#jogamp @ irc.freenode.net - 20151023 05:05:18 (UTC)

20151023 05:05:18 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20151022050518.html
20151023 05:05:18 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20151023050518.html
20151023 06:28:23 * elect (~GBarbieri@anon) has joined #jogamp
20151023 06:33:20 <elect> hi
20151023 07:15:52 * bigpet (uid25664@anon) Quit (Quit: Connection closed for inactivity)
20151023 07:34:51 <zubzub> sgothel: for the api doc, I just c/p from the gluegen manual :)
20151023 07:35:00 <zubzub> (command line invocation args)
20151023 07:49:56 * monsieur_max (~maxime@anon) has joined #jogamp
20151023 07:52:38 <zubzub> sgothel: when I do an ant maven.install
20151023 07:52:43 <zubzub> and then attempt to use gluegen
20151023 07:52:46 <zubzub> I get
20151023 07:53:00 <zubzub> java.lang.NoClassDefFoundError: antlr/RecognitionException
20151023 07:53:19 <zubzub> when looking inside the gluegen jar, it seem it's not present either
20151023 07:54:14 <zubzub> I assume it's simply not part of the dependencies listed in the gluegen maven artifact?
20151023 08:05:16 <zubzub> ok seems gluegen maven artifact depens on an ancient ant version
20151023 08:05:25 <zubzub> 2.7.7(?)
20151023 08:05:42 <zubzub> adding the latest available ant major version from maven resolves it
20151023 08:05:43 <zubzub> 2.7.4
20151023 08:05:54 <zubzub> *major compatible version
20151023 08:28:46 <sgothel> ^^ uh .. so we have a bug w/ latest ant er?
20151023 08:29:25 <sgothel> no the other way around
20151023 08:30:11 <sgothel> re verbose API doc: ha, that is funny. and I thought 'another long text writer' :)
20151023 08:31:01 <sgothel> Jul 2, 2015 - Apache Ant 1.9.6 Released
20151023 08:31:52 <sgothel> oh .. maven and project version differs: headache :)
20151023 08:31:58 <sgothel> ^^ for ant
20151023 08:35:00 <sgothel> http://www.rawstory.com/2015/10/iceland-does-what-the-us-wont-26-top-bankers-sent-to-prison-for-role-in-financial-crisis/ <- *strike*
20151023 08:42:31 <zubzub> ANTLR Parsing Error: line 130:5974: unexpected token: int token name:"int"
20151023 08:42:32 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20151023 08:42:32 <zubzub> :(
20151023 08:42:45 <zubzub> oh well at least the annotation is working and picking up the cfg
20151023 08:42:48 * jvanek (jvanek@anon) has joined #jogamp
20151023 08:42:49 <zubzub> and header
20151023 08:43:51 <sgothel> or are we talking a diff antlr ?
20151023 08:44:00 <sgothel> 'int' for sure is defined in our grammar
20151023 08:44:03 <zubzub> yeah
20151023 08:44:08 <zubzub> that's why I find it strange
20151023 08:44:14 <zubzub> it also not the first 'int' in the header
20151023 08:44:25 <sgothel> err .. we use our local antlr withing gluegen
20151023 08:44:42 <sgothel> yes, it is old - the new one probably won't even parse our grammar
20151023 08:44:49 <sgothel> (antlr version that is)
20151023 08:45:07 <zubzub> the mavinized jar of gluegen does not seem to contain antlr
20151023 08:45:33 <sgothel> right .. a runtime only thing so far
20151023 08:45:40 <zubzub> that's why I added one myself
20151023 08:45:44 <zubzub> probably the wrong one
20151023 08:45:58 <zubzub> hence the int error
20151023 08:46:09 <sgothel> version 2.7.7
20151023 08:46:21 <sgothel> gluegen/make/lib/antlr.txt
20151023 08:47:19 <zubzub> version 2.7.7 is not available in maven it seems or as you say, don't match their own version
20151023 08:47:27 <sgothel> so now you are testing your work w/ maven, i.e. it already works well w/ ant ?
20151023 08:48:02 <sgothel> the maven issue here, would be another bug report
20151023 08:48:16 <sgothel> I hope we don't need to change too much for that one .. hmm
20151023 08:48:16 <zubzub> yeah I'll file one
20151023 08:48:31 <zubzub> sgothel: I dont think so, we unpack the antlr jar and put it in the gleugen jar
20151023 08:48:33 <sgothel> so you can make the nice commits :)
20151023 08:49:46 <sgothel> so gluegen does not contain its make/lib/antlr.jar ?
20151023 08:50:09 <sgothel> then we either add it - or, sweating, need to update our grammar ..
20151023 08:50:45 <zubzub> sgothel: yeah seems like it doesnt contain it, I get a classloading exception when I try to invoke gluegen from maven
20151023 08:51:26 <sgothel> @Mark: Gluegen compile time module seems to be not tested so far?
20151023 08:51:35 <sgothel> ^^ maven
20151023 08:52:25 <rmk0> .o.
20151023 08:52:36 <rmk0> this is news to me, it was working back in february
20151023 08:54:12 <rmk0> er, unless i'm thinking of something else
20151023 08:55:02 <sgothel> @Xerxes: Xj3D: Using JOGL's native windowing toolkit (NEWT) with no more Swing heavy/lightweight issues.
20151023 08:55:13 <sgothel> @Xerxes: we shall add this to our jogl page
20151023 08:56:01 <sgothel> @zuzub: would be great to have your clean commits then 'Bug xxxx: lalala' .. and we fix this maven/antlr issue later on
20151023 08:56:55 <sgothel> Xj3D: .. hmm .. still using sourceforge :)
20151023 08:57:28 <xranby_> the Xj3D update also include an fork of Aviatrix3D/j3d render libraries to support JogAmp JOGL 2.3.2
20151023 09:02:24 <rmk0> does the compile-time gluegen component need extra jars that won't be on the classpath when using maven?
20151023 09:03:11 <zubzub> nothing that the ordinary gluegen doesnt use
20151023 09:03:50 <zubzub> (in other words, it does not add extra dependencies)
20151023 09:07:44 <rmk0> to be clear, i do absolutely nothing to the maven jars... they are byte for byte identical to the jar files that are included in each jogamp-all-platforms.7z
20151023 09:07:53 <rmk0> and maven does nothing but put them on the classpath
20151023 09:08:08 <rmk0> so there's likely something extra going on outside of maven that's papering over a problem
20151023 09:08:52 <zubzub> the reason it doesnt work is because the gluegen maven jar does not define that it has a dependency on antlr
20151023 09:09:15 <rmk0> ah!
20151023 09:09:32 <zubzub> and since the gluegen maven jar does not contain the antlr classes themself, they are not on the classpath
20151023 09:09:38 <rmk0> that would be a problem :D
20151023 09:09:48 <rmk0> easily fixable... what version of antlr does it need?
20151023 09:09:51 <zubzub> 2.7.7
20151023 09:10:11 <zubzub> if I add it as an extra dep in my project, the antlr dep exceptions go away
20151023 09:10:44 <rmk0> which of the hundreds of antlr packages on maven central do you suggest?
20151023 09:11:54 <rmk0> 2.7.7 seems pretty ancient... they don't appear to have published versions that old
20151023 09:12:14 <zubzub> rmk0: they do
20151023 09:12:19 <zubzub> but it's not on maven.org
20151023 09:12:27 <zubzub> http://mvnrepository.com/artifact/antlr/antlr/2.7.7
20151023 09:12:33 <zubzub> :)
20151023 09:12:59 <rmk0> http://search.maven.org/#search|gav|1|g%3A%22antlr%22%20AND%20a%3A%22antlr%22
20151023 09:13:13 <rmk0> i'll use that one, then
20151023 09:13:23 <zubzub> yup
20151023 09:21:29 * rmk0 (~rmk0@anon) Quit (Ping timeout: 265 seconds)
20151023 09:21:44 * rmk0 (~rmk0@anon) has joined #jogamp
20151023 09:21:44 * rmk0 (~rmk0@anon) Quit (Changing host)
20151023 09:21:44 * rmk0 (~rmk0@anon) has joined #jogamp
20151023 09:21:53 <rmk0> ow
20151023 09:28:53 <zubzub> ok my further problems seem to stem from function pointers not being configured
20151023 09:37:35 <zubzub> sgothel: gluegen is able to define callbacks in java and pass them to c, right?
20151023 09:37:48 <zubzub> (if not, I'm not able to dogfood it)
20151023 09:38:44 <rmk0> sgothel: https://github.com/JogAmp/jogamp-scripting/pull/3
20151023 10:01:28 * bruce-_ (~x@anon) has joined #jogamp
20151023 10:07:50 * bruce- (~x@anon) Quit (*.net *.split)
20151023 10:15:57 <zubzub> damn the emitter is biting me in the ass again
20151023 10:36:32 * eclesia (~husky@anon) has joined #jogamp
20151023 10:36:56 <eclesia> hi
20151023 11:27:12 <sgothel> we can work on a 'jni->java callback config' .. TODO -> bug/feature
20151023 11:28:43 <sgothel> zuzub: if our unit tests work do far, we shall work on merging your commit
20151023 11:28:49 <sgothel> refine it
20151023 11:28:54 <sgothel> then add more stuff you may need
20151023 11:29:18 <sgothel> i.e. don't get lost ..
20151023 11:29:31 <sgothel> @thx Mark
20151023 11:31:13 <zubzub> sgothel: but the callback thing isn't really supported out of the box?
20151023 11:31:21 <zubzub> well with some simple configs that is
20151023 11:31:25 <zubzub> (no custom code injection)
20151023 11:31:29 <sgothel> its complicated .. use case dependent
20151023 11:31:43 <zubzub> :/
20151023 11:31:47 <sgothel> so I have 2 use cases from jogl .. and yours then
20151023 11:31:54 <sgothel> we can build a config for such
20151023 11:32:19 <sgothel> i.e. mine are pre-defined callbacks @ init, i,.e. during a register callback thing
20151023 11:32:51 <zubzub> mine are wayland callbacks :)
20151023 11:33:00 <zubzub> ie function pointers as method arguments
20151023 11:33:21 <sgothel> i.e. we need to define the arguments passed, the JNI method .. and where to callback, static way, or a pre- or post- arbitrary function
20151023 11:33:22 <zubzub> and I need to implement them on the java side
20151023 11:33:41 <sgothel> you can use function pointers
20151023 11:33:56 <sgothel> I am talking about the actual callback, i.e. from something native -> java
20151023 11:34:00 <zubzub> yeah
20151023 11:34:19 <sgothel> after we finished this task here, we work on the callback OK?
20151023 11:34:24 <zubzub> ok :p
20151023 11:34:37 <sgothel> baby steps :)
20151023 11:38:40 <sgothel> you case is similar, i.e. register a native callback function using a register-function, where the native callback function shall pass args to java?
20151023 11:39:06 <zubzub> yup
20151023 11:39:32 <zubzub> jna, bridj etc implement that case by defining a single method class/interface
20151023 11:39:55 <zubzub> that single method corresponds to the method signature of the method pointer
20151023 11:40:17 <sgothel> I will make a proposal ..
20151023 11:40:37 <sgothel> probably w/o need of extra interface using method signatures
20151023 11:41:09 <zubzub> anyway, it seems the annotation thing outputs code nicely in my project (gluegen configuration woes aside)
20151023 11:41:42 <sgothel> yeah .. would be good if I can have a look at it (clean commits .. nag nag) .. then we walk forward
20151023 11:42:52 <sgothel> the callback thing was on my plate for a long time :)
20151023 11:44:40 <zubzub> the callback thing seems to be a though problem for all java-native binding frameworks
20151023 11:44:45 <zubzub> swig doesnt support
20151023 11:44:52 <zubzub> bridj doesnt support it for armhf
20151023 11:45:01 <zubzub> only one I know that does support it nicely is jna
20151023 11:54:19 <bruce-_> xranby_: do you have any experiences with teavm?
20151023 12:47:57 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=1258 <- Add Multiple JNI/Native to Java Callback Generation Techniques
20151023 12:50:24 <xranby_> bruce-_: no hands on experience, it looks capable like a faster alternative to using GWT
20151023 12:50:46 <xranby_> and that it work using bytecode sounds awesome
20151023 12:51:30 <zubzub> sgothel: if you have that implemented
20151023 12:51:32 <zubzub> as an extra bonus
20151023 12:51:38 <zubzub> if my wayland java bindings work
20151023 12:51:45 <zubzub> you can reuse all those configs
20151023 12:51:50 <zubzub> to let gluegen generate wayland bindings
20151023 12:51:53 <zubzub> for use in newt
20151023 12:51:55 <xranby_> i have not explored teawm because i have never targeted deployment to javascript
20151023 12:52:06 <sgothel> yes ofc! thats on my mind as well :)
20151023 12:52:28 <bruce-_> yeah, I should try to find some time over the weekend. I'd be nice to write a WebgL drawer/renderer such that I can compile sketches made in my Java framework to web publishable thing
20151023 12:52:54 <sgothel> yikes (java->javascript)
20151023 12:53:15 <xranby_> example here is a libgdx compiled using teavm http://teavm.org/live-examples/libgdx-demo-invaders/
20151023 12:53:18 <sgothel> isn't there this new bytecode assembler thingy ongoing?
20151023 12:53:26 <sgothel> for 'javascript' ?
20151023 12:53:30 <bruce-_> asm.js?
20151023 12:53:47 <sgothel> forgot names .. maybe that ..
20151023 12:53:54 <sgothel> and isn't it allowing native code invocation?
20151023 12:54:22 <sgothel> I guess none of these tools are .. hmm
20151023 12:54:50 <bruce-_> yeah yeah. but the alternative surely is not running applets in a browser
20151023 12:55:32 <sgothel> yup .. I thought about it, and this brings me back to the browser thingy in java :)
20151023 12:55:57 <sgothel> i.e. webkit (or whatever) .. to be used directly w/ NEWT/JOGL
20151023 12:55:58 <xranby_> https://github.com/konsoletyper/teavm-libgdx teavm can only use javascript thus for 3d they get a webgl context context = (WebGLRenderingContext)element.getContext("webgl");
20151023 12:57:10 <xranby_> jocam could possibly support a teavm newt backend
20151023 12:57:13 <xranby_> jogl
20151023 12:57:27 <xranby_> if someone want to explore this route
20151023 13:00:24 <bruce-_> I should also check libgdx a bit better, but the 'game' aspect scares me for some reason
20151023 13:02:12 * packet0 (~packet0@anon) has joined #jogamp
20151023 13:03:54 <zubzub> sgothel: I assume gluegne can only handle structs(unions?) that are typedef'ed?
20151023 13:05:47 <xranby_> bruce-_: libgdx focus on es 2 and es 3 only
20151023 13:06:09 <zubzub> (the EmitStruct directive that is)
20151023 13:06:47 <xranby_> when you use libgdx you work using libgdx interfaces that exposes the common subset of es 2 and gl2 or es 3 and gl 3 using the libgdx GL20 and GL30
20151023 13:08:35 <xranby_> it also contain several classes that ease pushing tons of sprites
20151023 13:08:43 <xranby_> to make games easy
20151023 13:08:59 <xranby_> and a ui toolkit for menus
20151023 13:11:40 * bigpet (uid25664@anon) has joined #jogamp
20151023 13:21:11 <zubzub> sgothel: glue does not seem to take java keywords into account when generating java code (ie, argument names named "interface" because on the c side they are called "interface").
20151023 13:21:22 <zubzub> I did not find how to fix this in the manual
20151023 13:21:48 <zubzub> https://jogamp.org/gluegen/doc/manual/#SecSubstitution
20151023 13:21:50 <zubzub> I find this
20151023 13:22:00 <zubzub> but it seems to target a different use case
20151023 13:36:11 * ctrueden (~ctrueden@anon) has joined #jogamp
20151023 13:53:04 <zubzub> sgothel: I do not seem to find how to map a pointer struct field (gluegen ignores these)
20151023 13:53:16 <zubzub> it seems the gluegen manual on the website is not complete/up-to-date?
20151023 13:53:32 <zubzub> I find config keywords in the unit tests not explained in the manual
20151023 13:53:43 <zubzub> eg RenameJavaSymbol
20151023 13:55:26 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20151023 14:20:32 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20151023 15:34:30 * elect (~GBarbieri@anon) Quit (Ping timeout: 240 seconds)
20151023 15:39:55 * elect (~elect@anon) has joined #jogamp
20151023 16:04:17 * packet0 (~packet0@anon) Quit (Changing host)
20151023 16:04:17 * packet0 (~packet0@anon) has joined #jogamp
20151023 16:05:30 * eclesia (~husky@anon) has left #jogamp
20151023 16:06:37 <sgothel> @zuzub: we support 'struct lala {}' i.e. w/o typedef
20151023 16:06:51 <sgothel> but I am unsure about unions .. hmm
20151023 16:07:42 <sgothel> @zuzub: we need a grouping bug entry for all your issues/req-features
20151023 16:07:47 <sgothel> so nothing gets lost
20151023 16:08:14 <sgothel> i.e. 'Support Features as Required by Wayland Binding'
20151023 16:08:36 <sgothel> then we will 'spawn off the bigger ones ..' - sub bug entries
20151023 16:09:11 <sgothel> hmm .. 'java keywords' .. yeah :)
20151023 16:10:19 <sgothel> pointer struct field should be supported, look at the test*.h
20151023 16:45:36 <elect> sven, advantages of having a neutral format enum is that, for example, you keep you code cleaner, you have a better overview over the whole set of formats and you can easily detect if a format have some specific characteristics, such as compression
20151023 17:03:47 <elect> What was the point of declaring GLPixelAttributes.format final? In this way we must instantiate a new object every time we want to modify it..
20151023 17:14:32 <sgothel> A GLPixelAttributes instance is immutable since changing it would break the utilizing code/functionality
20151023 17:14:44 <sgothel> it is a key of GL fmt+type
20151023 17:15:15 <sgothel> within GL space there is nothing 'neutral' but GL oriented
20151023 17:15:40 <sgothel> however, it may make sense to better declare the supported set of formats, yes
20151023 17:16:02 <sgothel> this could be an enum using the original GL values - however, this also creates some overhead
20151023 17:16:13 <elect> I am thinking about creating enums having an int value corresponding to the GL value
20151023 17:16:25 <elect> ah
20151023 17:16:30 <sgothel> IMHO overhead ..
20151023 17:16:30 <elect> we wrote the same ^^
20151023 17:16:41 <elect> ok, I'll try without first
20151023 17:17:08 <sgothel> the abstract path will come in later using PixelFormat or a later ImageFormat
20151023 17:17:29 <sgothel> lets make the GL impl. code as tight as possible
20151023 17:17:51 <elect> yep, this is exactly what I am doing
20151023 17:17:54 <elect> right now
20151023 17:17:59 <sgothel> great
20151023 17:18:33 <elect> I am basically throwing away what is old/useless/replaced
20151023 17:19:09 <sgothel> @zuzub: I know you already started trying to use GlueGen for your binding .. i.e. going ahead. So you tell me when I may be able to merge your clean commits (nag nag :)
20151023 17:20:30 <sgothel> you could create a branch w/ the respective bug-number, i.e. bug1256
20151023 17:25:02 * packet0 (~packet0@anon) Quit ()
20151023 17:33:10 <elect> sven, we don't have any utility to perform int-vector-wise operation
20151023 17:33:45 <elect> like division, max, greaterThan, equal, compMul, shiftRight, ceilMultiple, ecc
20151023 17:33:49 <elect> ??
20151023 17:33:55 <elect> only FloatUtil?
20151023 17:34:08 <sgothel> VectorUtil and FloatUtil
20151023 17:34:14 <sgothel> but for floats
20151023 17:35:03 <sgothel> feel free to add int variants to VectorUtil if you need it
20151023 17:35:18 <sgothel> i.e. scaleVec2i(..) for example
20151023 17:35:21 <elect> it's already huge that class
20151023 17:35:25 <sgothel> i -> int
20151023 17:35:34 <sgothel> since I use an IDE ..
20151023 17:35:37 <elect> what about an Int/FloatVectorUtil
20151023 17:35:46 <elect> or something similar
20151023 17:36:39 <sgothel> IntVectorUtil OK
20151023 17:38:30 <sgothel> even though .. Vec<Type><Something> looks better hmm .. i.e. classes in order
20151023 17:38:32 <elect> this would mean rename VectorUtil to FloatVectorUtil
20151023 17:38:44 <elect> VectorFloatUtil
20151023 17:39:01 <elect> I can replicate all the methods in both classes
20151023 17:39:02 <sgothel> We can keep this refactoring in mind - not yet
20151023 17:39:09 <sgothel> i.e. we would have quite some ..
20151023 17:39:52 <sgothel> But yes, creating VectorIntUtil using 'int' would be nice
20151023 17:39:58 <elect> ok
20151023 17:40:12 <sgothel> dunno whether its useful to add all methods .. since many are useless for integers
20151023 17:40:19 <elect> ok
20151023 17:40:33 <elect> I start only with what I need then
20151023 17:40:38 * gouessej (5279cd9d@anon) has joined #jogamp
20151023 17:40:41 <sgothel> awesome
20151023 17:40:44 <gouessej> Hi
20151023 17:40:55 <sgothel> Hi Julien
20151023 17:40:56 <elect> you want low memory footprint I image..
20151023 17:41:03 <elect> salut
20151023 17:41:04 <sgothel> yeah
20151023 17:41:07 <elect> ^^
20151023 17:41:10 <elect> I know you
20151023 17:41:39 <sgothel> needs discipline to curb your OO explosions :)
20151023 17:42:12 <sgothel> i.e. for very high level stuff: fine - but more 'on the fly stuff' nope
20151023 17:42:55 <elect> ok
20151023 17:42:56 <sgothel> KUDOS Julien: always replying and busy w/ stuff
20151023 17:43:24 <gouessej> I've just updated JogAmp's Ardor3D Continuation
20151023 17:43:34 <sgothel> ah great
20151023 17:43:43 <sgothel> when will you push your git repo to jogamp?
20151023 17:43:43 <gouessej> I forgot a break when implementing a RFE in August
20151023 17:44:24 <gouessej> it's just a matter of entering a command line, isn't it?
20151023 17:44:35 <sgothel> sort of - sure :)
20151023 17:44:59 <sgothel> you have ssh access and all that AFAIK
20151023 17:45:09 <gouessej> I can do it immediately if you guide me so that we don't waste any time
20151023 17:45:37 <sgothel> now I need to brush up my git :)
20151023 17:45:47 <sgothel> so lets see .. branches and all of that
20151023 17:46:13 <sgothel> git://jogamp.org/srv/scm/ardor3d.git <- read
20151023 17:47:08 <sgothel> ssh://jgouesse@jogamp.org/srv/scm/ardor3d.git <- write
20151023 17:47:42 <sgothel> git remote add jogamp ssh://jgouesse@jogamp.org/srv/scm/ardor3d.git
20151023 17:47:51 <sgothel> git fetch --all
20151023 17:47:54 * Eclesia (~eclesia@anon) has joined #jogamp
20151023 17:48:28 <sgothel> your jogamp repo is available over the 'token' jogamp now
20151023 17:48:30 <sgothel> hmm
20151023 17:48:40 <sgothel> then: git push master jogamp
20151023 17:48:55 <sgothel> master is the branch
20151023 17:49:03 <sgothel> then do it for all other branches
20151023 17:49:09 <sgothel> then for all tags
20151023 17:49:20 <sgothel> 'git push --tags jogamp' <- I guess
20151023 17:49:45 <sgothel> that should populate things ..
20151023 17:51:00 <gouessej> Permission denied (publickey)
20151023 17:51:14 <sgothel> ssh-rsa AAAAB3NzaC1yc2EAAAADAQABAAACAQDS1coXg2skdDhMO1L91jjiQP4j/waCMyKcMAfHIknLjC5FKeN7sewDP1mhCjnikfSsKT3DkmoXWd3S/gUyQfYZVHbnx3KeqiWu0kkG+8OuxizxH41nEXsBaYW4pYFGp+r6tdJeIKUl5nxD/iBuWvSFjfqRDTsjdvfFzzB9Jgx6PpdZZ9fuplC3z+Vi1BDvYzC8ueuVdj+gdx/TXlkpp0/473skMy/LJ9xux24a5VNGt2XlKaFG6Ox5NOPAeUhHt/JfA17Om4aTrhXZ1hCt55e9YC6DU2twQ3Hth4EQ9DGjwsI/2/2nA/6HGmanAo23c/d7Y3kGJgA+dC5nrhSYf2XzxJMeFw9GV1rm/81FtEkrRf/pXSvH95Y3NDQdOtxa/ZJf9qA0K1Ro+eg2y/lgUFeIskYXFOHo4c
20151023 17:51:21 <sgothel> your public key ..
20151023 17:51:29 <sgothel> can you login in ?
20151023 17:51:33 <sgothel> via ssh
20151023 17:51:37 <gouessej> I understand the very first line but why should I do a fetch?
20151023 17:51:43 <gouessej> yes I can
20151023 17:51:56 <gouessej> I just did it to update JogAmp's Ardor3D Continuation
20151023 17:52:01 <sgothel> ah .. I just do that, so git knows the state
20151023 17:52:25 <sgothel> 'Permission denied (publickey)' ?
20151023 17:52:30 <gouessej> yes
20151023 17:52:32 <sgothel> thought you could login?
20151023 17:52:34 <sgothel> via ssh
20151023 17:52:37 <gouessej> yes
20151023 17:52:43 <gouessej> you can see the updates
20151023 17:52:49 <gouessej> it's very recent
20151023 17:53:25 <sgothel> if your remote is: ssh://jgouesse@jogamp.org/srv/scm/ardor3d.git
20151023 17:53:32 <sgothel> it uses ssh ..
20151023 17:53:39 <sgothel> and that is an ssh error message
20151023 17:53:44 <sgothel> but you can login .. funny
20151023 17:54:21 <gouessej> no it's ok now
20151023 17:54:35 <sgothel> good
20151023 17:54:40 <sgothel> what was it?
20151023 17:55:43 <gouessej> I used your command lines whereas I was already logged in jogamp.org
20151023 17:55:58 <sgothel> uh
20151023 17:56:10 <sgothel> yeah .. that was intended for 'doing from home' :)
20151023 17:56:17 <gouessej> ok
20151023 17:56:47 <gouessej> [gouessej@localhost Ardor3D (master)]$ git push master jogamp error: src refspec jogamp does not match any. error: impossible de pousser des références vers 'master'
20151023 17:56:48 <sgothel> oh .. there was another push option, i.e. to clone
20151023 17:57:06 <gouessej> impossible to push the references to 'master'?
20151023 17:57:44 <sgothel> ah it is vice versa
20151023 17:57:48 <sgothel> git push jogamp master
20151023 17:58:08 <sgothel> you can mirror ..
20151023 17:58:12 <sgothel> git push --mirror jogamp
20151023 17:58:18 <sgothel> i.e. 1:1 to jogamp
20151023 17:58:27 <sgothel> if your origin has it all (branches, tags ..)
20151023 18:00:42 <gouessej> I can use "git push --mirror jogamp" after "git push jogamp master", can't I?
20151023 18:01:00 <sgothel> hmm
20151023 18:01:05 <sgothel> one sec
20151023 18:01:43 <sgothel> now you can :)
20151023 18:02:04 <sgothel> I allowed non-fast-forward temporary .. i.e. overwrite
20151023 18:02:22 <gouessej> thanks
20151023 18:07:40 <gouessej> Is it ok now?
20151023 18:08:41 <sgothel> looks good
20151023 18:09:25 <gouessej> Ok, I'll just have to use "git push jogamp master" to push the other changes in the future
20151023 18:13:18 <gouessej> Someone talked about WebGL yesterday. You can use WebGL4J in the meantime, it's quite lightweight, it's based on GWT, it already supports WebGL 2: https://github.com/sriharshachilakapati/WebGL4J
20151023 18:14:45 <sgothel> the asm.js seems to be highly performant .. if one could enhance, add other external libs ..
20151023 18:14:58 <sgothel> but not really my kind of beer :)
20151023 18:15:29 <sgothel> http://ejohn.org/blog/asmjs-javascript-compile-target/ <- quite 'mature' blog entry :)
20151023 18:15:37 <sgothel> read: old
20151023 18:16:06 <sgothel> I guess this brings us back to integrate the browser into JOGL/NEWT :)
20151023 18:16:34 <sgothel> after GraphUI is done .. and our outstanding issues are solved, maybe something
20151023 18:17:24 <gouessej> Do you know Web Assembly?
20151023 18:17:44 <sgothel> nope .. just read about it, and how it may allow better optimizations
20151023 18:18:11 <sgothel> so one more AOT target
20151023 18:19:14 <sgothel> http://jogamp.org/git/?p=ardor3d.git;a=summary <- maybe remove the 'external' refs
20151023 18:19:22 <sgothel> i.e. jogamp/master and origin/master :)
20151023 18:19:30 <gouessej> In my humble opinion, maybe it will be possible to modify Bck2Brwsr to produce Web Assembly from Java bytecode instead of Javascript
20151023 18:19:45 <sgothel> they mention this on their page, right
20151023 18:20:00 * rmk0 (~rmk0@anon) Quit (Ping timeout: 250 seconds)
20151023 18:20:35 <gouessej> Do I have to use "git remote remove"?
20151023 18:20:58 <gouessej> No these are two branches :s
20151023 18:21:12 <sgothel> a special push command .. hmm
20151023 18:21:43 <sgothel> http://gitready.com/beginner/2009/02/02/push-and-delete-branches.html
20151023 18:22:31 <sgothel> that would be ':jogamp/master' I guess
20151023 18:22:33 <sgothel> etc
20151023 18:22:48 <sgothel> git push jogamp :jogamp/master
20151023 18:23:53 <gouessej> done, but I have to delete them locally too...
20151023 18:24:38 <gouessej> git tells me that they aren't found
20151023 18:24:43 <gouessej> Is it ok now?
20151023 18:28:30 <elect> is the Flusher still necessary?
20151023 18:30:34 <gouessej> Yes in my humble opinion
20151023 18:30:42 <gouessej> Why wouldn't it be necessary?
20151023 18:31:34 <gouessej> You can use it to release the native resource of a texture data
20151023 18:31:48 * rmk0 (~rmk0@anon) has joined #jogamp
20151023 18:31:49 * rmk0 (~rmk0@anon) Quit (Changing host)
20151023 18:31:49 * rmk0 (~rmk0@anon) has joined #jogamp
20151023 18:32:00 <gouessej> It's very important when you use a direct NIO buffer as a storage
20151023 18:32:11 <gouessej> as it's up to you to manage the native heap
20151023 18:34:59 <gouessej> elect: I have to update my game so that the italian players can see your translation ;)
20151023 18:35:05 <elect> ^^
20151023 18:35:24 <elect> how do you the flusher?
20151023 18:35:38 <elect> and which native resources do you refer to?
20151023 18:41:29 <gouessej> You can pass a flusher when creating a TextureData object
20151023 18:41:45 <gouessej> It's called when TextureData.flush() is called
20151023 18:42:22 <gouessej> Unlike the indirect NIO buffer, the direct NIO buffer don't use the managed Java heap
20151023 18:42:28 <gouessej> they use the native heap
20151023 18:42:37 <gouessej> it's unmanaged
20151023 18:43:11 <gouessej> the garbage collector doesn't take care of it, it releases the native memory only if it risks to run out of Java memory
20151023 18:43:23 <gouessej> "buffers"
20151023 18:43:25 <elect> how would you create a direct NIO buffer through a TextureData?
20151023 18:45:01 <gouessej> you can pass a NIO buffer to the constructor
20151023 18:45:39 <gouessej> When you create a texture object, it creates a TextureData object too
20151023 18:48:13 * rmk0 (~rmk0@anon) Quit (Ping timeout: 252 seconds)
20151023 18:48:58 <gouessej> When you use TextureIO, it creates an indirect NIO buffer as far as I know
20151023 18:49:18 * rmk0 (~rmk0@anon) has joined #jogamp
20151023 18:49:18 * rmk0 (~rmk0@anon) Quit (Changing host)
20151023 18:49:18 * rmk0 (~rmk0@anon) has joined #jogamp
20151023 18:49:24 <gouessej> I don't use those classes in JogAmp's Ardor3D Continuation and I use my own excellent deallocator
20151023 18:51:17 <elect> :)
20151023 18:52:33 <gouessej> It's a complicated but well documented class: http://sourceforge.net/p/tuer/code/HEAD/tree/pre_beta/src/main/java/engine/misc/DeallocationHelper.java
20151023 18:53:57 <gouessej> it's fully tested: http://sourceforge.net/p/tuer/code/HEAD/tree/pre_beta/src/test/java/engine/misc/TestDeallocationHelper.java
20151023 18:59:20 <Eclesia> still working on this game ^^
20151023 19:01:03 <gouessej> Eclesia: I won't give up :p
20151023 19:01:18 <gouessej> Eclesia: 9 years :)
20151023 19:10:06 * rmk0 (~rmk0@anon) Quit (Ping timeout: 240 seconds)
20151023 19:11:02 <gouessej> I'm silly, I have forgotten to translate the pause menu
20151023 19:24:04 <Eclesia> gouessej: don't give up, but try to finish it before you get too old :)
20151023 19:24:17 <gouessej> Eclesia: lol
20151023 19:26:12 * Eclesia wondering what unlicense-lib will look like when it reaches 9years old
20151023 19:28:33 * rmk0 (~rmk0@anon) has joined #jogamp
20151023 19:36:12 * rmk0 (~rmk0@anon) Quit (Ping timeout: 268 seconds)
20151023 19:39:27 * elect (~elect@anon) Quit (Ping timeout: 256 seconds)
20151023 19:45:25 <gouessej> Eclesia: Something great for sure :)
20151023 19:45:39 <Eclesia> let's hope
20151023 19:47:38 * Eclesia wish for a public domain VM
20151023 19:49:50 <gouessej> I've found a bug just before planning to upload a new version :(
20151023 20:01:53 * rmk0 (~rmk0@anon) has joined #jogamp
20151023 20:01:53 * rmk0 (~rmk0@anon) Quit (Changing host)
20151023 20:01:53 * rmk0 (~rmk0@anon) has joined #jogamp
20151023 20:04:52 <gouessej> I've found the root cause of the bug, the level disappears when I retry a mission. I ask my cleanup mechanism to destroy the level model without setting its reference to null. I expected a crash of the JVM
20151023 20:22:35 * Eclesia (~eclesia@anon) has left #jogamp
20151023 20:23:15 <gouessej> I'm currently uploading the new version. I'll have to improve the German translation later
20151023 20:23:47 <sgothel> *duck and hide* :)
20151023 20:24:25 <sgothel> oh well, is it all in one file or so ? then I can help you doing EN -> DE
20151023 20:24:50 <sgothel> you use git?
20151023 20:25:34 <sgothel> oh .. you speak German? tsts, and we never used it :)
20151023 20:25:56 <gouessej> Sven, you know I understand your mother tongue
20151023 20:26:09 <gouessej> I studied it ten years
20151023 20:26:49 <sgothel> Ah .. three languages then, another language genius, KUDOS
20151023 20:27:14 <sgothel> Well .. I failed miserably in french .. only half a year :-/
20151023 20:27:20 <gouessej> http://sourceforge.net/p/tuer/code/HEAD/tree/pre_beta/src/main/resources/i18n/MessagesBundle_de.properties
20151023 20:27:28 <sgothel> and in Chinese so far .. sniff
20151023 20:27:57 <gouessej> As I speak very rarely in German, I have forgotten tons of words but I was able to understand almost all your words in Belgium and in USA
20151023 20:28:10 <gouessej> about 80%
20151023 20:28:19 <sgothel> oh .. I used German there?
20151023 20:28:23 <gouessej> a few
20151023 20:28:28 <sgothel> how impolite of me
20151023 20:28:32 <gouessej> no
20151023 20:28:41 <gouessej> I told you that it was ok for me
20151023 20:28:49 <sgothel> ah .. me getting old :)
20151023 20:29:01 <sgothel> see, example: SHORTCUTS=Schnellzugriffe
20151023 20:29:36 <sgothel> this is the kind of translation .. while correct, my stomach turns .. somewhat
20151023 20:29:41 <gouessej> It's difficult to know whether it is the good word in the context
20151023 20:29:51 <sgothel> today, so many english words are in use here ..
20151023 20:30:00 <gouessej> in Italian too
20151023 20:30:05 <gouessej> in French too
20151023 20:30:15 <sgothel> so when talking computer stuff .. no German please :)
20151023 20:30:30 <gouessej> "Shortcuts"?
20151023 20:30:51 <sgothel> I thought in France it is obligatory to use French
20151023 20:31:02 <sgothel> yeah .. just leave it Shortcuts
20151023 20:31:27 <sgothel> I send you one version by email ..
20151023 20:34:45 * gouessej (5279cd9d@anon) Quit (Ping timeout: 246 seconds)
20151023 21:09:58 <sgothel> that was a fun exercise .. and I realize its often hard to translate something w/o context :)
20151023 21:13:09 <sgothel> http://www.zerohedge.com/news/2015-10-23/meet-stunningly-catastrophic-patricia-worlds-strongest-storm-ever-about-hit-mexico
20151023 22:01:57 * ctrueden (~ctrueden@anon) has left #jogamp
20151024 00:03:57 * xranby_ (~familjen@anon) Quit (Quit: Leaving.)
20151024 00:59:22 * xranby_ (~familjen@anon) has joined #jogamp
20151024 03:31:08 * packet0 (~packet0@anon) has joined #jogamp
20151024 04:59:21 * packet0 (~packet0@anon) Quit ()
20151024 05:05:18 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20151024050518.html