#jogamp @ irc.freenode.net - 20140630 05:05:29 (UTC)

20140630 05:05:29 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20140629050528.html
20140630 05:05:29 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20140630050529.html
20140630 06:43:03 * eclesia (~husky@anon) has joined #jogamp
20140630 06:43:09 <eclesia> good morning
20140630 07:27:17 * monsieur_max (~maxime@anon) has joined #jogamp
20140630 07:33:51 <sgothel> Alternative for HLSL 'float4x4 lerp(float4x4, float4x4, float)' in GLSL ?
20140630 07:34:09 <sgothel> Not available on GLSL: 'mat4 mix(mat4, mat4, float)'
20140630 07:48:10 <eclesia> has anyone played with vp8 codec ?
20140630 07:56:12 <zubzub> 23:40 < rmk0> point would be to separate everything into maven modules, with the correct interdependencies
20140630 07:56:20 <zubzub> if you can/want please use gradle ;)
20140630 07:56:44 <zubzub> the end result will still be maven artifacts
20140630 07:56:59 <zubzub> but you will have a lot of build flexibility if you use gradle
20140630 07:57:28 <zubzub> things where you normally have to write a maven plugin, can be easily be 'inlined' in a gradle build file
20140630 07:58:08 <sgothel> my point here: while I like the idea of cleaning up things and be more modularized .. if this is not yet 'perfect' - I surely will not like the idea to use maven as the official build system, since maven's dependencies itself are huge - more than what we use today (ant, ..)
20140630 07:58:46 <sgothel> having said that, it's OK if somebody maintains a parallel, non-official, build system
20140630 07:59:08 <sgothel> but that means - it must be maintained properly
20140630 07:59:47 <zubzub> you mean maven dependencies on the/a build server?
20140630 07:59:59 <sgothel> everything
20140630 08:00:08 <zubzub> ideally you just use the maven fat jar that you dld from the site
20140630 08:00:13 <zubzub> and don't depend on system jars
20140630 08:00:15 <sgothel> when I launch a maven build it pulls stuff from the internet
20140630 08:00:22 <sgothel> and that is something we cannot tolerate
20140630 08:00:31 <zubzub> yes that's maven
20140630 08:00:36 <zubzub> maven dlds the internet ;)
20140630 08:00:40 <zubzub> I hate it as well
20140630 08:00:45 <zubzub> but you can do an offline build as well
20140630 08:00:57 <zubzub> so it doesn't do that
20140630 08:01:05 <zubzub> that said, gradle doesn't do it iirc
20140630 08:01:29 <sgothel> maybe, however .. it will be a parallel option build system for now - iff anybody really likes to dive in deep and do all the hard work maintaining it
20140630 08:01:40 <sgothel> and the latter is a strong requirement
20140630 08:01:59 <zubzub> yeah, it'll be a hard time finding anyone doing it for fun for free :p
20140630 08:02:13 <sgothel> nothing is 'for free'
20140630 08:02:15 <zubzub> If I was unemployed and single maybe :p
20140630 08:02:41 <sgothel> i.e. it should be one who has a strong benefit from it, ofc .. otherwise there would be no motivation
20140630 08:02:50 <zubzub> true
20140630 08:02:57 <sgothel> 'just for fun' .. this task is ridiculous
20140630 08:03:13 <zubzub> I still need to fix wayland-java, improve the jni, and then try to make newt use it
20140630 08:03:46 <sgothel> so Mark (@Mark) indeed, is the only one I could think of right now .. and Harvey ofc
20140630 08:04:07 <sgothel> you do it manually ?
20140630 08:04:10 <sgothel> the JNI ?
20140630 08:04:24 <sgothel> I just enhanced GlueGen for structs and arrays .. etc
20140630 08:09:19 <zubzub> the wayland-jni?
20140630 08:09:26 <zubzub> *wayland-java jni
20140630 08:09:31 <sgothel> yup
20140630 08:09:36 <zubzub> yes, well I did not write the original jni
20140630 08:09:47 <zubzub> but it's partially automatic as well
20140630 08:09:52 <sgothel> ah
20140630 08:09:57 <zubzub> where a wayland protocol xml is parsed and java source code is generated
20140630 08:10:06 <sgothel> nice
20140630 08:10:23 <zubzub> but libwayland has some basic boilerplate implementation of it's core protocol
20140630 08:10:32 <zubzub> and to map that to java, there is some hand written jni
20140630 08:11:15 <zubzub> I was investigating that jni last week and noticed it was written by someone not very familiar with java...
20140630 08:11:31 <zubzub> in a sense that 'it works', but the code is not very pretty/optimal
20140630 08:11:51 <zubzub> and for some reason it segfaults now (so there goes the 'it works')
20140630 08:11:59 <zubzub> (more like, 'it used to work'_
20140630 08:12:53 <zubzub> and when trying to debug, I noticed arch linux does not provide debug symbols in there package manager
20140630 08:13:11 <zubzub> so that was a major point of frustration/waste of time
20140630 08:13:18 <zubzub> *their
20140630 08:13:47 <sgothel> benefits of code generation / compiler .. hmm, optimize it once in generator :)
20140630 08:14:17 <sgothel> yeah, need to add attributes like 'in' 'out' .. to be able to skip copying data for arrays (JNI stuff)
20140630 08:15:02 <zubzub> you mean like direct bytebuffers?
20140630 08:16:00 <sgothel> anything .. i.e. ANSI-C doesn't know about whether a non-const argument maybe used for input values as well
20140630 08:17:51 <eclesia> zubzub, sgothel: "If I was unemployed and single maybe" "nothing is 'for free'" "for fun" <-- lol . looks like me, doing thing for free and for fun, but i'm not unemployed :D
20140630 08:19:11 <sgothel> well .. it's fun for me to work on all this - ofc! .. however, there is some use-case scenario behind it :)
20140630 08:19:20 <sgothel> otherwise my wife would kick me hard :)
20140630 08:19:46 <sgothel> and I myself as well .. hehe
20140630 08:20:02 <eclesia> isn't that the fun part ? (being kicked) ^^
20140630 08:20:18 <sgothel> 'for fun' was meant like 'for no good reason .. like business case, huge improvements ..'
20140630 08:20:33 <sgothel> yeah, the benefit of partnership
20140630 08:28:55 <eclesia> so nobody had fun with a VP8 decoder ?
20140630 08:29:57 <sgothel> using it ? maybe - looking into code ? nope, but there is a reference of a review .. hmm
20140630 08:30:53 <sgothel> http://x264dev.multimedia.cx/archives/541
20140630 08:30:58 <sgothel> http://x264dev.multimedia.cx/archives/377
20140630 08:31:09 <sgothel> discussed it a few month ago w/ Harvey ..
20140630 08:31:36 <sgothel> AFAIK he uses something in a java only implementation .. which is OSI licensed AFAIK
20140630 08:32:29 <eclesia> nice docs, thanks
20140630 08:35:14 <eclesia> same question on VP6 :D
20140630 08:35:46 <sgothel> https://ixquick.com/ :)
20140630 08:36:00 <eclesia> héhé
20140630 09:08:58 * magaio (~magaio@anon) Quit (Ping timeout: 240 seconds)
20140630 09:09:23 * magaio (~magaio@anon) has joined #jogamp
20140630 10:07:47 * [Mike] (~Mike]@anon) Quit ()
20140630 10:08:11 <rmk0> i'm not going to seriously advocate for switching jogl/gluegen to a maven build, but most of the arguments against it that i just heard here are fallacies
20140630 10:08:44 <rmk0> firstly the idea that it has "a lot of dependencies" is sort of silly, and not correct
20140630 10:09:12 <rmk0> ant has a lot of built-in tasks. if those tasks were separated out into plugins and the program downloaded versions of them the first time it started, is that "a lot of dependencies"?
20140630 10:09:20 <rmk0> implying that the system is giant and complex
20140630 10:10:44 <rmk0> when you have to write thousands of lines of imperative build scripts, with no way of formally stating the dependencies between parts (meaning the whole pile has to be recompiled every time you make a minor change to any part of the code) is idiotic
20140630 10:11:18 <rmk0> when you have everything in one giant program like ant, all you can do is say "this code probably builds on versions of ant >= x"
20140630 10:12:17 <zubzub> rmk0: that's why you use gradle ;)
20140630 10:12:45 <rmk0> i looked at it earlier... it looks like you write imperative build scripts, and never specify version numbers for plugins
20140630 10:12:48 <rmk0> am i wrong?
20140630 10:12:53 <zubzub> yes
20140630 10:13:11 <zubzub> :)
20140630 10:13:12 <rmk0> the tutorials make it seem exactly like ant
20140630 10:13:37 <zubzub> to me it looks exactly like maven
20140630 10:13:55 <zubzub> except you don't have to write a seperate maven plugin if your build needs to do special steps
20140630 10:14:10 <zubzub> you can just write it in the build file itself or seperate it out
20140630 10:14:28 <zubzub> + you need to write a whole lot less lines than maven
20140630 10:14:41 <rmk0> i don't really consider that an advantage... whole problem with building software is that everyone thinks their own build is somehow special and requires extra steps
20140630 10:15:01 <rmk0> meaning the first thing anyone has to do when trying to work on a project they didn't write is learn someone else's build scripts
20140630 10:15:02 <zubzub> eh?
20140630 10:15:16 <zubzub> well to me that's aproblem with maven
20140630 10:15:29 <zubzub> trying to find the source code of a custom plugin
20140630 10:15:34 <zubzub> that somebody wrote
20140630 10:15:40 <zubzub> that's on some repository
20140630 10:15:51 <zubzub> with gradle, the 'plugin' is just part of teh build script itself
20140630 10:16:07 <rmk0> that last qualifier is the major point... as you know, everyone is encouraged to publish to maven central
20140630 10:16:30 <rmk0> i've been using it for about five years now, and haven't once needed to use anything that isn't in maven central
20140630 10:16:32 <zubzub> you can still write plugins if you want
20140630 10:16:35 <rmk0> meaning the source jars are right there
20140630 10:16:38 <zubzub> and use them in a 'maven way'
20140630 10:16:41 <zubzub> and publish them
20140630 10:16:57 <rmk0> it's not the ability to write plugins that's the issue... i don't want to learn other people's build scripts
20140630 10:17:08 <rmk0> got better ways to be wasting my time!
20140630 10:17:16 <zubzub> then don't learn it :)
20140630 10:17:41 <zubzub> I don't understand how maven could be better in that respect
20140630 10:18:28 <rmk0> mainly because they're so viciously authoritarian... people follow the conventions or suffer for failing to
20140630 10:18:42 <rmk0> it shouldn't work, but for some reason it does... i never have to "learn" how a maven project is built
20140630 10:19:27 <rmk0> is basically their entire philosophy... "your build isn't special and you're not a unique snowflake, get over it"
20140630 10:19:52 <zubzub> yes
20140630 10:19:57 <zubzub> so it gradle...
20140630 10:20:11 <zubzub> http://www.gradle.org/docs/current/userguide/java_plugin.html
20140630 10:20:14 <rmk0> i'm not disputing that, i can only go by how the tutorials made it look
20140630 10:20:34 <zubzub> you can write a gradle build file
20140630 10:20:37 <rmk0> i'm mainly arguing against ant
20140630 10:20:44 <zubzub> ah yes
20140630 10:20:52 <zubzub> well don't have to convince me :p
20140630 10:20:53 <rmk0> because... ant is basically one step above writing a fucking shell script
20140630 10:21:02 <zubzub> yeeep
20140630 10:21:11 <rmk0> leads to all the problems we see with building jogl and gluegen now... is extremely difficult to work on the project
20140630 10:21:29 <zubzub> we had the same problem with qtjambi
20140630 10:21:31 <rmk0> i stopped sitting through 2-3 minute compiles for minor changes back in 2005
20140630 10:23:53 <rmk0> reading chapter 6 of the gradle user guide
20140630 10:24:18 <rmk0> ... are people violently encouraged not to use the things they're describing?
20140630 10:24:28 <zubzub> http://www.gradle.org/docs/current/userguide/tutorial_java_projects.html#N103C9
20140630 10:24:40 <zubzub> gradle is a generic build system
20140630 10:24:48 <rmk0> yeah, saw that
20140630 10:25:07 <rmk0> so... which version of the java plugin is executed when i say "apply plugin: 'java'"?
20140630 10:25:07 <zubzub> I don't read all the low level stuff they describe
20140630 10:25:25 <zubzub> I guess the one that comes bundled with your maven install
20140630 10:25:32 <zubzub> *gradle
20140630 10:25:33 <rmk0> nasty
20140630 10:26:03 <zubzub> ie it's part of gradle afaik
20140630 10:26:10 <zubzub> but don't take my word on it
20140630 10:26:42 <zubzub> http://www.gradle.org/docs/current/javadoc/org/gradle/api/plugins/JavaPlugin.html
20140630 10:27:13 <rmk0> point i'm making is that if version numbers aren't specified in the build, then you're arbitrarily depending on "whatever the user has"
20140630 10:27:17 <rmk0> leading to nondeterminism
20140630 10:27:34 <zubzub> well for custom plugin do you specify the version
20140630 10:28:15 <zubzub> http://www.gradle.org/docs/current/userguide/plugins.html
20140630 10:28:43 <zubzub> or you can at least
20140630 10:30:34 <rmk0> again... point isn't really whether you can or can't (and i can't see how you could specify any for the built-in plugins anyway; if they're built into gradle then presumably the build fails on any other version)
20140630 10:30:45 <rmk0> if you're not required to, then nobody will
20140630 10:30:55 <rmk0> maven will scream at you if you fail to provide version numbers
20140630 10:31:33 <rmk0> if developers aren't required to do the right thing by a machine, they will generally always do the wrong thing
20140630 10:32:45 <rmk0> i don't want to be seen to be advocating for maven specifically... there's a lot about it i dislike
20140630 10:33:02 <rmk0> but i do want to be seen to be advocating for the kind of determinism that results from doing things the way maven does things in general
20140630 10:33:16 <rmk0> strict adherence to detailed version numbers, no custom logic in builds... no "scripting"
20140630 10:34:04 <rmk0> in my experience, that stuff always leads to disaster (grab any C++ project off github and see if you can build it, see if you can fix any of the inevitable compilation errors, etc)
20140630 10:34:34 <rmk0> see what minor environmental changes result in build failures because the system never asked you to account for them
20140630 10:34:40 <rmk0> etc
20140630 10:34:44 <rmk0> i'm sure you know all this already
20140630 10:35:04 <rmk0> you sound like you've built software more than once :D
20140630 10:36:22 <rmk0> i only use maven because it's the first build system i've used to date that rejected the approach that everyone else seems to take
20140630 10:36:39 <rmk0> well, that's a bit strong... "only use" isn't quite right
20140630 10:36:47 <rmk0> i stayed with it for those qualities, after having used ant for a year or so
20140630 10:36:55 <rmk0> not to mention the build systems in other languages
20140630 10:37:07 <rmk0> (one of which i authored)
20140630 10:37:16 <rmk0> .. build systems, not languages
20140630 10:40:02 * xranby (~xranby@anon) has joined #jogamp
20140630 10:41:49 <xranby> The new US #FACTA tax law and international agreements go in full effect tomorrow on 1'st July 2014
20140630 10:42:44 <sgothel> @Mark: not just that there are dependencies .. - they all sure have - but downloading them from the internet would be a no-go for me for a build server
20140630 10:43:17 <xranby> I have guess this mean that US is the first company to start tax non-US company’s with US partial owners and non-US individuals with past connections with US
20140630 10:43:24 <xranby> US is the first government...
20140630 10:43:47 <rmk0> sgothel: you do realize it downloads the versions of plugins specified in the build and doesn't do it again?
20140630 10:43:56 <rmk0> just to be clear
20140630 10:44:13 <rmk0> i can appreciate that "even once" could be a problem
20140630 10:44:19 <sgothel> no fight Mark .. just mentioning :)
20140630 10:44:37 <sgothel> i.e. I like to be in control of what is being used for a build, sure there is a way to deal w/ it ..
20140630 10:44:40 * rmk0 puts on boxing gloves
20140630 10:45:05 <xranby> http://world.time.com/2013/12/20/swiss-banks-tell-american-expats-to-empty-their-accounts/
20140630 10:45:39 <sgothel> IMHO the realistic thing is about maintainership, so as it is now - you @Mark, would be the one I guess realizing it
20140630 10:45:53 <rmk0> yeah, can appreciate that too
20140630 10:45:55 <sgothel> and when it's there .. we see :)
20140630 10:45:58 <sgothel> hehe
20140630 10:46:42 <sgothel> Hi Xerxes .. more political crap going on these days .. after the sell out of sovereignty of EU countries, now the US .. and so forth ? :)
20140630 10:46:44 <rmk0> i'd like to work on jogl/gluegen more, but the build system makes it hard
20140630 10:46:53 <rmk0> that's not a criticism of jogl/gluegen, but of ant
20140630 10:47:06 <sgothel> I know .. but not liking a build system is no reason either
20140630 10:47:43 <sgothel> i.e. if it hinders one from doing things .. then we sure have a problem
20140630 10:48:51 <xranby> sgothel: i know sweden is about to sign the agreement but i as a citicen cant read the agreement unless they have signed it
20140630 10:48:52 <sgothel> however: I agree - setting up a project 'ad-hoc' w/ ant .. is tricky :)
20140630 10:48:58 <xranby> its all secret mumbo jumo
20140630 10:49:06 <rmk0> i think it's a fundamental problem with systems like ant
20140630 10:49:25 <rmk0> they basically don't have any semantic information and the build is just a long list of imperative instructions
20140630 10:49:30 <xranby> i kind of think sweden already have signed it because swedish bank communitys are already aware of it
20140630 10:49:38 <sgothel> well, EU is not democratic .. thats it
20140630 10:49:40 <rmk0> meaning that the build can't be separated, and external tools can't be more intelligent about what gets built
20140630 10:49:52 <sgothel> @Mark: I agree !
20140630 10:50:12 <rmk0> hence having to turn off things like automatic building in eclipse (losing real-time type errors and so on), because eclipse can't do anything less than a full build without knowledge of the scripts
20140630 10:50:16 <sgothel> @Mark: It's only .. that it needs LOOOVE :) lots of love
20140630 10:50:17 <rmk0> yeah, you know all this already
20140630 10:50:20 <xranby> the quesion is.. are we all "american" US-persons tomorrow?
20140630 10:50:39 <sgothel> we are all: EU+US .. 'western' conglomerate :)
20140630 10:50:45 <sgothel> since a while actually
20140630 10:50:49 <xranby> ok then all tax go to us starting tomorrow
20140630 10:51:24 <sgothel> before the EU vote .. well, parlament did pass some TTIP laws .. i.e. who can sue whom .. etc (companies -> gov.)
20140630 10:53:16 <sgothel> and Ukraine and the other east countries has signed that association thing, which includes agreeing upon military operations which they cannot influence. Besides, this violates an old agreement lead to dissolution of the USSR. So all HAIL THE WEST :) (cynical)
20140630 10:54:31 <sgothel> assuming those folks (politicians) are not stupid, I guess they speak w/ one voice .. and came to the conclusion that war and pressure is the answer for the next few years, and a strong western group is required to survive. Oh dear ..
20140630 10:56:09 <sgothel> ok .. back to my little world :)
20140630 10:57:31 <xranby> jaja.. we will see if this is the first step before all national currencys gets revaluated
20140630 10:57:53 <xranby> revaluated based on the current assets of all countrys
20140630 10:58:24 <xranby> many people think this is the first step before a big dollar collapse
20140630 10:58:32 <sgothel> I won't expect fairness here :)
20140630 10:59:00 <sgothel> and currency doesn't really matter for the wealthy anyways, they store wealth otherwise
20140630 10:59:30 <sgothel> just the _use_ of currency .. and hence demand, might be of interest - see 'petro dollar'
20140630 10:59:36 <xranby> and all who are in debt now will be in more debt tomorrow thus more dependent on the people who are super whealty
20140630 11:00:22 <xranby> thus it will only strengthen the grip of people who are in controll today
20140630 11:00:33 <sgothel> the nice discussion regarding freedom being killed by welfare .. rendering them non-free, dependent etc. Bismark's great idea those days to stay in power :)
20140630 11:00:37 <xranby> the people who are in controll will have a stronger grip on the average person
20140630 11:01:15 <sgothel> so one would need to separate welfare from the powers/politics
20140630 11:02:10 <xranby> ok lets see what happens tomorrow
20140630 11:02:25 <sgothel> little government / subsidizing (forgot the proper term) like in Switzerland .. still works
20140630 11:02:56 <sgothel> subsidiary .. something like that
20140630 11:03:35 <eclesia> we'll have a new revolution in france, like always ^^. I just hop it will come soon, before it's too late
20140630 11:04:20 <sgothel> oh dear .. just chatted about french revolution w/ daughter a few days ago (school) .. bloody business full of corruption :)
20140630 11:04:49 <sgothel> heck, just look at the middle east - all funded w/ our tax dollars .. poor souls
20140630 11:05:07 <sgothel> imaginary freedom
20140630 11:05:19 <xranby> eclesia: have FACTA been discussed at all inf rench media? http://www.french-news-online.com/wordpress/?p=34207#axzz367TCq5Ya
20140630 11:05:27 <monsieur_max> related with the discussion : http://biblioklept.org/2013/06/08/huxley-vs-orwell-the-webcomic-2/
20140630 11:05:43 <xranby> at least your governemnt have probably already signed an agreement on your behalf
20140630 11:05:43 <sgothel> IMHO must be tackled by the whole world ability to actually being capable to FEED everybody properly, etc etc
20140630 11:07:25 <xranby> "FATCA is an American law that requires all foreign financial institutions to turn over lists of people with connections to the U.S. (US Persons) and their financial information. Generally, this kind of thing is illegal in most countries and a clear violation of basic privacy rights assured by many nation-state charters and constitutions. "
20140630 11:07:42 <xranby> not any more.. because your government have signed the deal on your behalf
20140630 11:07:59 <sgothel> well, US citizen always need to pay tax to the US :)
20140630 11:08:14 <xranby> the thing is that the term US persons is verry loose defined
20140630 11:08:14 <sgothel> now .. this information 'just' becomes legal to the IRS .. haha
20140630 11:08:27 <xranby> a green card is enough
20140630 11:08:36 <sgothel> interesting :)
20140630 11:08:53 <sgothel> awaiting my IRS invoice ..
20140630 11:09:17 <eclesia> xranby: normal medias don't talk about it. but honestly, things are really going to blow up here, the current president is hated by the population with less then 2% of 'favorable'. same goes for the first minister, a raciste who openly said jews are above other.
20140630 11:09:47 <sgothel> at least that matches political correctness .. haha
20140630 11:10:15 <sgothel> (yeah, I know things are f**cked *p for a long time)
20140630 11:11:39 <sgothel> @Max: Thx for the cartoon link - I prefer to suffer from things I love :)
20140630 11:12:15 <eclesia> maybe I should go back where I lived before, in french polynesia, or perhaps new zealand...
20140630 11:12:32 <eclesia> far away from all this crap
20140630 11:12:39 <sgothel> lol .. New Zealand is sold out for a long time already :)
20140630 11:14:15 <eclesia> well then, back to tahiti ^^
20140630 11:14:58 <sgothel> how is life there ? Internet access ?
20140630 11:15:40 <eclesia> very low connection, if you have a 20Ko/s you're at the max speed :D
20140630 11:15:50 <eclesia> Kb*
20140630 11:15:55 <sgothel> so maybe for retirement :)
20140630 11:16:12 <eclesia> it doesn't bother me, I hve my lib :D
20140630 11:16:19 <sgothel> haha
20140630 11:16:41 <eclesia> well, maven downloading the internet might be a problem ...
20140630 11:16:57 <sgothel> right in the mud of Fukoshima's dirty laundy .. hmm
20140630 11:17:51 <sgothel> however, looks like a nice place .. will read about it later
20140630 11:19:41 <xranby> http://www.scmp.com/business/economy/article/1308540/us-tax-law-no-red-light-chinese-green-card-seekers "They're applying often unaware of Fatca and America's taxation on worldwide income for green card holders,"
20140630 11:21:13 <sgothel> haha .. you cannot seize anything in China if they hide there
20140630 11:22:00 <xranby> we will see ... http://www.china-briefing.com/news/2014/02/05/china-and-the-u-s-foreign-account-tax-compliance-act.html
20140630 11:23:06 <xranby> http://www.scmp.com/business/banking-finance/article/1542111/beijing-back-united-states-over-new-fatca-law-against-tax
20140630 11:23:10 <sgothel> but why should China agree to such bullshit, they *own* the US :)
20140630 11:23:36 <xranby> i think they do it if it will lead to currency revaluation
20140630 11:23:58 <xranby> 1 $ is worth 6 Yuan will become 1 Yuan is worth 6 $
20140630 11:24:12 <sgothel> that would be very bad for China
20140630 11:24:29 <xranby> china is buying up tons of gold
20140630 11:24:33 <sgothel> rising costs are already killing folks there
20140630 11:24:54 <sgothel> so a more slow currency adaption - leveling playfield - is needed
20140630 11:25:16 <xranby> the US national bank have collapsed two times in the past...
20140630 11:25:22 <xranby> i think we are on the brink of the third time
20140630 11:28:12 <sgothel> then .. it will be war
20140630 11:30:34 <sgothel> (thx to the east, they didn't pick up the bait - which was put there for provocation quite a few times in the last month/year)
20140630 11:33:33 <sgothel> China did an IGA, or will do .. i.e. share data of citizens .. as if they didn't do that for a long time. AFAIK Germany does same w/ US for a long time.
20140630 11:53:29 <eclesia> ho good news for me, webm is just an mkv with vp8 codec + vorbis, yeah less work to do
20140630 11:56:08 <rmk0> how is matroska?
20140630 11:56:18 <rmk0> as in... good? awful?
20140630 11:56:45 <rmk0> i've not heard anything bad about it, which is unusual
20140630 11:57:02 <eclesia> hm, I only have a draft of mkv, but it's a container format, like riff
20140630 11:57:05 <sgothel> mkv .. nice container (speaking as a user)
20140630 11:57:22 <sgothel> i.e. contains subtitles .. all languages .. etc
20140630 11:58:01 <rmk0> it seems like it supports everything
20140630 11:58:03 <eclesia> bad side of this container, you can't parse it without knwoing the structure
20140630 11:58:36 <eclesia> ebml (somekinf of binary xml used in mkv) isn't very friendly to understand
20140630 11:59:44 * rmk0 eyes ebml
20140630 12:00:11 <rmk0> "There is currently no equivalent to a DTD or Schema to define known elements for a document. But we plan on adding such a level."
20140630 12:00:18 <rmk0> massive red warning light switches on
20140630 12:00:45 <eclesia> what I was saying ^^
20140630 12:00:58 <rmk0> yeah...
20140630 12:01:12 <rmk0> you're required to know the structure, but there's no way to specify the structure
20140630 12:01:27 <rmk0> that should have been the first thing they produced...
20140630 12:01:48 <rmk0> oh well
20140630 12:02:09 <eclesia> riff isn't better
20140630 12:02:36 <eclesia> the best you can do is split the blocks and that's all
20140630 12:03:08 <eclesia> then you have to know block id --> bytes structure
20140630 12:58:52 <rmk0> sgothel: i've pushed the aliasing changes to http://github.com/io7m/gluegen and http://github.com/io7m/jogl, but i think they're going to need review/testing
20140630 12:59:17 <rmk0> on the master branch... didn't realize until too late that i'd done it on master, but i guess it's not a big problem
20140630 12:59:27 <sgothel> awesome, thank you - will check later/tomorrow
20140630 12:59:38 <sgothel> it's your master branch :)
20140630 12:59:46 <rmk0> all this does is remove the suffixes (-core, -android, etc), so jogl-core.jar becomes jogl.jar
20140630 12:59:55 <rmk0> and makes the build produce natives for each
20140630 13:00:29 <sgothel> you mean for finding the native jar name .. not really remove jogl.jar ? :)
20140630 13:00:44 <rmk0> yeah, just for finding names
20140630 13:00:53 <sgothel> great
20140630 13:00:54 <rmk0> it doesn't produce fewer jars or anything
20140630 13:01:07 <rmk0> just two of them were renamed, due to the lack of suffixes
20140630 13:01:21 <rmk0> jogl-core.jar → jogl.jar, nativewindow-core.jar → nativewindow.jar
20140630 13:02:00 <sgothel> why was that necessary ? i.e. core was for atomic semantics only ..
20140630 13:02:10 <sgothel> ah
20140630 13:02:26 <sgothel> get it .. atomic only .. reuse default native jar file .. good
20140630 13:02:52 <rmk0> there may be consequences i've not thought of! did test it and didn't see issues, but you know how that goes...
20140630 13:03:07 <sgothel> sure sure .. it's not all defined as an DTD :)
20140630 13:03:25 <sgothel> organic we are .. hehe
20140630 13:03:31 <rmk0> if the aliasing removal works, i'll move on to the classloader stuff
20140630 13:04:01 <rmk0> would like to get maven and gradle test builds set up after all of this
20140630 13:04:02 * odin_ (~Odin@anon) has joined #jogamp
20140630 13:04:16 <rmk0> like, projects set up that try to compile against jogl using every possible combination of packages
20140630 13:04:19 <sgothel> you could try it w/ jogl-noawt w/ maven ..
20140630 13:04:22 <rmk0> make sure they all work
20140630 13:05:12 <rmk0> yeah... reminds me i need to edit the maven deployment scripts to use the new natives and so on
20140630 13:05:37 <sgothel> I am fine w/ removal of '-core'
20140630 13:05:52 <sgothel> you will have review/merge by tomorrow
20140630 13:06:01 <rmk0> take your time, i've got lots to do here
20140630 13:12:31 * rmk0 spots issue with gluegen change
20140630 14:13:27 * hharrison (~chatzilla@anon) has joined #jogamp
20140630 14:14:05 <hharrison> eclesia: we wrote our own VP8 encoder in java to record screencasts of our sim program
20140630 14:14:18 <hharrison> It's rudimentary, but works well enough for us
20140630 14:15:01 <hharrison> In case you're interested:
20140630 14:15:02 <hharrison> https://github.com/AusencoSimulation/JaamSim/tree/master/src/main/java/com/jaamsim/video/vp8
20140630 14:15:06 <eclesia> hharrison: nice, ... is it private work ?
20140630 14:15:56 <hharrison> All that stuff is written by my coworker, he really enjoys the video encode/decode stuff
20140630 14:16:21 <eclesia> "Copyright (C) 2012 Ausenco Engineering Canada Inc." :(
20140630 14:16:28 <hharrison> all the jaamsim stuff is GPL3
20140630 14:17:06 <hharrison> If people asked nicely, and had a real use case, probably I could get things more permissively licensed...but sorry, not going to be public domain
20140630 14:17:18 <hharrison> Best I think I could convince people of here would be Apache2
20140630 14:17:26 <sgothel> @Harvey: As w/ other modules .. if you could dual license it (BSD) we could add it as a decoder / encoder for our AV stuff :)
20140630 14:17:33 <sgothel> yes APL2 would be great
20140630 14:17:34 <hharrison> Company lawyers love waiver of liability
20140630 14:17:55 <hharrison> Currently we only have an encoder
20140630 14:18:07 <hharrison> We've actually been waiting for a real VP9 spec to appear
20140630 14:18:28 <sgothel> https://github.com/AusencoSimulation/JaamSim/blob/master/src/main/java/com/jaamsim/video/vp8/Decoder.java ?
20140630 14:18:30 <eclesia> I write decoders mostly, that's why your encoder was interesting ^^
20140630 14:18:34 <hharrison> We just don't feel like reading the libvpx code to figure it out, we want a real documented bitstream
20140630 14:19:13 <hharrison> Oh, we have a docoder, but it is really just to test the encoder, I don't think it is in a shape you'd want to really use it
20140630 14:19:28 <hharrison> Can't remember where we landed on that ;-D
20140630 14:19:57 <eclesia> I will have to write mine anyway because of the license, so I better not even have a look at yours to avoid troubles.
20140630 14:20:31 <hharrison> Just don't ship our code directly, feel free to be inspired by it though
20140630 14:20:52 <hharrison> It's not a huge amount of work as it turns out
20140630 14:21:23 <eclesia> perhaps if I am blocked for a week on something, i'll take a peek ^^
20140630 14:21:33 <hharrison> Wish I could convince my coworker to hang out here, hopefully after I drag him to siggraph he'll know people here
20140630 14:22:02 <hharrison> the vp8 bitstream spec was actually not bad (except for some frustrating omissions)
20140630 14:22:27 <hharrison> Really hoping VP9 learns something from how Opus was done, _you need a damn spec_
20140630 14:24:52 * eclesia has not started the opus decoder yet
20140630 14:30:42 <hharrison> rmk0: looking at the native jars lookup, is the intention that we'll end up with a single natives jar per package?
20140630 14:30:53 <hharrison> gluegen-natives.jar, jogl-natives.jar...etc?
20140630 14:31:17 <rmk0> nope, will still be jogl-natives-linux-amd64.jar, etc
20140630 14:31:47 <hharrison> right, just not jogl-all-natives, jogl-noawt-natives....etc
20140630 14:32:30 <rmk0> http://waste.io7m.com/2014/06/30/jars.txt
20140630 14:32:37 <rmk0> is what the build produces now, not including the atomics
20140630 14:33:42 <rmk0> http://waste.io7m.com/2014/06/30/atomics.txt
20140630 14:34:19 <rmk0> basically removed the "suffixes" from the addNativeLibraryJar, which was what allowed the jogl-all-noawt.jar to use the jogl-all natives
20140630 14:35:29 <rmk0> is why jogl-core.jar was renamed to jogl.jar, because "-core" was implemented via suffixes
20140630 14:36:00 <rmk0> not sure any of this is making sense... the original behaviour was a bit tricky to understand as it was
20140630 14:42:49 <hharrison> I guess I'm just trying to figure out why there is jogl-all-natives... and jogl-all-noawt-natives
20140630 14:44:38 <hharrison> could we get to a place where there is only jogl-natives.{os.arch}.jar
20140630 14:45:09 <sgothel> diff packages contain diff sets of native libs
20140630 14:45:34 <sgothel> the all targets/packages contain more .. reducing downloads, incr. convenience - i.e. usable sets
20140630 14:45:54 <sgothel> atomics are more for special interests .. embedded etc
20140630 15:36:46 * hharrison (~chatzilla@anon) Quit (Remote host closed the connection)
20140630 15:53:51 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20140630 16:01:15 * eclesia (~husky@anon) Quit (Quit: Leaving.)
20140630 16:11:02 * bbbruce (~bx@anon) Quit (Ping timeout: 245 seconds)
20140630 16:12:50 * bbbruce (~bx@anon) has joined #jogamp
20140630 16:53:23 * monsieur_max (~maxime@anon) has joined #jogamp
20140630 18:23:23 * [Mike] (~Mike]@anon) has joined #jogamp
20140630 18:39:32 * [Mike] (~Mike]@anon) Quit ()
20140630 18:40:21 * [Mike] (~Mike]@anon) has joined #jogamp
20140630 18:41:02 * ___m___ (~Mike]@anon) has joined #jogamp
20140630 18:45:00 * [Mike] (~Mike]@anon) Quit (Ping timeout: 255 seconds)
20140630 18:47:56 * ___m___ is now known as [Mike]
20140630 21:26:22 * ___m___ (~Mike]@anon) has joined #jogamp
20140630 21:27:44 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20140630 21:30:50 * zzuegg_g (zzuegg@anon) has joined #jogamp
20140630 21:31:22 * zzuegg (zzuegg@anon) Quit (Killed (rajaniemi.freenode.net (Nickname regained by services)))
20140630 21:31:22 * zzuegg_g is now known as zzuegg
20140630 21:35:56 * [Mike] (~Mike]@anon) Quit (Ping timeout: 240 seconds)
20140630 21:58:09 * hija (~hija@anon) has joined #jogamp
20140630 22:13:17 * hija (~hija@anon) has left #jogamp
20140630 22:59:59 * ___m___ is now known as [Mike]
20140701 01:27:45 * kermyt (~kermyt@anon) Quit (Ping timeout: 255 seconds)
20140701 01:30:40 * kermyt (~kermyt@anon) has joined #jogamp
20140701 04:20:55 * kermyt (~kermyt@anon) Quit (*.net *.split)
20140701 04:21:02 * odin_ (~Odin@anon) Quit (*.net *.split)
20140701 04:21:02 * jk4 (~jk4@anon) Quit (*.net *.split)
20140701 04:21:02 * magaio (~magaio@anon) Quit (*.net *.split)
20140701 04:21:02 * zubzub (~zubzub@anon) Quit (*.net *.split)
20140701 04:21:02 * xranby (~xranby@anon) Quit (*.net *.split)
20140701 04:21:03 * zzuegg (zzuegg@anon) Quit (*.net *.split)
20140701 04:21:08 * masterzen (~masterzen@anon) Quit (*.net *.split)
20140701 04:21:08 * bbbruce (~bx@anon) Quit (*.net *.split)
20140701 04:21:08 * [Mike] (~Mike]@anon) Quit (*.net *.split)
20140701 04:21:08 * sgothel (~sgothel@anon) Quit (*.net *.split)
20140701 04:23:24 * kermyt (~kermyt@anon) has joined #jogamp
20140701 04:23:25 * zzuegg (zzuegg@anon) has joined #jogamp
20140701 04:23:25 * [Mike] (~Mike]@anon) has joined #jogamp
20140701 04:23:25 * bbbruce (~bx@anon) has joined #jogamp
20140701 04:23:25 * odin_ (~Odin@anon) has joined #jogamp
20140701 04:23:25 * xranby (~xranby@anon) has joined #jogamp
20140701 04:23:25 * magaio (~magaio@anon) has joined #jogamp
20140701 04:23:25 * masterzen (~masterzen@anon) has joined #jogamp
20140701 04:23:25 * jk4 (~jk4@anon) has joined #jogamp
20140701 04:23:25 * zubzub (~zubzub@anon) has joined #jogamp
20140701 04:23:25 * sgothel (~sgothel@anon) has joined #jogamp
20140701 05:05:29 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20140701050529.html