#jogamp @ irc.freenode.net - 20151018 05:05:17 (UTC)

20151018 05:05:17 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20151017050517.html
20151018 05:05:17 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20151018050517.html
20151018 05:12:31 * bigpet (uid25664@anon) has joined #jogamp
20151018 06:25:14 <elect> http://mvnrepository.com/artifact/org.jogamp.jogl/jogl-all-main, all the releases report a wrong date
20151018 06:25:58 <elect> btw which is the difference between the first three JOGL ? http://mvnrepository.com/artifact/org.jogamp.jogl
20151018 06:50:17 <elect> anyway, it looks like in the maven repo there are no src or am I wrong?=
20151018 07:01:40 <elect> ah solved
20151018 07:43:07 * elect (~elect@anon) Quit (Ping timeout: 256 seconds)
20151018 07:47:49 <xranby_> http://mvnrepository.com/artifact/org.jogamp.jogl/jogl-all-main <- fun get to know some projects using jogamp that are deployed on maven
20151018 07:48:30 <xranby_> 2.0.2 is still the most popular version used by maven people http://mvnrepository.com/artifact/org.jogamp.jogl/jogl-all-main/2.0.2/usages
20151018 07:48:46 <xranby_> ... that have deployed their projects to maven
20151018 08:51:43 * monsieur_max (~maxime@anon) has joined #jogamp
20151018 08:59:39 * elect (~elect@anon) has joined #jogamp
20151018 09:01:17 <elect> someone successfully used Gradle here with jogl?
20151018 09:01:42 <elect> coz I get crash
20151018 09:01:43 <elect> http://stackoverflow.com/questions/33195461/netbeans-gradle-and-jogl-crash
20151018 09:43:58 <zubzub> xranby_: the name of the project
20151018 09:44:11 <zubzub> it's the same name for a low latency gc
20151018 09:45:28 <zubzub> or am I seeing it wron?
20151018 09:46:12 <zubzub> or is it simply called 'mobile project'
20151018 11:09:38 <xranby_> zubzub: that is correct it will be called `the Mobile Project`
20151018 11:29:20 <xranby_> elect: how about a bugreport?
20151018 11:30:05 <xranby_> since you get it during run
20151018 11:30:26 <xranby_> if you push your gradle script then i can try run it as well
20151018 11:34:58 <xranby_> elect: your stack overlow examplle is incomplete because the gradle script with the run task is not published in the stack overflow or to your git
20151018 11:35:25 <xranby_> you have only published unrelated code
20151018 11:35:55 <xranby_> the settings.gradle s also missing
20151018 11:36:26 <xranby_> add all missing files to your project and then we can look into what is crashing and why
20151018 11:40:55 <xranby_> zubzub: you are testing on the same machine that you managed to crash in libxcb the other day.. right?
20151018 11:41:05 <xranby_> zubzub: using the same test program
20151018 11:42:00 <xranby_> hence we need to make sure you xorg version is ok, and your ubuntu window manager is ok
20151018 11:42:24 <xranby_> and check if there is some sort of usecase that work on ther platforms break this setup
20151018 11:42:33 <xranby_> i doubt it has anything to do with gradle
20151018 12:13:56 * Eclesia (~eclesia@anon) has joined #jogamp
20151018 12:14:16 <Eclesia> hi
20151018 12:15:28 <sgothel> hi
20151018 12:15:43 <sgothel> elect is working on a bad reputation for some tools incl. jogl? :)
20151018 12:18:07 <sgothel> I also read the JGLI thingy .. and I almost concur w/ Julien's assessment, however - maybe the 'sweat and tears' hard work of compiling things from C++ to Java manually can result in a module to be used for JOGL
20151018 12:18:34 <sgothel> @elect: my advise is: don't use your shortcomings as an argument
20151018 12:19:09 <sgothel> (like: can't install, can't compile, can't develop with .. )
20151018 12:25:05 <Eclesia> just for images ? jgli ? nothing else ?
20151018 12:25:37 <sgothel> AFAIK .. yes, however - I see this as a more abstract/general issue here
20151018 12:26:23 <sgothel> i.e. complaining .. whining .. and using lack of success of basic operations as an argument to go other ways etc
20151018 12:26:51 <sgothel> and yes, we did help quite a lot enabling developing w/ JOGL
20151018 12:27:29 * Eclesia has no issues with images+jogl, works like a charm.
20151018 12:27:59 <sgothel> regarding JGLI in particular, to trust some codebase we have to assume that the code has been understood
20151018 12:28:07 <sgothel> otherwise there can be no maintenance
20151018 12:28:20 <sgothel> and a code drop w/o maintenance cannot be accepted
20151018 12:34:19 <xranby_> elect: for the record, i use gradle + jogl + libgdx and that work like a charm
20151018 14:03:41 <elect> xranby_, I'll try the bugreport asap. Which script you mean? I see only settings and build.grade
20151018 14:04:00 <elect> I inserted the settings.gradle in the question btw
20151018 14:35:05 <zubzub> xranby_: no idea what you're talking about :)
20151018 14:35:11 <zubzub> must be someone else that did that
20151018 15:05:06 <elect> sgothel, I think there are a couple of things you quite misunderstood, maybe not on purpouse or maybe because opportunation. Let me repeat some points, if anything is still unclear, I'll be glad to explain. Porting things from C++ to Java was quite easy and fast, I just spent some hidle time about planning to find the best structure which could match Java, but nothing that stopped me for more than some minutes and nothing that can't
20151018 15:05:06 <elect> be changed in the soon future if I find out something better.
20151018 15:05:06 <elect> What was "sweat & tears" was instead trying to implement all the features such as 3d texture structure and having an old and new API enums for defining formats above the current texture implementation of jogl, mainly from 2009 and that in its turn borrows some (small I think) code from another library, devil, probabily even older. Just to be precise, older does not mean necessarily bad, but neither the contrary.
20151018 15:05:11 <elect> I think I never used "can't install" or "can't compile" as a shortcoming argument to start jgli. If wrong, I will be happy to be corrected. I did use the "can't develop" option. But I motivated it. It is unfair blaming the action itself without trying to argue the reasons I bringt for.
20151018 15:05:15 <elect> Let me use a metaphor, it's like we have an old building. And we need building more flat. If conditions like too old, lacking of good energetic insulation, small fundaments and so on are met, it's better to take it down and start from scratch. But we don't want to modify the facade, because the citizien are used to it, it's nice because classic or other reasons. In this case, the facade is taken up, but all the rest will be
20151018 15:05:22 <elect> completely new. This will lead to having a same-but-new-under-the-hood building, that will be totally suitable to hold all the new floors you plan to build it on and will be also good in terms of heating and it will get a still longer live.
20151018 15:05:26 <elect> I am sorry to not be able to macth your skill, but you should take in account that some operations that may be appear basic to you may appear differently to other people, like me.
20151018 15:05:29 <elect> I am sorry but from me it looks absolutely normal (and logic I'd say) choosing another way if that one you are going does not take you anywhere or it looks much harder and messy compared to another one, especially if that seems not having the issues of the first one.
20151018 15:05:33 <elect> I tried starting from jogl, but it didn't work out for me, I didn't complain or whining, I took a look around and decided it was worth to give a try starting another way and it turned out I found it much better. I can't find anything bad in the process itself. Things may appear one way if you look them from outside. But they might also change once you are deep inside.
20151018 15:05:38 <elect> Anyway, regarding the code understanding, gli is pretty straightforward. The advantages of gli is that I don't have to care about getting a logic line in the MS specs. G-truc did the hard work of figure out the format from the fields of the header and this is what I really love.
20151018 15:05:42 <elect> When I started, I wanted to do a very complete lib that could solve once for all a lot of the limitations of the current jogl implementation. This means having everything working (as much as possible), every format, every texture target, type and so on, and not solving just the problem for few formats, maybe hardcoding some shit, and then postposing the problem to the posterity. I did not want that and I refuse to do it, because it
20151018 15:05:47 <elect> is a very bad usage of time in terms of efficiency. Another main advantage of gli over jogl, is that it uses internally an API-independent format (well actually much more, target, type, swizzle, etc), that acts also as a common medium when you need to handle texture from different APIs. This will give us much more flexibility to "build new floors", such as KMG, Vulkan, etc..
20151018 15:05:54 <elect> Anyway, since it seems you, guys, don't like this way and I really dislike, honestly, the idea of upsetting both of you, therefore I will therefore come back on my steps and I will continue working on the current jogl trying to get the best of it based my skill.
20151018 15:07:15 <elect> Eclesia, http://gli.g-truc.net/0.7.0/index.html
20151018 15:10:14 <sgothel> good .. phew :)
20151018 15:10:23 <sgothel> lots of text elect :)
20151018 15:11:09 <sgothel> it is OK to learn other APIs and implementations .. sure, anyway one likes - no problem.
20151018 15:11:35 <sgothel> the only point was .. how can this work being utilized in a way, that we can maintain it and directly benefit
20151018 15:12:11 <sgothel> so while you learned the shortcomings of current Texture class - and we have now 2 bug entries for its re-work ..
20151018 15:12:39 <sgothel> there is room to enhance our stuff. the question here now is, will you do it - or at least help doing so?
20151018 15:12:49 <sgothel> _and_ in a high quality manner w/ unit tests etc
20151018 15:13:01 <Eclesia> elect: my own coment on this part "is that it uses internally an API-independent format" . definitly wrong, the api I see in jgli is 100% opengl oriented, no hope to make a grid or netcdf fit in there.
20151018 15:13:21 <Eclesia> grib*
20151018 15:13:35 <sgothel> I assume the problem Julien saw here, and me as well - is that the promotion and work on the port may not show this path
20151018 15:14:39 <sgothel> so w/ the 'work on JOGL enhancement' is the core issue here ..
20151018 15:15:37 <sgothel> @Eclesia: Yes, we also have 2 separate abstractions of 'pixel rectangle' things, one generic in Nativewindow and the Texture class .. and its internals
20151018 15:15:46 <sgothel> also a thing we may can merge somewhat
20151018 15:21:40 <sgothel> @elect: see it from a different angle. you started w/ this task by a self assigned bug report .. enhancing JOGL. Well, now you ended up w/ an own project of yourself :)
20151018 15:22:25 <sgothel> what I like to see is - whatever task it is you do - to make it perfect once, so we have one new great contributor!
20151018 15:22:56 <sgothel> and yes, we talk about test test test, maintenance and quality
20151018 15:23:41 <sgothel> talking about 'building new floors', one shall stand on a quite stable one in the first place
20151018 15:23:53 <sgothel> .. i.e. to begin with
20151018 15:24:13 <sgothel> and this is credibility, one reason people will choose jogamp
20151018 15:25:46 <sgothel> again: my wish would be that you can put all the knowledge you have learned from JGLI into jogl .. or even make it a part of it (iff it has almost zero additional footprint)
20151018 15:26:40 <sgothel> and this will need you enthusiasm and work power to make it great (tests, maintenance etc)
20151018 15:28:31 <sgothel> another advise: its is also not a valid argument to state 'A is 10 years old, hence shit', but 'A does not cover use cases C and D' is.
20151018 15:28:53 <sgothel> :)
20151018 15:48:13 <elect> yeah, sorry for the wall of text, but I wanted to be sure you got my whole point of view
20151018 15:48:16 <elect> :p
20151018 15:49:20 <elect> porting a lib to java, considered simple as one image lib may be, it means for sure facility also in the maintenance
20151018 15:50:10 <elect> I am available to do it, although I may lack experience to do a work you like
20151018 15:50:48 <elect> Eclesia, the api match more or less 90% opengl, much less opengl es
20151018 15:51:44 <elect> some target does not exist (yet) in opengl, such as TARGET_RECT_ARRAY
20151018 15:51:58 <elect> and a couple of other variables
20151018 15:52:17 <elect> opengl is, of course, the lib it matches at best
20151018 15:52:21 <elect> but no 100%
20151018 15:52:47 <elect> and moreover, having them separately assures flexibility in the future
20151018 15:55:47 <elect> btw, I pointed in advance that "old doesnt means necessarily bad", but it *could* instead mean "unsuitable" for the future and even the present
20151018 15:56:14 <elect> anyway
20151018 15:56:37 <elect> I have half an idea about how to get further from the current implementation
20151018 15:57:00 <elect> I could make a DxgiImage, parallel to the DdsImage
20151018 15:57:12 <elect> and company
20151018 16:01:35 <Eclesia> elect: different image implementation for different formats is a bad idea.
20151018 16:02:25 <elect> I also have the feeling
20151018 16:03:20 <elect> but DdsImage.java is already a 929 lines class
20151018 16:03:35 <elect> I dislike more the idea to have a 2k lines class
20151018 16:03:49 <elect> I dont know
20151018 16:04:03 <Eclesia> it's not a matter of number of lines
20151018 16:04:41 <elect> not until you stay under a threshold
20151018 16:05:05 <sgothel> "I am available to do it, although I may lack experience to do a work you like" <- so, will you do it, or not?
20151018 16:05:12 <elect> I'll try
20151018 16:06:01 <sgothel> see, this is a good independent piece of work, even thought core to JOGL, the requested results are not blocking anybody here right now
20151018 16:06:56 <sgothel> "I'll try' .. is something I don't understand - but well, OK, so we don't expect anything from you and see what will happen.
20151018 16:07:21 <elect> I will try to implement it
20151018 16:07:33 <elect> but I don't know if I will success
20151018 16:07:41 <sgothel> we are talking philosophy now, elect :)
20151018 16:08:09 <sgothel> of course you will succeed if we cooperate - b/c we want you to succeed .. that simple and that hard
20151018 16:08:31 <sgothel> but: no whining allowed, but hard work only :)
20151018 16:08:34 <Eclesia> elect: could you start by a format I don't have ? ^^
20151018 16:09:19 <sgothel> 'whining' something one may does instead of actually doing 'it' :)
20151018 16:09:36 <elect> then let's cooperate
20151018 16:10:05 <sgothel> 'I will try' is like saying 'I don't want'
20151018 16:10:34 <sgothel> some 'tough talk' as learned while playing baskeball :)
20151018 16:10:35 <elect> Eclesia, you tell me https://github.com/elect86/jgli/blob/master/Jgli/src/jgli/Format.java ^^
20151018 16:10:58 <elect> I want to be honest, I see it hard to implement the way I wanted
20151018 16:11:16 <elect> so I will need to go down to several compromises
20151018 16:11:34 <elect> and I am worried this would mean an handicapped lib
20151018 16:11:54 <elect> but I give my availability anyway
20151018 16:12:29 <Eclesia> elect: why not have a look at the way I made it ? already have tenth of image formats available you know
20151018 16:13:08 <sgothel> Texture and its framework .. must not be core to what you do, but can be hooked into it. It would be good however, to enhance all our GL buffer switch/case blocks while validating them (format/data-type)
20151018 16:13:13 <Eclesia> elect: but first you should start by thinking a bit : 'what is an image' (not an opengl image)
20151018 16:13:20 <elect> I took a look, but it looks like you designed it in a way to follow your goals of course
20151018 16:13:53 <Eclesia> my goals ? I don't have any :D coding for fun only
20151018 16:13:57 <sgothel> http://jogamp.org/deployment/archive/master/gluegen_901-joal_627-jogl_1470-jocl_1107/javadoc/jogl/javadoc/com/jogamp/nativewindow/util/PixelRectangle.html
20151018 16:14:04 <elect> you told me something being available also when no x11 is available
20151018 16:14:08 <sgothel> http://jogamp.org/deployment/archive/master/gluegen_901-joal_627-jogl_1470-jocl_1107/javadoc/jogl/javadoc/com/jogamp/nativewindow/util/PixelFormat.html
20151018 16:14:22 <sgothel> ^^ a generic approach
20151018 16:14:43 <sgothel> intended for CPU based ops .. however, quite generic
20151018 16:15:04 <elect> anyway, if you have a list I will be glad to take a look at it
20151018 16:15:07 <Eclesia> generic approach yes but not slower for gl either, since the ByteBuffer is available
20151018 16:16:08 <elect> I meant goals of your lib :p
20151018 16:16:21 <sgothel> thinking about it .. maybe its about 2 representations: CPU + [GL, ..] and the transition between both
20151018 16:16:34 <elect> I can work on that, sure, sgothel
20151018 16:16:49 <Eclesia> elect: well dds,bmp,.. are simplistic structure images, you should have a look at the opposite, complexe structure images such as Dicom,NetCDF and Grib
20151018 16:17:12 <Eclesia> not implementing them, just to have an idea :)
20151018 16:19:14 <Eclesia> dicom are medical images, netcdf are mostly oceanographic images, and grib(or hdf) for meteorology
20151018 16:41:49 <xranby_> elect: can you add the gradle files to your git?
20151018 16:42:58 <xranby_> elect: i want to reproduce the bug using git clone https://github.com/elect86/Jglm ; cd Jglm/Jglm ; gradle run
20151018 16:43:51 <xranby_> but right now i cant do/reproduce what you ask in your stackowerflow because the files are not part of the project
20151018 16:45:59 <xranby_> your crash log is interesting.. but i need to reproduce it, hece i need you to push all the gradle files to the git https://gist.github.com/elect86/727c445c176686bec10e
20151018 16:56:08 <sgothel> ^^ one sideeffect of a release .. hands-on Xerxes is on the hunt :)
20151018 16:56:33 <rmk0> release the xranby
20151018 16:57:08 <sgothel> *unleashed*
20151018 16:57:39 <Eclesia> run for your gradles !
20151018 17:10:07 <xranby_> bring your robes, embroidered insignia patches and we can all have our secret society
20151018 17:10:30 <xranby_> however we will eat your bugs!
20151018 19:07:44 <xranby_> sgothel: http://jogamp.org/log/stats/jogamp.org.2015.html the statistics is broken hmm
20151018 19:07:45 <xranby_> Error: Plugin load for plugin 'hostinfo' failed with return code: Error: Can't locate Net/XWhois.pm in @INC ...
20151018 20:27:17 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20151018 20:33:51 * Eclesia (~eclesia@anon) Quit (Quit: Leaving.)
20151018 21:15:29 <elect> xranby_, yeah sure
20151018 21:15:37 <elect> I have no time now to deal with backup
20151018 21:15:43 <elect> so I put it here https://www.wetransfer.com/downloads/455bc03045396658adeaa77a2199bba020151018211207/0b731bc7818dd152e84794df03e513fb20151018211207/9e858a
20151018 21:15:48 <elect> night! :)
20151018 21:20:15 * elect (~elect@anon) Quit (Ping timeout: 256 seconds)
20151018 23:29:11 * guillaum1 (~gl@anon) has joined #jogamp
20151019 04:16:39 * guillaum1 (~gl@anon) Quit (Ping timeout: 240 seconds)
20151019 04:37:23 * guillaum1 (~gl@anon) has joined #jogamp
20151019 05:05:18 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20151019050518.html