#jogamp @ irc.freenode.net - 20140903 05:05:41 (UTC)
20140903 05:05:41 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20140902050541.html
20140903 05:05:41 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20140903050541.html
20140903 06:59:54 * jvanek (jvanek@anon) has joined #jogamp
20140903 07:28:51 * monsieur_max (~maxime@anon) has joined #jogamp
20140903 09:00:53 * xranby (~xranby@anon) Quit (Quit: Leaving.)
20140903 09:29:01 * xranby (~xranby@anon) has joined #jogamp
20140903 09:30:43 <xranby> I am happy today, JogAmp JOGL renders OpenGL 3 out of the box using Mesa nouveau Gallium driver on external DisplayLink based USB monitors!
20140903 09:31:04 <xranby> this is freesoftware driver synergy at its best
20140903 09:35:02 * rmk0 (~rmk0@anon) has joined #jogamp
20140903 09:35:02 * rmk0 (~rmk0@anon) Quit (Changing host)
20140903 09:35:02 * rmk0 (~rmk0@anon) has joined #jogamp
20140903 09:35:08 <xranby> https://twitter.com/xranby/status/507099354113777664/photo/1
20140903 09:38:47 <bbbruce> nice!
20140903 11:49:02 * xranby (~xranby@anon) Quit (Remote host closed the connection)
20140903 12:09:43 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20140903 12:10:07 * jvanek (jvanek@anon) has joined #jogamp
20140903 12:16:49 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20140903 12:19:31 * doev (~doev@anon) has joined #jogamp
20140903 12:20:43 * doev (~doev@anon) Quit (Read error: Connection reset by peer)
20140903 12:56:00 * jvanek (jvanek@anon) has joined #jogamp
20140903 13:23:24 <sgothel> Hi Mark .. hope all is great.
20140903 13:23:50 <sgothel> Hi Xerxes .. nice, how works the GPU/USB blitting ?
20140903 13:25:49 <sgothel> ok .. another round of URI encoding .. to accept all the special chars in the path like ?!.. and the like
20140903 13:43:54 <zzuegg> wait, does that mean you can have a fully transparent newt window?
20140903 13:47:55 <sgothel> zzueg: ofc .. look at our applet demos ..
20140903 13:48:05 <sgothel> there is one .. hanging over browser window
20140903 13:53:35 <sgothel> http://jogamp.org/deployment/jogamp-current/jogl-test-applets.html <- 1st column, 3rd row :)
20140903 13:54:56 <zzuegg> ah, you where faster then me. so its possible only in applets, or also with a newt window? would make a great intro :)
20140903 13:55:13 <sgothel> it's a NEWT demo
20140903 13:55:53 <sgothel> com.jogamp.opengl.test.junit.jogl.demos.es2.newt.TestGearsES2NEWT <- look here, main arguments
20140903 13:56:22 <zzuegg> nice, thanks a lot
20140903 14:03:57 * xranby (~xranby@anon) has joined #jogamp
20140903 14:04:37 <xranby> bbbruce: the USB monitors appears as extra DVI outputs for the Nvidia cards when using the nuveau driver
20140903 14:04:39 <xranby> DVI-2-1 connected 800x480+1920+0 160mm x 230mm
20140903 14:04:39 <xranby> 800x480 59.5*+
20140903 14:04:39 <xranby> DVI-1-2 connected 800x480+1920+524 160mm x 230mm
20140903 14:04:39 <xranby> 800x480 59.5*+
20140903 14:04:46 <xranby> like that.
20140903 14:05:02 <sgothel> (03:23:50 PM) sgothel: Hi Xerxes .. nice, how works the GPU/USB blitting ?
20140903 14:05:25 <xranby> sgothel: 30 fps https://twitter.com/xranby/status/507099354113777664/photo/1
20140903 14:05:37 <sgothel> technically ?
20140903 14:05:45 <xranby> sgothel: all magic happens using the linux kernel + xorg + mesa
20140903 14:06:01 <xranby> when using the freesoftware drivers
20140903 14:06:05 <sgothel> so who implements the USB protocol for the display ?
20140903 14:06:13 <xranby> nvidia card + usb monitors
20140903 14:06:26 <sgothel> I assume: frame -> special usb protocol -> usb-display ?
20140903 14:06:44 <sgothel> is there a standard protocol now ?
20140903 14:06:56 <sgothel> whats its name ?
20140903 14:07:48 <xranby> the linux kernel rovide
20140903 14:07:49 <xranby> udl 24545 2
20140903 14:07:49 <xranby> drm_usb 13134 1 udl
20140903 14:07:49 <xranby> syscopyarea 12529 1 udl
20140903 14:07:49 <xranby> sysfillrect 12701 1 udl
20140903 14:07:49 <xranby> sysimgblt 12640 1 udl
20140903 14:07:57 <xranby> provide
20140903 14:08:29 <sgothel> so, there is now a usb/display protocol standard I assume ?
20140903 14:08:45 <xranby> yes its the standard linux kernel driver api
20140903 14:08:47 <sgothel> a product list .. how to search for this new protocol ..
20140903 14:08:58 <sgothel> that is not a standard :)
20140903 14:08:59 <xranby> drm 303102 11 udl,ttm,drm_usb,drm_kms_helper,nouveau
20140903 14:09:19 <sgothel> (I mean usb protocol standard)
20140903 14:09:43 <sgothel> any references ?
20140903 14:10:21 <sgothel> or is it .. that somebody simply hacked that drm_usb for some types of such devices ?
20140903 14:10:43 <sgothel> I remember .. we had this chat .. and folks where doing it
20140903 14:11:07 <xranby> yes the story is that displaylink who make the chips never revealed the protocol
20140903 14:11:15 <xranby> then reverseenginneering took place
20140903 14:11:32 <sgothel> displaylink protocol over usb ?
20140903 14:12:00 <xranby> creating libdlo http://libdlo.freedesktop.org/wiki/
20140903 14:12:13 <sgothel> ofc .. using such protocols (at least in hardware) requires license fees AFAIK
20140903 14:13:16 <sgothel> 30fps @ resolution x color-depth ?
20140903 14:14:02 <sgothel> only nouveau, or any drm driver (amd, intel, ..) ?
20140903 14:14:15 <xranby> i have only tested nouveau
20140903 14:14:46 <xranby> right now i have two 800x480 screens connected
20140903 14:15:06 <sgothel> cool, DL over USB .. nice performance .. USB2 I assume
20140903 14:15:28 <xranby> USB2
20140903 14:15:35 <sgothel> well, there will be a lag I guess, but this is not for gamers I guess :)
20140903 14:15:53 <sgothel> great
20140903 14:16:10 <xranby> trying to find out the colordepth
20140903 14:16:15 <sgothel> multiseat .. let me see the price tags .. @ http://displaylink.com/shop
20140903 14:16:17 <xranby> i can only find properties of the resolution
20140903 14:16:21 <xranby> but it looks to be at least 16bit
20140903 14:17:04 * xranby trying to query xrandr
20140903 14:18:15 <xranby> sgothel: the cool part was that it simply worked in a natural way
20140903 14:18:30 <xranby> makes it look like your primary GPU has more DVI outputs
20140903 14:18:35 <xranby> all exposed by xrandr
20140903 14:18:51 <xranby> and hw acceleration and screensharing worked
20140903 14:19:15 <xranby> no extra code on jogamps side to use it
20140903 14:20:16 <sgothel> well, I see prices around 300 USD .. i.e. close to a 'normal' DVI/DL device
20140903 14:20:52 <xranby> these small screens i buy costs around 50$ each
20140903 14:21:53 <xranby> so.. yay.. i can use jogamp in our modular instrument now
20140903 14:22:40 <sgothel> shop for those $50 devices please ?
20140903 14:23:26 <xranby> we buy them in boxes of 10 each from the lilliput factory in china
20140903 14:23:45 <sgothel> I guess I could buy 10 :)
20140903 14:23:54 <sgothel> URL ?
20140903 14:24:36 <sgothel> friend complains about expensive small monitors .. for hardware monitoring etc (heating controller etc), so this could be it
20140903 14:25:11 <sgothel> http://lilliput.cn/ ?
20140903 14:25:21 <rmk0> lo
20140903 14:25:44 <sgothel> hi Mark .. already was concerned you are not here (since you are always here) :)
20140903 14:25:48 * rmk0_ (~rmk0@anon) has joined #jogamp
20140903 14:25:56 <xranby> sgothel: yes, i did a double check of the price.. we pay $75 for single samples
20140903 14:25:59 * rmk0_ is now known as rmk0-alt
20140903 14:26:06 * rmk0-alt (~rmk0@anon) Quit (Changing host)
20140903 14:26:06 * rmk0-alt (~rmk0@anon) has joined #jogamp
20140903 14:26:10 <xranby> of the UM-70
20140903 14:26:36 <rmk0> recovering from a bit of food poisoning
20140903 14:26:43 <rmk0> wasn't too bad, only lasted a day or so
20140903 14:26:53 <rmk0> spent one night lying flat on the floor, gripping the floor in case i fell off
20140903 14:26:55 <sgothel> puhh ..
20140903 14:27:23 <rmk0> seems completely gone now, but i'm mainly recovering from not really eating for a few days
20140903 14:27:26 <rmk0> body doesn't like that!
20140903 14:27:42 <sgothel> good you can eat again
20140903 14:28:03 <sgothel> @Xerxes: cannot find um-70 .. all those javascript requirements .. grrr
20140903 14:28:18 <xranby> http://lilliput.cn/products/detail/UM70/
20140903 14:28:20 <sgothel> found it ..
20140903 14:28:23 <sgothel> :)
20140903 14:29:45 <sgothel> UM-1010 .. also interesting (touch pad)
20140903 14:30:38 <sgothel> hmm .. if you already order them in bulk .. can buy a few from you ? I miss the 'buy now' button :)
20140903 14:31:24 <xranby> sgothel: sure.. i will bring some with me to the next bof
20140903 14:31:36 <xranby> i have plenty of them here
20140903 14:32:13 <sgothel> you simply wrote them by email .. and then need to negotiate price ?
20140903 14:32:23 <sgothel> thanx for that offer!
20140903 14:33:12 <xranby> yes you simply write them a letter
20140903 14:33:15 <xranby> email
20140903 15:05:12 * jvanek (jvanek@anon) Quit (Ping timeout: 245 seconds)
20140903 15:07:43 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=908 re-opened
20140903 15:08:11 <sgothel> now, need to find a good URI 'native file path character' encoder/decoder, to drop our IOUtil hacks ..
20140903 15:17:29 * jvanek (jvanek@anon) has joined #jogamp
20140903 15:30:57 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20140903 15:54:04 * jvanek (jvanek@anon) has joined #jogamp
20140903 15:57:26 * jvanek (jvanek@anon) Quit (Client Quit)
20140903 16:02:13 * monsieur_max1 (~maxime@anon) has joined #jogamp
20140903 16:04:48 * rmk0-alt_ (~rmk0@anon) has joined #jogamp
20140903 16:08:30 * masterzen (~masterzen@anon) Quit (Ping timeout: 260 seconds)
20140903 16:08:33 * rmk0-alt (~rmk0@anon) Quit (Ping timeout: 260 seconds)
20140903 16:08:36 * monsieur_max (~maxime@anon) Quit (Ping timeout: 260 seconds)
20140903 16:08:50 * rmk0-alt_ is now known as rmk0-alt
20140903 16:08:57 * masterzen (~masterzen@anon) has joined #jogamp
20140903 16:08:57 * rmk0-alt (~rmk0@anon) Quit (Changing host)
20140903 16:08:57 * rmk0-alt (~rmk0@anon) has joined #jogamp
20140903 16:12:37 <rmk0> sgothel: re bug 1046: i don't disagree the jar file is completely mangled, but should it crash the whole jvm?
20140903 16:12:41 <rmk0> i mean... is that ... expected?
20140903 16:13:03 <rmk0> the crash occurs when loading a library that's not part of jogl, too...
20140903 16:13:09 <sgothel> @Mark: yes .. since loading wrong native libs .. well
20140903 16:13:16 <sgothel> not our domain ..
20140903 16:13:27 <sgothel> i.e. can't deal w/ wrong native libs here ..
20140903 16:13:30 <rmk0> when you say the wrong natives, which do you mean?
20140903 16:14:07 <sgothel> I dunno .. I stopped right there, but it seems not to be the right thing guaranteed
20140903 16:14:24 <sgothel> i.e. if loading the a wrong native lib .. anything may happen
20140903 16:14:36 <rmk0> does loadLibrary not do any better than a segfault if it tries to load a library that's of the wrong architecture?
20140903 16:14:46 <sgothel> good question
20140903 16:14:59 <sgothel> I never tested this thoroughly ..
20140903 16:15:04 <rmk0> i'd assume it'd fail immediately, rather than doing what it appeared to do - silently fail to load a library and then crash when loading something else (that's not even part of jogl)
20140903 16:15:15 <rmk0> is sort of why i put the bug there in the first place... couldn't work out who's fault it was
20140903 16:15:48 <sgothel> well, in such case, we need a dedicated test case and more narrow question IMHO :)
20140903 16:15:54 <rmk0> hehe
20140903 16:16:06 <rmk0> the jar file is definitely completely wrong, no doubt about that, but i'd assume better than a jvm segfault
20140903 16:16:19 <sgothel> I was working down the bug reports, and stopped as early as possible :)
20140903 16:16:24 <rmk0> hehe, yes
20140903 16:16:33 <rmk0> will try to do a bit of testing later... wasn't even my jar file
20140903 16:16:42 <rmk0> someone came in trying to make a onejar type setup, before they saw the wiki docs
20140903 16:16:49 <sgothel> we have an elf tester .. but I don't like to probe any native lib there
20140903 16:16:49 <rmk0> crashed their jvm and mine
20140903 16:17:08 <rmk0> i'd not
20140903 16:17:17 <sgothel> then pls make a new bug report, w/ recipe .. etc
20140903 16:17:30 <rmk0> the dlopen call should definitely fail for elf binaries that it doesn't understand
20140903 16:17:36 <sgothel> yup
20140903 16:18:07 <sgothel> it is just that I don't have a 'proper' test case, and no time to test all variations myself, since I don't use fat jars
20140903 16:18:14 <rmk0> right, me neither
20140903 16:18:15 <sgothel> but having one .. I will work on it .. sure
20140903 16:18:23 <rmk0> will try and construct some things
20140903 16:18:39 <rmk0> currently approaching the blender people to see what needs to be done to improve collada support
20140903 16:18:44 <sgothel> I dislike them, due to the 'non officially signed' and reproducible .. manually mixed up content .. blablab :)
20140903 16:18:51 <rmk0> yeah, i don't like them either
20140903 16:19:02 <rmk0> they complicate things... make it hard to know what you've actually got
20140903 16:19:12 <rmk0> prefer a directory full of jars that can be checked individually
20140903 16:19:31 <rmk0> easier for people to get the program running if you failed to include some sort of frontend for their weird platform, too
20140903 16:19:34 <sgothel> what we could do .. is producing fat jar files w/ certain reasonable configurations .. but well
20140903 16:32:51 <rmk0-alt> i'd not bother, personally
20140903 16:33:02 <rmk0-alt> what's really needed is for people to read the rules...
20140903 16:33:16 <rmk0-alt> everyone's going to use a different type of onejar system (there were at least four last time i checked)
20140903 16:36:54 * hharrison (~chatzilla@anon) has joined #jogamp
20140903 16:39:51 <hharrison> sgothel: semver actually tests for the interned String thing?
20140903 17:43:12 * rmk0 (~rmk0@anon) Quit (Read error: Connection reset by peer)
20140903 17:45:11 <sgothel> @Harvey: Yuheee .. yes, it compares compatibility on bytecode level
20140903 17:45:28 <sgothel> @Mark: ^^ Actually our 1st success story :)
20140903 17:46:03 <sgothel> https://jogamp.org/chuck/view/fwd/job/jogl/1336/label=linux-x86_64-amd/testReport/com.jogamp.opengl.test.junit.jogl.acore/TestVersionSemanticsNOUI/testVersionLatest/
20140903 17:46:57 <sgothel> Changed Field GLES3, desc Ljava/lang/String;, access
20140903 17:47:05 <rmk0-alt> \o/
20140903 17:47:08 <sgothel> (not so detailed .. but was a great hint)
20140903 17:48:29 <hharrison> neat
20140903 17:51:57 <rmk0-alt> hm... that's a point
20140903 17:52:11 <rmk0-alt> is the output of the differences between two versions public?
20140903 17:52:26 <rmk0-alt> would be amusing to have a small changelog that lists differences in that way
20140903 17:52:37 <rmk0-alt> i mean... as part of the usual release notes
20140903 17:55:55 <sgothel> yes ..
20140903 17:56:21 <sgothel> http://jogamp.org/deployment/v2.2.0/archive/ChangeLogs/
20140903 17:56:22 <sgothel> ?
20140903 17:56:39 <sgothel> jogamp stuff you mean ?
20140903 17:56:55 <rmk0-alt> nice... didn't know that existed
20140903 17:57:08 <rmk0-alt> er
20140903 17:57:12 <sgothel> https://jogamp.org/wiki/index.php/SW_Tracking_Report_Objectives_for_the_release_2.2.1 <- well, that the comprehensive one IMHO :)
20140903 17:57:39 <rmk0-alt> meant the semver data, but i suppose if the changes are already documented in "human" to this extent...
20140903 17:57:47 <sgothel> however .. all trackable, git-sha1 <-> bugzilla ..
20140903 17:58:07 <sgothel> oh .. more comprehensive semver data ?
20140903 17:58:19 <sgothel> well, the semver test result is deployed, yes
20140903 17:58:32 <sgothel> http://jogamp.org/deployment/v2.2.0/archive/test-results/
20140903 17:58:43 <sgothel> we don't hide a thing :)
20140903 17:59:29 <rmk0-alt> think i did a poor job of explaining what i meant
20140903 17:59:44 <sgothel> sorry .. me esl :)
20140903 17:59:47 <rmk0-alt> i mean like... 2.2.1 comes out... where do i look to see all the changes in terms of the semver tests?
20140903 18:00:08 <sgothel> either here: http://jogamp.org/deployment/v2.2.0/archive/test-results/
20140903 18:00:28 <sgothel> or you go from: http://jogamp.org/wiki/index.php/Release_2.2.0
20140903 18:00:53 <sgothel> to https://jogamp.org/chuck/view/fwd/job/jogl/1329/ (Build information)
20140903 18:00:59 <sgothel> and see test results online
20140903 18:01:59 <sgothel> i.e. it's a unit test .. maybe we should also write the results to one simple file - good idea!
20140903 18:22:14 * rmk0 (~rmk0@anon) has joined #jogamp
20140903 18:22:14 * rmk0 (~rmk0@anon) Quit (Changing host)
20140903 18:22:14 * rmk0 (~rmk0@anon) has joined #jogamp
20140903 18:27:15 <sgothel> @Mark: ok ?
20140903 18:27:39 <rmk0> weird stuff happening with my isp connection...
20140903 18:28:18 <sgothel> b/c you are .*\.fast\..* now ? :)
20140903 18:28:25 <rmk0> hehe
20140903 18:28:34 <sgothel> as in *.in-addr.fast.co.uk' :)
20140903 18:28:48 <rmk0> my isp seems to be having problems with their provider
20140903 18:29:23 <sgothel> now having 2 uplinks here w/ load balancing .. bad thing: same provider :-/
20140903 18:30:36 <sgothel> oh .. since dyn.com is now charging for DNS requests .. (we are above a million per month etc) .. I need to push a zone file to our server ISP ..
20140903 18:30:55 <sgothel> when I do this this month .. I probably need a bit of help / validation, so nothing gets cut off .. ahem
20140903 18:31:25 <sgothel> dyn.com are like blackmailing for their services now .. need to get away from 'em
20140903 18:31:59 <sgothel> @Mark: You know this stuff as well right ? .. or whoever reads this here.
20140903 18:32:24 <sgothel> i.e. hetzner (server ISP) allows pushing a zone file .. so need to switch
20140903 18:32:25 <rmk0> hm, have had a bit of experience with it
20140903 18:32:33 <rmk0> not doing that, specifically, but managing dns
20140903 18:32:38 <sgothel> great, I have mine here too .. but well
20140903 18:32:49 <sgothel> always to have a look at things ..
20140903 18:32:58 <sgothel> we can play w/ jausoft.com first though :)
20140903 18:33:19 <sgothel> after 2.2.1 ofc
20140903 19:01:24 * void256 (~void@anon) has joined #jogamp
20140903 19:30:37 * gouessej (5279c957@anon) has joined #jogamp
20140903 19:30:47 <gouessej> Hi
20140903 19:31:12 <gouessej> rmk0: Have you seen that? http://jogamp.org/log/irc/jogamp_20140902050541.html#l128
20140903 19:32:56 * rmk0 eyes it
20140903 19:34:22 <rmk0> hm, yeah the issue was that there are two forms of ssh "transport" in maven
20140903 19:34:44 <rmk0> we needed maven to be able to talk to the ssh-key-agent process so that it didn't need to know private key passwords and so on
20140903 19:35:32 <rmk0> think there are two things you need to do, let me check...
20140903 19:36:59 <rmk0> ah, yeah, you need to add the ssh "wagon" to the pom:
20140903 19:37:01 <rmk0> http://waste.io7m.com/2014/09/03/wagon.txt
20140903 19:37:10 <rmk0> no idea why they picked such a stupid term... "wagon"
20140903 19:37:26 <gouessej> ok
20140903 19:37:28 <rmk0> and then instead of using:
20140903 19:37:34 <rmk0> http://ardor3d.com:8081/nexus/content/repositories/releases
20140903 19:37:38 <rmk0> you should be able to use:
20140903 19:38:00 <rmk0> er
20140903 19:38:04 <rmk0> well, the uri starts with:
20140903 19:38:06 <rmk0> scpexe://ardor3d.com/
20140903 19:38:12 <rmk0> but i'm not sure where you'd want to be copying files
20140903 19:38:29 <rmk0> it looks like you're running nexus... can you really just scp files into it?
20140903 19:38:47 <rmk0> i... assume you can, otherwise you'd probably not be asking
20140903 19:39:40 <rmk0> the "scpexe" is necessary to tell maven to use whatever ssh binary you have installed, because that's the only thing capable of talking to an existing ssh-key-agent process at the moment
20140903 19:40:01 <rmk0> using ssh:// won't work, for that reason, if you're using keys
20140903 19:40:21 <gouessej> I'd like to copy the file directly in jogamp.org/deployment/ardor3d
20140903 19:40:36 <rmk0> oh, right
20140903 19:40:43 <rmk0> should all be fine, then
20140903 19:41:19 <rmk0> think that'd be... scpexe://jogamp.org/srv/www/jogamp.org/deployment/ardor3d/
20140903 19:42:28 <gouessej> ok but what does scpexe mean?
20140903 19:43:04 <rmk0> that ssh wagon allows maven to speak ssh using either a pure java ssh library, or the system's existing ssh binary
20140903 19:43:17 <rmk0> ssh:// tells it to use the java library, scpexe:// tells it to use the ssh binary
20140903 19:43:28 <rmk0> but only the ssh binary is capable of talking to a key agent, so it's necessary if you're using keys
20140903 19:43:47 <rmk0> which... i assume you are, to log into jogamp.org?
20140903 19:43:53 <rmk0> seem to remember that being required
20140903 19:45:23 <gouessej> ok thanks
20140903 19:45:40 <rmk0> hope they implement the key agent communication soon
20140903 19:46:35 <rmk0> without it, you'd have to tell maven your private key passphrase
20140903 19:46:37 <rmk0> not ideal!
20140903 19:46:40 <gouessej> I hope they do it correctly. Do you think that you'll succeed with blender collada?
20140903 19:47:14 <rmk0> getting some promising responses
20140903 19:47:26 <rmk0> the main problem they have is knowing exactly how the animation data should actually appear in collada files
20140903 19:47:28 <rmk0> the spec is weird
20140903 19:47:38 <rmk0> is why it's not been implemented to date
20140903 19:47:47 <rmk0> there actually is a maintainer... joined when 2.66 came out
20140903 19:48:13 <rmk0> he says if i can put together some example files that show how animation should be represented, he'll try it
20140903 19:48:32 <rmk0> i'm going to get hold of 3ds max or one of those programs and try exporting some skeletal animations to see how they represent it
20140903 19:51:49 <gouessej> cool
20140903 19:52:41 <gouessej> I should check whether it still exports only the first frame when importing meshes with key frame animations
20140903 19:52:53 <rmk0> yeah...
20140903 19:53:40 <gouessej> but my scripts no longer work. I'll have to find an example that I can import without having to add another script
20140903 19:54:23 <rmk0> seems to still be the first frame here
20140903 19:54:44 <rmk0> as in... it just exports the bones in whatever pose they happen to be in at the first frame
20140903 19:54:59 <rmk0> i think we're dealing with the same issue, essentially... animation is really not supported at all
20140903 19:55:15 <gouessej> :(
20140903 19:55:36 <rmk0> well, is good at least... if i get my issue sorted out, you'll get yours sorted out automatically
20140903 19:58:22 <gouessej> I don't want people to be forced to use proprietary formats for interchange
20140903 19:58:35 <rmk0> me neither
20140903 19:59:00 <rmk0> i'm distressed that people have allowed the situation to get this bad
20140903 19:59:19 <rmk0> you'd think an open source 3D system would put a priority on getting data out of it in open formats
20140903 19:59:46 <rmk0> there's a lot of hatred for collada around
20140903 20:00:29 <rmk0> it is pretty complex and strange
20140903 20:00:46 <gouessej> I'm sad to see some efforts to support FBX in JMonkeyEngine
20140903 20:00:56 <gouessej> just to be "convenient"
20140903 20:01:48 <rmk0> autodesk have no excuse for not releasing format documentation... is blatantly hostile lock-in tactics
20140903 20:02:09 <hharrison> We ended up 'inventing' our own 'jsm' format to package the animator information in jaamsim
20140903 20:02:26 <rmk0> where did the animator information come from?
20140903 20:02:36 <hharrison> blender
20140903 20:02:41 <rmk0> ...
20140903 20:02:46 <rmk0> how did you get it out of blender?
20140903 20:02:52 <hharrison> had to write a python plugin to export it the way we wanted
20140903 20:03:01 * rmk0 chokes on blood and feathres
20140903 20:03:03 <rmk0> ... feathers
20140903 20:03:07 <hharrison> should all be in the jaamsim repo
20140903 20:03:24 <rmk0> how many times have you had to rewrite the plugin?
20140903 20:03:33 <hharrison> .....excpet now I notice I didn't push out the plugin in the repo (todo item added)
20140903 20:03:38 <hharrison> 0
20140903 20:03:42 <rmk0> !
20140903 20:03:51 <hharrison> python plugin using the internal API
20140903 20:03:56 <hharrison> not a C plugin
20140903 20:03:59 <rmk0> bizarre... maybe the data structures they use for that are more stable
20140903 20:04:17 <hharrison> We don't look at the data structures, just the API calls
20140903 20:04:18 <rmk0> i had to rewrite mine three times, once for each minor blender update, and all that exported was mesh data
20140903 20:04:36 <rmk0> hm... maybe that was the difference... i had to actually inspect meshes
20140903 20:04:39 <hharrison> As long as you hamstirng yourself to what the API provides, you are OK
20140903 20:04:51 <rmk0> i thought i did!
20140903 20:05:27 <hharrison> I will get that plugin out to the opensource side this week
20140903 20:05:32 <rmk0> \o/
20140903 20:05:38 <rmk0> would really like to take a look at that
20140903 20:05:48 <rmk0> i'd still like them to fix collada, but it might save me a lot of work if they don't
20140903 20:05:51 <hharrison> Wouldn't celebrate too hard, it's pretty bespoke to our needs
20140903 20:06:00 <rmk0> that's fine, it's more the API usage i'd care about
20140903 20:06:05 <rmk0> i'd be defining my own formats anyway
20140903 20:06:24 <rmk0> i just want the bone transforms for each keyframe, really
20140903 20:06:24 <hharrison> Maybe jsm interests you ;-)
20140903 20:06:36 <hharrison> That's pretty much all we added
20140903 20:06:46 <hharrison> and a binding to a timeline
20140903 20:06:53 <rmk0> where's jsm?
20140903 20:06:59 <gouessej> Is there a single bug report for the bug about animation data not exported in Blender?
20140903 20:07:14 <hharrison> inside the jaamsim project
20140903 20:07:25 <hharrison> It's.....purpose built
20140903 20:07:32 <hharrison> not exactly reusable at the moment
20140903 20:07:57 <hharrison> but the format was pretty spartan, mesh/vertex/transfom
20140903 20:08:20 <rmk0> gouessej: not yet, will likely create one in order to give the developer i spoke to information
20140903 20:08:42 <rmk0> hharrison: yeah, the engine i've developed just wants triangulated meshes and the bones
20140903 20:08:59 <rmk0> anything regarding materials and so on is defined elsewhere
20140903 20:09:06 <gouessej> rmk0: Let me know when it is done. I'll have to mention it in my tutorial to explain some implementation choices
20140903 20:09:12 <rmk0> right
20140903 20:25:33 * monsieur_max (~maxime@anon) has joined #jogamp
20140903 20:27:38 * gouessej (5279c957@anon) Quit (Quit: Page closed)
20140903 20:49:02 <hharrison> sgothel: is there any way to make jogl reliably explode if you call methods from a context it does not expect?
20140903 20:49:35 <hharrison> specifically, can you make jogl blow up if you call certain methods when you are not on the AWT EDT?
20140903 20:49:49 <hharrison> instead of just testing for it and silently failing?
20140903 20:50:32 <sgothel> scratch head ..
20140903 20:51:27 <sgothel> let me see what we do have in our magic cookie box :)
20140903 20:51:55 <rmk0> i'd put money on cookies
20140903 20:51:57 <hharrison> Oh, dammit, it's even worse than that, this isn't even jogl code, completely not your fault
20140903 20:52:24 <sgothel> still .. I like an optional constraint checker ..
20140903 20:52:26 <sgothel> good one
20140903 20:52:27 <hharrison> But...the feature request still stands I guess jogl.blowupifIsurpriseyou
20140903 20:52:38 <sgothel> @Mark probably will fall in love w/ it :)
20140903 20:52:48 <rmk0> \(^_^)/
20140903 20:53:08 <hharrison> jogl.dontworkaroundmybugs
20140903 20:53:12 <hharrison> ;-)
20140903 20:53:21 <sgothel> http://jogamp.org/deployment/archive/master/gluegen_817-joal_559-jogl_1330-jocl_1001/javadoc/jogl/javadoc/javax/media/opengl/GLAutoDrawable.html#isThreadGLCapable%28%29
20140903 20:53:36 <sgothel> so this is a new method .. doing some validation
20140903 20:54:28 <sgothel> now we just need to add a constraint SPI in .. ? GLContext ? or ? GLAutoDrawable ? .. I would guess .. err .. both ?
20140903 20:55:15 <sgothel> IMHO simple as 'interface Valildator { void validation() throws RuntimeException; } ' ?
20140903 20:55:36 <sgothel> *void validate(GLContext ctx)*
20140903 20:56:03 <sgothel> invoked .. err .. at makeCurrent() I guess
20140903 20:56:27 <sgothel> before* and/or after actual makeCurrent() .. hmm
20140903 20:56:41 <sgothel> *before: after locking, but before GL makeCurrent ..
20140903 20:56:44 <sgothel> makes sense ?
20140903 20:57:03 <sgothel> then you can throw your exception ..
20140903 21:00:49 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20140903 21:01:20 <sgothel> tsts .. @myself: add those '@since' tags in API ..
20140903 21:03:40 <hharrison> Yeah, that seems reasonable
20140903 21:04:07 <sgothel> ok .. will add bug report / and impl. it later
20140903 21:04:25 <sgothel> better API names ?
20140903 21:04:51 <hharrison> Always best if JOGL can tell us when we get it wrong
20140903 21:05:03 <sgothel> like .. validateMakeCurrentPre(), validateMakeCurrentPost(), ..
20140903 21:05:15 <sgothel> no .. here you would need to plug it in ..
20140903 21:05:36 <sgothel> i.e. make your own validation and let if fail if you wish so ..
20140903 21:05:56 <sgothel> usually we do validate things .. and fail fast, exceptions are fwd now
20140903 21:07:38 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=1039#c8 <- GLAnimatorControl.UncaughtExceptionHandler etc ..
20140903 21:08:05 <sgothel> pre and post validation as described above ?
20140903 21:08:32 <hharrison> Yeah, it wouldn't have helped us here I guess, this was purely user error
20140903 21:08:36 <sgothel> i.e. post would only be required, if validating GL states etc .. maybe good
20140903 21:09:11 <sgothel> pre would be nice to test surface constraints .. etc, all application logic ofc .. sort of
20140903 21:09:53 <hharrison> I have no idea why only OSX would show this...but calling Frame.setvisible, not from the AWT thread basically only works once
20140903 21:10:14 <hharrison> everything else is just broken.....but how the hell did this work for so long everywhere else....mysterious
20140903 21:15:33 <hharrison> Oh well, at least I can only blame myself
20140903 21:17:15 <sgothel> the weirdo AWT spec says ..
20140903 21:17:25 <sgothel> the 1st setVisible call can be done on !EDT
20140903 21:17:43 <sgothel> from there on .. all subsequent mutable stuff shall be done on AWT-EDT
20140903 21:17:57 <sgothel> however .. our unit tests do it all on AWT-EDT ..
20140903 21:35:05 <sgothel> wow .. 1st spam received via a github comment on a commit ..
20140903 22:01:48 * zzuegg is now known as zzuegg|afk
20140903 22:05:43 * void256 (~void@anon) Quit (Ping timeout: 272 seconds)
20140903 23:39:01 <hharrison> Well, with the OSX stuff working, I think jaamsim will be moving to allowing Java7 features
20140903 23:54:30 * hharrison (~chatzilla@anon) Quit (Quit: ChatZilla 0.9.90.1 [Firefox 32.0/20140825202822])
20140904 00:11:40 * kermyt_ (~kermyt@anon) has joined #jogamp
20140904 00:12:16 * kermyt (~kermyt@anon) Quit (Ping timeout: 255 seconds)
20140904 00:12:17 * kermyt_ is now known as kermyt
20140904 01:32:09 * rmk0 (~rmk0@anon) Quit (Ping timeout: 260 seconds)
20140904 01:32:47 * rmk0 (~rmk0@anon) has joined #jogamp
20140904 02:24:00 * kermyt (~kermyt@anon) Quit (Ping timeout: 252 seconds)
20140904 02:26:23 * kermyt (~kermyt@anon) has joined #jogamp
20140904 05:05:41 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20140904050541.html