#jogamp @ irc.freenode.net - 20151020 05:05:18 (UTC)


20151020 05:05:18 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20151019050518.html
20151020 05:05:18 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20151020050518.html
20151020 05:47:59 * jvanek (jvanek@anon) has joined #jogamp
20151020 06:06:13 * netlike (~netlike@anon) has joined #jogamp
20151020 06:08:56 * netlike is now known as netlike_
20151020 06:35:24 * elect (~GBarbieri@anon) has joined #jogamp
20151020 07:28:56 * monsieur_max (~maxime@anon) has joined #jogamp
20151020 07:44:30 <zubzub> sgothel: yeah I'm going to mash something up whenever I find the time for it
20151020 07:45:04 <zubzub> still thinking where I should put the stub headers
20151020 07:45:13 <zubzub> as they should not be included at runtime
20151020 08:35:08 * eclesia (~husky@anon) has joined #jogamp
20151020 08:35:13 <eclesia> o/
20151020 09:17:33 * packet0 (~packet0@anon) has joined #jogamp
20151020 09:18:50 * packet0 (~packet0@anon) Quit (Remote host closed the connection)
20151020 09:57:10 <zubzub> sgothel: https://github.com/Zubnix/gluegen-cta-processor
20151020 09:57:16 <zubzub> not tested, quick poc
20151020 09:58:05 <zubzub> should be that much more work to get it fully working
20151020 09:58:10 <zubzub> *should not be
20151020 10:00:21 <sgothel> @zuzub: pls make it inside GlueGen, i.e. extend the current annotation processor
20151020 10:00:39 <sgothel> then - you tell me what the impact of your code is, i.e. possible regression, etc
20151020 10:00:54 <sgothel> in the pull req .. best, best w/ a test
20151020 10:01:12 <sgothel> I showed you one unit test already .. copy .. enhance ..
20151020 10:02:00 <sgothel> while your work is nice as-is, I guess, you understand that I need you as a maintainer of that gluegen enhancement, or at least a very good submission
20151020 10:02:13 <sgothel> (not a separate project)
20151020 10:02:16 <zubzub> yup
20151020 10:02:22 <sgothel> thank you
20151020 10:02:34 <sgothel> IMHO extending our current annotation thingy should not be that hard
20151020 10:03:27 <zubzub> probably, you could also add another one I think
20151020 10:03:42 <sgothel> ah .. maybe, you choose the best technique
20151020 10:03:56 <sgothel> less is more though :)
20151020 10:05:09 <zubzub> Architecture wise I think a seperate processor is the cleanest
20151020 10:05:18 <zubzub> with least amount of risk
20151020 10:05:34 <zubzub> as no existing code has to change to make it work
20151020 10:05:37 <zubzub> afaik
20151020 10:05:50 <zubzub> I'll investigate this evening
20151020 10:06:09 <zubzub> (I was blocked @ work so I quickly wrote the processor ^)
20151020 10:07:00 * netlike_ (~netlike@anon) Quit (Quit: Leaving)
20151020 10:08:10 <sgothel> sure .. do it as you see it best
20151020 10:08:35 <sgothel> we here .. still fighting w/ cold .. and fever, quite tiring
20151020 10:09:23 <sgothel> yesterday son was Ok'ish already .. w/o fever, now back to higher temps a bit .. will see
20151020 12:09:29 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20151020 12:09:51 * monsieur_max (~maxime@anon) has joined #jogamp
20151020 12:40:56 <zubzub> sgothel: do you have any preferences where you want the annotation and processor to be put?
20151020 12:41:16 <zubzub> I see for the the struct it's all in structgen
20151020 12:41:33 <zubzub> but creating a package gluegengen sounds a bit silly
20151020 12:42:23 <sgothel> right
20151020 12:42:52 <sgothel> so just in com.jogamp.gluegen ?
20151020 12:43:07 <sgothel> structgen package was not my choice ..
20151020 12:43:17 <sgothel> we could move all into com.jogamp.gluegen.annotation ?
20151020 12:43:28 <sgothel> Note: API change is fine now
20151020 12:43:44 <zubzub> ok
20151020 12:44:03 <sgothel> then .. also move the unit tests :)
20151020 12:49:23 <zubzub> ok done
20151020 12:49:29 <zubzub> any preferences for an annotatoin name?
20151020 12:49:33 <zubzub> @GlueGenerated
20151020 12:49:34 <zubzub> ?
20151020 12:50:05 <eclesia> @StickyStuff
20151020 12:50:23 <zubzub> it's source retentation only, so it doesn't stick ;)
20151020 12:50:26 <zubzub> *retention
20151020 12:50:52 <monsieur_max> might interest people here : https://letsencrypt.org/
20151020 12:51:12 <sgothel> @GlueGen
20151020 12:51:17 <sgothel> IMHO ..
20151020 12:53:58 <zubzub> k
20151020 12:55:24 <sgothel> @MMax: interesting .. nice nice, I guess
20151020 13:09:33 <zubzub> https://github.com/Zubnix/gluegen/commit/8bcec34b495ff63399ec732b8501665361821034
20151020 13:09:41 <zubzub> for some reason intellij decided to remove trailing spaces
20151020 13:09:46 <zubzub> (in build.xml)
20151020 13:10:04 <zubzub> so the diff there is a bit polluted
20151020 13:10:38 <zubzub> the real change is on line 768
20151020 13:25:01 <zubzub> BUILD FAILED
20151020 13:25:02 <zubzub> Target "test" does not exist in the project "GlueGen".
20151020 13:25:03 <zubzub> :/
20151020 13:26:31 <zubzub> looks like an error in the docs
20151020 13:26:34 <zubzub> should be junit.run
20151020 13:37:16 <sgothel> processor ?
20151020 13:37:35 <sgothel> isn't that ambiguous and we chose annotation ?
20151020 13:38:33 <sgothel> you use annotation somewhere .. hmm
20151020 13:38:39 <sgothel> pls processor -> annotation
20151020 13:39:08 <zubzub> put the processor in the annotation package?
20151020 13:39:17 <sgothel> err ..
20151020 13:39:22 <zubzub> or what are you saying?
20151020 13:39:29 <sgothel> I thought 'annotation' package is more than enough :)
20151020 13:40:24 <zubzub> so move everything to a single annotation package?
20151020 13:40:32 <sgothel> yes sir
20151020 13:40:49 <sgothel> structgen + gluegen annotation processors :)
20151020 13:41:17 <sgothel> btw .. we do the cleanup w/ IDE as well, just not the build.xml yet :)
20151020 13:41:47 <sgothel> since I view diff w/o whitespace .. all good
20151020 13:42:21 <sgothel> .. and the test for the new @GlueGen annotation ?
20151020 13:43:17 <sgothel> using annotation as common package makes things more easy on the eyes and brain .. as well
20151020 13:43:27 <zubzub> sgothel: not finished yet :)
20151020 13:43:51 <zubzub> sure sgothel I don't care :0
20151020 13:44:00 <sgothel> good good
20151020 13:44:02 <zubzub> *:)
20151020 13:44:16 <zubzub> all I want is a simple way to integrate gluegen in my build
20151020 13:44:46 <sgothel> I understood
20151020 13:45:08 <sgothel> so then we have 'jogamp.gluegen.annotation.debug' for all ..
20151020 13:45:30 <sgothel> which can be placed in a common static class for example
20151020 13:45:42 <sgothel> i.e. all the statics ..
20151020 13:46:26 <sgothel> 'annotation.output' .. etc
20151020 13:47:02 <sgothel> would be good if all our annotation processors will follow this ..
20151020 13:47:18 <sgothel> if applicable .. ofc
20151020 13:47:55 <sgothel> and when you are done, maybe you can rebase and make a clean pull req. :)
20151020 13:48:20 <zubzub> yup yup, that's the plan
20151020 13:48:22 <sgothel> one for whitespace cleanup will give you bonus points :)
20151020 13:49:37 <sgothel> for a real world test, maybe you can use our other gluegen unit test to be produced w/ new annotation? com.jogamp.gluegen.test.junit.generation
20151020 13:49:53 <sgothel> and place it in a subdir, i.e. com.jogamp.gluegen.test.junit.generation.annotated
20151020 13:50:03 <sgothel> to be tested in parallel w/ the static generation
20151020 13:50:09 <sgothel> something like that
20151020 13:50:59 <sgothel> Test1p2ProcAddressEmitter <- to be tested via .. one more abstraction needed then
20151020 13:51:14 <sgothel> so you can reuse existing tests
20151020 14:02:49 <zubzub> sgothel: I plan on using it in my own projects for a real world test
20151020 14:02:58 <zubzub> something about dogfood :)
20151020 14:03:39 <zubzub> but I will keep your unit test pointers in mind
20151020 14:04:11 <zubzub> I have to write those anyway
20151020 14:04:12 <sgothel> 'eat your own dogfood' :)
20151020 14:04:19 <zubzub> yeah
20151020 14:04:41 <sgothel> well, if the annotation can work w/ those tests - you can feel save :)
20151020 14:05:17 <zubzub> yup
20151020 14:16:46 * elect (~GBarbieri@anon) Quit (Ping timeout: 240 seconds)
20151020 14:27:09 <zubzub> sgothel: how do I test the generator? ie I don't understand how the 'original' gluegen is invoked in those tests
20151020 14:27:55 <zubzub> or is it invoked through ant?
20151020 14:28:20 <sgothel> yes see build-test.xml
20151020 14:28:25 <zubzub> ic
20151020 14:29:13 <sgothel> so a Test2p2ProcAddressEmitterByAnnotation would be something :)
20151020 14:29:14 <zubzub> the cstruct annotation tests are currently not in the generator package
20151020 14:29:27 <zubzub> *generation
20151020 14:29:41 <sgothel> hu?
20151020 14:30:00 <zubzub> they never were
20151020 14:30:17 <zubzub> they are in a seperate package next to the generation package
20151020 14:30:22 <sgothel> unit tests are in its own jar file ..
20151020 14:30:24 <zubzub> yes
20151020 14:30:25 <sgothel> jaja
20151020 14:30:43 <zubzub> what I mean is that they are not in ...junit.generation.annotation
20151020 14:30:44 <zubzub> but
20151020 14:30:48 <zubzub> ...junit.annotation
20151020 14:30:54 <sgothel> if you like .. clean it up
20151020 14:30:57 <zubzub> yeah
20151020 14:31:02 <zubzub> would be more consistent
20151020 14:31:31 <sgothel> great .. you know .. all this evolutionary enhancements .. needs a cleanup w/ fresh eyes sometimes :)
20151020 14:52:23 <sgothel> OSX issues again .. headaches :-/
20151020 15:04:20 <zubzub> ooh the cstruct annotation processor already calls gluegen directly, it just uses a cstruct emitter
20151020 15:04:29 <zubzub> so I can just abstract that out
20151020 15:04:33 <zubzub> reuse all the things
20151020 15:04:51 <zubzub> and let the user pass any emitter it wants
20151020 15:05:37 <zubzub> I think :/
20151020 15:06:43 <zubzub> hmm it also passes in it's own cfg options, so I need to seperate those out too
20151020 15:36:15 <zubzub> I'm a bit puzzled at the added value of the cstruct annotation processor (not I will remove it)
20151020 15:36:45 <zubzub> all it does is add struct cfg properties
20151020 15:36:47 <zubzub> that's it
20151020 15:52:00 <sgothel> sounds good, i.e. we may can mark CStruct obsolete then .. all using @GlueGen
20151020 15:52:27 <sgothel> ofc .. a cfg file may be needed, i.e. option/arg must exist
20151020 15:52:50 <sgothel> less is more :)
20151020 15:52:57 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20151020 16:02:50 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20151020 16:04:01 * eclesia (~husky@anon) has left #jogamp
20151020 16:32:25 * monsieur_max (~maxime@anon) has joined #jogamp
20151020 17:14:06 * elect (~elect@anon) has joined #jogamp
20151020 17:23:26 * Eclesia (~eclesia@anon) has joined #jogamp
20151020 17:27:42 * elect (~elect@anon) Quit (Remote host closed the connection)
20151020 17:30:42 <Eclesia> sgothel: hi, I found something a little strange, maybe because of a misunderstanding. NEWT can generate several event types. here I have an issue with the 'CLICKED','PRESS','RELEASE' type. I have a widget which respond on PRESS events, and another one on 'CLICKED'. the normal flow of events send by newt is PRESS -> RELEASE -> CLICKED. my widget which do some work on PRESS takes a bit of time which cause the RELEASE and CLICKED events
20151020 17:30:56 <Eclesia> But in this case the CLICKED event isn't send anymore.
20151020 17:31:19 <Eclesia> so I suspect the CLICKED is send only when the time between PRESS and RELEASE is 'short
20151020 17:31:27 <Eclesia> am I right ?
20151020 17:37:04 * elect (~elect@anon) has joined #jogamp
20151020 17:44:12 <sgothel> I thought we don't have any more CLICKED .. ?
20151020 17:44:41 <sgothel> oh still there ..
20151020 17:44:53 <Eclesia> ^^
20151020 17:45:57 <sgothel> ah .. that was key press release _typed_ .. hehe, memories
20151020 17:45:58 <Eclesia> sgothel: it wil be removed ? should I ignore this kind of event ?
20151020 17:46:15 <elect> sven, do you want to use api independent enums or stick completely with the gl ones?
20151020 17:46:58 <sgothel> EVENT_MOUSE_CLICKED: is synthesized if within when - pState0.lastButtonPressTime < MouseEvent.getClickTimeout()
20151020 17:47:15 <sgothel> 300ms hardcoded ..
20151020 17:47:51 <sgothel> @elect: for the GL stuff, we use the existing GL enums if possible, i.e. 'static int'
20151020 17:48:05 <elect> ok
20151020 17:48:07 <sgothel> for the Nativewindow .. abstract image stuff, we have our own
20151020 17:48:30 <Eclesia> sgothel: that include my processing time of the event. so the real click might have take only a few ms.
20151020 17:48:38 <elect> shall we create a new branch?
20151020 17:48:38 <sgothel> i.e. short and precise .. no need to force a GL abstraction if useless
20151020 17:49:03 <sgothel> @elect: no - you may create a working branch in your repo for me to pull later
20151020 17:49:24 <elect> ok
20151020 17:49:25 <sgothel> bit we don't create an official 'feature branch' .. waste of time
20151020 17:49:45 <sgothel> @elect: you got my elaborations from yesterday here?
20151020 17:49:49 <elect> sure
20151020 17:49:53 <sgothel> great
20151020 17:50:11 <sgothel> @Eclesia .. we can refine that ..
20151020 17:50:14 <elect> but I'd like asking some clarifications
20151020 17:50:37 <sgothel> shoot quick, family is waiting - or better tomorrow
20151020 17:50:42 <elect> go
20151020 17:50:44 <elect> dont worry
20151020 17:50:50 <Eclesia> go, family first :p
20151020 17:51:13 <sgothel> for now: I would think .. focus on the GL part first w/ results if possible - tomorrow I will be yours :)
20151020 17:51:21 <elect> ^^
20151020 17:51:32 <sgothel> @Eclesia: you can make a proposal to configure that (pull req.)
20151020 17:52:06 * Eclesia is still not sure on how to define a 'click'
20151020 18:04:56 <zubzub> 17:52 <+sgothel> sounds good, i.e. we may can mark CStruct obsolete then .. all using @GlueGen
20151020 18:04:59 <zubzub> 17:52 <+sgothel> ofc .. a cfg file may be needed, i.e. option/arg must exist
20151020 18:05:01 <zubzub> 17:52 <+sgothel> less is more :)
20151020 18:05:04 <zubzub> awesome
20151020 18:05:07 <zubzub> I had the same idea :)
20151020 18:47:37 * elect (~elect@anon) Quit (Ping timeout: 256 seconds)
20151020 19:56:12 * elect (~elect@anon) has joined #jogamp
20151020 20:41:49 * elect (~elect@anon) Quit (Ping timeout: 252 seconds)
20151020 20:50:17 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20151020 23:16:47 * xranby_ (~familjen@anon) Quit (Ping timeout: 256 seconds)
20151020 23:20:28 * Eclesia (~eclesia@anon) Quit (Quit: Leaving.)
20151020 23:40:18 * xranby_ (~familjen@anon) has joined #jogamp
20151021 05:05:18 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20151021050518.html