657v thanks for flying freenode!
20130319 23:12:04 * ___m___ (~Mike]@anon) has joined #jogamp
20130319 23:12:05 * [Mike] (~Mike]@anon) Quit (Read error: Connection reset by peer)
20130319 23:28:15 * DemoscenePassiv (~Lutsche@anon) Quit (Ping timeout: 256 seconds)
20130320 01:18:45 <sgothel> Thanks to Mike pointing at our pointer-size array behavior: http://jogamp.org/git/?p=gluegen.git;a=commit;h=1610dbb4e4d1a1ecf31eb837a47e0f5be04afdac
20130320 01:19:01 <sgothel> this also fixes one NV GL extension mapping :)
20130320 01:19:49 <sgothel> public void glVDPAUMapSurfacesNV(int numSurfaces, LongBuffer surfaces); -> public void glVDPAUMapSurfacesNV(int numSurfaces, PointerBuffer surfaces);
20130320 01:21:07 <sgothel> since the type 'GLvdpauSurfaceNV' is an anonymous pointer-size type
20130320 02:45:13 * ___m___ (~Mike]@anon) Quit ()
20130320 03:13:07 * [Mike] (~Mike]@anon) has joined #jogamp
20130320 03:26:36 * hharrison (~chatzilla@anon) has joined #jogamp
20130320 03:30:21 <hharrison> Java3D 1.6.0-pre7 released, for those that care :-)
20130320 03:42:14 <sgothel> sweet
20130320 03:43:01 <sgothel> merged a few things .. added more PNG tests and cleanup..
20130320 03:57:00 <hharrison> I emailed 7 people that have been testing releases...still shocked to have actual users
20130320 03:58:10 <hharrison> You have any stats for how many times -pre6 was downloaded?
20130320 03:58:38 <sgothel> hmm .. the log/stats doesn't list files properly .. hmm
20130320 03:59:04 <hharrison> oh well....at least 7 :-)
20130320 03:59:34 <sgothel> http://jogamp.org/log/stats/jogamp.org.201303.html
20130320 03:59:59 <sgothel> probably a config error .. of awstats
20130320 04:03:22 <hharrison> Holy crap I didn't realize jogamp was serving that kind of bandwidth per month
20130320 04:04:00 <sgothel> http://jogamp.org/log/stats/jogamp.org.201212.html last year complete
20130320 04:04:50 <sgothel> quite .. popular, however, take bandwidth w/ a grain of salt - some leachers dl a lot
20130320 04:05:12 <sgothel> but the 'unique visitors' value is quite nice
20130320 04:05:15 <hharrison> people pointing a webstart link at your server instead of self-hosting
20130320 04:05:23 <sgothel> maybe
20130320 04:05:59 <hharrison> I might have an interesting demo ready in time for siggraph
20130320 04:06:19 <sgothel> sounds great!
20130320 04:06:35 <hharrison> Our application @work is starting to have good luck importing content direct from autocad files
20130320 04:06:47 <sgothel> nice nice
20130320 04:07:13 <sgothel> I guess some reverse engineering .. even though specs are open .. or maybe not
20130320 04:07:16 <hharrison> I'll see what kind of release/agreement I can get to show off some content....we're talking trains right down to individual bolts
20130320 04:07:42 <hharrison> autocad->3dmax->collada
20130320 04:07:55 <sgothel> a nice train game, routing .. very famous for kids here
20130320 04:08:43 <sgothel> since we gain more traction .. funny bug messages appear .. uh, need volunteers to screen them :)
20130320 04:09:00 <sgothel> like Bug 706 .. not working at all :)
20130320 04:09:26 <sgothel> are you familiar w/ quaternions ? -> bug 703
20130320 04:09:40 <sgothel> should have a weekly bug-walk
20130320 04:10:11 <hharrison> When it comes to math, that _is_ something I can help with
20130320 04:10:38 <sgothel> .. I remember, cool, thx for looking at 703 then
20130320 04:11:07 <sgothel> could have a quaternion / pmvmatrix w/ GLSL demo .. hmm
20130320 04:11:09 <hharrison> I'll do a quick compare to the slerp we have in our quarternion class
20130320 04:11:55 <hharrison> https://github.com/AusencoSimulation/JaamSim/blob/master/src/main/java/com/jaamsim/math/Quaternion.java
20130320 04:12:24 <sgothel> right .. that nice treasure of yours
20130320 04:12:52 <sgothel> if you like to migrate it .. maybe w/ such a demo or unit test - awesome
20130320 04:13:23 <sgothel> (-> GPL3 it is .. so you better do that :)
20130320 04:18:45 <hharrison> Yeah, I'll handle this one, just wanted to have a link there in case anyone wanted to play along at home :-)
20130320 04:21:07 <sgothel> perfect, thx - reusing lerp in slerp looks nice (all those names sound like a soft drink :)
20130320 04:27:21 <hharrison> Actually, from first glance, it just looks like a simple typo
20130320 04:48:08 <hharrison> Hmmm, actually, I'll have to look a little closer, it's a slightly different implementation (but I think the same)
20130320 05:05:53 -CatOut- Continue @ http://jogamp.org/log/irc/jogamp_20130320050553.html