#jogamp @ irc.freenode.net - 20141007 05:05:54 (UTC)
20141007 05:05:54 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20141006050554.html
20141007 05:05:54 -jogamp- This channel is logged @ http://jogamp.org/log/irc/jogamp_20141007050554.html
20141007 06:33:10 * eclesia (~husky@anon) has joined #jogamp
20141007 06:33:22 <eclesia> good underwater morning
20141007 06:50:53 * monsieur_max (~maxime@anon) has joined #jogamp
20141007 06:53:11 <eclesia> hi monsieur_max, had a good swim last night ?
20141007 06:55:32 <monsieur_max> haha :) my cats did not appreciate the show :)
20141007 07:00:15 <eclesia> I barely slept 3hours, ... the water pomp was running all night in the underground parking.
20141007 07:00:30 <eclesia> pump*
20141007 07:09:31 * doev (~doev@anon) has joined #jogamp
20141007 07:15:05 <monsieur_max> eclesia: ah, must have remember you ... last week ...
20141007 07:16:17 <eclesia> monsieur_max: just us 2 from montpellier and you still forget :p
20141007 07:17:31 <xranby> goodmorning,
20141007 07:17:49 <xranby> eclesia: monsieur_max: did you keep all stuff in the basement on pallets?
20141007 07:18:54 <xranby> we have had two incidents this year here with water destroying stuff in cardboard boxes standing directly on the ground
20141007 07:19:03 <xranby> pallets to the rescue
20141007 07:19:34 <eclesia> xranby: water reached the top of the basement, all cars there are drowned in my case. fortunatly mine was outside
20141007 07:19:45 <monsieur_max> :) yeah, basement located stuff is not directly on the ground, common mistake that you quickly learn not to reproduce !
20141007 07:20:12 <xranby> eclesia: in that case i dont think pallets would have helped /o\
20141007 07:20:13 <monsieur_max> eclesia: wow
20141007 07:30:04 <eclesia> by sympathy, would someone write a water like shader for me ? :D
20141007 07:42:23 <monsieur_max> eclesia: not skilled enough, but i can help you draw a very calm water plane, like a blue quad you know
20141007 07:45:52 * xranby (~xranby@anon) Quit (Ping timeout: 240 seconds)
20141007 07:46:25 <eclesia> monsieur_max: with difraction when looking thrught it ?
20141007 07:46:47 <eclesia> through*
20141007 07:47:36 <monsieur_max> eclesia: i wish i was skilled enough to code such a thing :)
20141007 07:53:30 <monsieur_max> eclesia: do you use opengl for work related projects ?
20141007 07:53:59 <eclesia> just for one, last year, only Opengl 1 ... very basic stuffs
20141007 07:58:12 <monsieur_max> oh ok :) I thought you were working in a 3D company, don't know why i imagined that
20141007 07:58:46 <eclesia> I make map applications
20141007 07:58:56 <eclesia> it's not that far
20141007 08:01:08 * xranby (~xranby@anon) has joined #jogamp
20141007 09:03:56 <monsieur_max> eclesia: does your company hire java developer ? ;)
20141007 09:04:30 <eclesia> monsieur_max: it does
20141007 09:16:42 * xranby (~xranby@anon) Quit (Ping timeout: 272 seconds)
20141007 09:30:41 * xranby (~xranby@anon) has joined #jogamp
20141007 09:30:56 * jvanek (jvanek@anon) has joined #jogamp
20141007 09:38:58 * hija (~hija@anon) has joined #jogamp
20141007 10:54:31 * monsieur_max (~maxime@anon) Quit (*.net *.split)
20141007 10:57:24 * monsieur_max (~maxime@anon) has joined #jogamp
20141007 11:08:09 * xranby (~xranby@anon) Quit (Ping timeout: 272 seconds)
20141007 11:10:11 * xranby (~xranby@anon) has joined #jogamp
20141007 11:10:14 <xranby> jogamp robovm status: https://github.com/robovm/robovm/issues/534 fail to compile class that loads using reflection -> ClassNotFound at runtime
20141007 11:11:05 <xranby> at least progress
20141007 11:11:42 <xranby> the jogamp jni librarys got picked up by robovm by specifying LD_LIBRARY_PATH=lib/linux-i586
20141007 11:12:27 <xranby> important to compile jogamp classes using the robovm -dynamic-jni option
20141007 11:13:11 <xranby> else robovm can only use static compilation and that would require jogamp to produce .a files
20141007 11:15:37 <xranby> testing a workaround by adding: jogamp.nativewindow.x11.X11GraphicsConfigurationFactory.registerFactory(); to demo main
20141007 11:16:52 * gouessej (5ee4b442@anon) has joined #jogamp
20141007 11:17:36 <gouessej> Hi
20141007 11:19:52 <gouessej> xranby: Do you need this bug to be fixed?
20141007 11:20:07 <gouessej> xranby: issue 534
20141007 11:20:45 <xranby> gouessej: yes, i will try some workaround, ideally i want to be able to compile jogamp projects using robovm and this is the current show stopper
20141007 11:21:11 <xranby> will first check if robovm upstream can fix it
20141007 11:21:16 <gouessej> xranby: Do you plan to give a try to avian too?
20141007 11:21:21 <xranby> yes
20141007 11:21:38 <gouessej> xranby: Avian + OpenJDK?
20141007 11:21:53 <monsieur_max> awesome :)
20141007 11:22:25 <xranby> gouessej: i am exploring all ahead of time static compilation options at the moment
20141007 11:22:41 <xranby> my target system is a raspberry pi connected to a usb lilliput screen
20141007 11:22:47 <gouessej> xranby: Which devices do you target?
20141007 11:22:51 <gouessej> ok
20141007 11:23:02 <xranby> the goal is to have a complete system bootup under 10s
20141007 11:23:03 <monsieur_max> ahead of time answering too ;)
20141007 11:24:08 <xranby> exploring ahead of time compilation on x86 systems before i try on the armv6 pi
20141007 11:26:33 <gouessej> xranby: Do you think we'll be able to run JOGL under Firefox OS one day? It seems to be very compromised
20141007 11:27:26 <xranby> if you build it yourself (gecko/iceweasel os etc) = yes .. if you use firefox os = no
20141007 11:27:46 <gouessej> xranby: iceweasel os??
20141007 11:28:14 <xranby> you know debian call firefox -> iceweasel
20141007 11:28:20 <xranby> due to trademark
20141007 11:28:36 <gouessej> yes
20141007 11:29:03 <xranby> my primary concern is to have jogamp running on platforms where native code is allowed
20141007 11:29:25 <xranby> i think the all locked down os will have a hard future
20141007 11:29:30 <monsieur_max> not jogamp but still 3D + oculus http://www.theverge.com/2014/10/6/6917343/watch-the-developer-become-god-in-this-precursor-to-the-metaverse
20141007 11:29:57 <gouessej> Even Firefox OS might have a hard future
20141007 11:30:16 <xranby> gouessej: yes firefox os will have a hahrd future
20141007 11:30:34 <xranby> custom roms for firefox os may be ok
20141007 11:31:19 <gouessej> Have you ever found one reliable custom rom for Firefox OS?
20141007 11:33:47 <xranby> gouessej: there is android roms available for the firefox os phone http://sayaksarkar.wordpress.com/2014/03/24/installing-android-on-geeksphone-keon/
20141007 11:34:20 <xranby> thus it is possible to replace the default locked down os
20141007 11:34:42 <gouessej> xranby: cool :)
20141007 11:35:33 <xranby> i am surprised that there is no widely used mech network phone os
20141007 11:35:55 <xranby> consider in hong kong everyone is using the firechat app to communicate
20141007 11:36:03 <xranby> that uses mech bluetooth network
20141007 11:36:24 <xranby> in order to communicate without the chinese government to lockdown the network
20141007 11:36:47 <xranby> firechat show that mech network do work
20141007 11:37:03 <xranby> but we need a phone os that can do it with encryption enabled by default
20141007 11:37:28 <xranby> also that would make us free from carriers
20141007 11:38:51 * xranby (~xranby@anon) Quit (Quit: Leaving.)
20141007 11:38:52 <gouessej> xranby: We'll need some infrastructure anyway
20141007 11:39:05 * xranby (~xranby@anon) has joined #jogamp
20141007 11:39:20 <gouessej> xranby: We'll need some infrastructure anyway, won't we?
20141007 11:39:53 <xranby> wifi routers for internet access at some locations would be enough
20141007 11:39:59 <gouessej> xranby: We need free bandwidth, neutral ISPs
20141007 11:40:21 <xranby> the mech network created by delegating information p2p between phones directly
20141007 11:40:24 <xranby> creates the network
20141007 11:40:48 <xranby> would probably lower energy as well.. no need to use long range radio
20141007 11:41:12 <gouessej> it wouldn't be enough in the countryside, maybe in towns
20141007 11:41:32 <gouessej> But at the end, if your ISP spies you, it ruins your efforts
20141007 11:41:50 <xranby> if you encrypt by default who cares who the information passes through
20141007 11:42:17 <xranby> there is no need to tust the network if the traffic is encrypted
20141007 11:42:19 <gouessej> it doesn't solve the problem of the DPI
20141007 11:42:26 <gouessej> deep packet inspection
20141007 11:42:38 <xranby> not sure i follow
20141007 11:43:11 <xranby> as long as you exchange keys face to face you can be sure there is no man in the middle
20141007 11:44:15 <gouessej> There is no purely technical solution. If we want a free society, we have to ensure that the ISPs respect the net neutrality. You can encrypt, it drives the spying harder.
20141007 11:44:50 <xranby> i would rather spend the same money i now give to carries to a local community mech network hackers
20141007 11:45:02 <gouessej> Yes, it's a short term solution
20141007 11:45:47 <xranby> most of this probably can be realized in the short term using a app on the phone
20141007 11:45:58 <gouessej> Yes, I don't deny it
20141007 11:46:02 <xranby> but if you want to be sure your phone is secure we need a phone os
20141007 11:46:16 <xranby> and some way to take controll of the baseband processor
20141007 11:46:49 <gouessej> Actually, if I want a safe phone, I need free software, free hardware and free bandwidth
20141007 11:54:47 <gouessej> xranby: they seem to agree with you: http://null-byte.wonderhowto.com/inspiration/is-firechat-future-internet-0157695/
20141007 11:55:02 <gouessej> "Imagine a day (not too far off) when everyone has a cell phone and has enabled an app like FireChat to provide a P2P mesh network. In that case, cell towers and other centralized access points will become unnecessary."
20141007 11:57:26 <xranby> yes that is something we need to try out
20141007 11:57:34 <xranby> in large scale
20141007 11:57:42 <xranby> so why not now
20141007 11:57:49 <xranby> we already got the hardware needed
20141007 12:02:43 <gouessej> Yes. Why not?
20141007 12:03:15 <gouessej> Please can anybody read my short wiki page and let me know whether it's understandable? http://jogamp.org/wiki/index.php/Foreign_Packaging
20141007 12:18:13 <xranby> gouessej: the text reads fine.. i do not have aproject to test on yet
20141007 12:18:41 <gouessej> xranby: T.U.E.R will use this stuff very soon
20141007 12:18:53 <gouessej> xranby: Good bye Java Web Start
20141007 12:19:08 <xranby> awesome robovm has ways to force compilation of certain class patterns
20141007 12:19:17 <xranby> https://github.com/robovm/robovm/issues/534
20141007 12:19:32 <xranby> -rvm:log=warn | grep WARN
20141007 12:19:43 <xranby> will list all classes robovm failed to load using reflection
20141007 12:20:01 <xranby> and then suggest what to do to force compilation
20141007 12:20:03 <xranby> example:
20141007 12:20:04 <xranby> [WARN] java.lang.Class: Class.forName() failed to load 'jogamp.opengl.x11.glx.X11GLXGraphicsConfigurationFactory'. Use the -forcelinkclasses command line option or add <forceLinkClasses><pattern>jogamp.opengl.x11.glx.X11GLXGraphicsConfigurationFactory</pattern></forceLinkClasses> to your robovm.xml file to link it in.
20141007 12:20:35 <gouessej> ok it would be a nice workaround
20141007 12:20:45 <gouessej> Yesterday, you spoke about a splashscreen
20141007 12:20:51 <gouessej> Will it be necessary?
20141007 12:22:54 <xranby> gouessej: sgothel suggested a splash screen
20141007 12:23:04 <xranby> its not necessary
20141007 12:23:17 <xranby> but it is a known workaround for slow starting applications
20141007 12:23:23 <xranby> to make the user happy
20141007 12:23:36 <xranby> my goal is to have the application start fast enough
20141007 12:23:48 <xranby> before the user gets bored
20141007 12:25:04 <gouessej> I'll probably go to the spalshscreen route if I plan to support Android with OpenJDK as I don't want to have tons of complicated things to do to make it work
20141007 12:25:17 <gouessej> I understand your position
20141007 12:25:28 <gouessej> 30 s, too much
20141007 12:25:35 <gouessej> 10 s, still a bit too much
20141007 12:26:48 <xranby> if i can get it to boot at 3s i woul dbe very happy
20141007 12:27:20 <xranby> https://www.youtube.com/watch?v=4Fjfqz6FxC8 <- raspberry pi 3s opengl es bootup
20141007 12:29:32 <gouessej> Ok good luck. I have to leave ;)
20141007 12:29:36 * gouessej (5ee4b442@anon) Quit (Quit: Page closed)
20141007 12:47:07 * hija (~hija@anon) Quit (Quit: hija)
20141007 12:47:42 * doev (~doev@anon) Quit (Ping timeout: 245 seconds)
20141007 13:17:34 <xranby> eclesia: https://www.shadertoy.com/view/Ms2SD1 <- fully procedural sea surface
20141007 13:17:39 <xranby> in a fragment shader
20141007 13:25:15 <eclesia> not sure if that's public
20141007 13:36:44 * hija (~hija@anon) has joined #jogamp
20141007 13:49:06 * doev (~doev@anon) has joined #jogamp
20141007 14:57:19 <sgothel> @Xerxes: 3s sounds nice
20141007 15:01:00 * eclesia (~husky@anon) Quit (Quit: Leaving.)
20141007 15:32:54 * doev (~doev@anon) Quit (Ping timeout: 258 seconds)
20141007 15:56:39 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20141007 16:33:10 * [Mike] (~Mike]@anon) has joined #jogamp
20141007 16:41:26 * monsieur_max (~maxime@anon) has joined #jogamp
20141007 17:08:27 * Eclesia (~eclesia@anon) has joined #jogamp
20141007 17:11:36 <Eclesia> I have a strange bug with my shadow map :
20141007 17:11:37 <Eclesia> http://jsorel.developpez.com/temp/shadowbug.png
20141007 17:12:11 * jvanek (jvanek@anon) Quit (Quit: Leaving)
20141007 17:12:27 <Eclesia> I think it's something related to the gl state but can't figure out what exactly
20141007 17:12:39 <Eclesia> any idea ?
20141007 17:58:27 * hija (~hija@anon) Quit (Quit: hija)
20141007 18:03:09 <rmk0> .o.
20141007 18:04:29 * xranby (~xranby@anon) Quit (Ping timeout: 260 seconds)
20141007 18:05:02 * xranby (~xranby@anon) has joined #jogamp
20141007 18:08:23 <rmk0> Eclesia: assume that white rectangle at the top left is the light's depth buffer?
20141007 18:08:27 <rmk0> top right, even
20141007 18:08:59 <Eclesia> rmk0: yes but it's a vec1 float buffer
20141007 18:09:12 <Eclesia> so nothing visible
20141007 18:09:18 <rmk0> hm, right
20141007 18:09:26 <rmk0> apitrace isn't particularly good at displaying those, is it...
20141007 18:09:46 <Eclesia> my problem is the square at cube intersection in the shadowmap
20141007 18:10:06 <rmk0> yeah, what i was leading towards was... how does that intersection appear in the map?
20141007 18:10:28 <rmk0> but you can't really see the map to inspect it
20141007 18:10:48 <Eclesia> I checked the content of the shadowmap, it's correct everything, depth value are in world unit so distant to light is ok, but not at the place where cubes intersect
20141007 18:11:23 <rmk0> mystified as to what could cause that
20141007 18:11:39 * rmk0 thinks
20141007 18:12:10 <Eclesia> I guess it's something like blend state or stencil maybe
20141007 18:12:35 <rmk0> could you send me an apitrace of it?
20141007 18:12:40 <rmk0> if the resulting trace isn't too huge
20141007 18:12:56 <rmk0> just run it for a few frames so that i can see the opengl calls and state
20141007 18:13:18 <Eclesia> hm.. ok just give me a few minutes
20141007 18:18:52 <Eclesia> rmk0 a quick reminder : apitrace -? java -jar ...
20141007 18:19:45 <rmk0> apitrace trace java -jar ...
20141007 18:21:11 <sgothel> http://io7m.com/contact/index.xhtml <- cute :)
20141007 18:21:24 <Eclesia> uploading
20141007 18:21:37 <rmk0> sgothel: hehe, made it out of actual newspapers
20141007 18:21:44 <rmk0> i didn't set a ransom for myself though
20141007 18:22:16 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=1088 <- experimented a lot w/ this issue
20141007 18:22:33 * rmk0 eyes it
20141007 18:22:34 <Eclesia> rmk0: http://jsorel.developpez.com/temp.java.trace
20141007 18:22:40 <rmk0> Eclesia: thanks
20141007 18:22:45 <Eclesia> oups http://jsorel.developpez.com/temp/java.trace
20141007 18:23:16 <Eclesia> rmk0: thanks to you, for taking the time to help me :)
20141007 18:23:27 <sgothel> it is related also to (https://www.opengl.org/registry/specs/ARB/robustness.txt Overview Paragraph 5)
20141007 18:24:32 <sgothel> so on OSX 10.9.5 .. using multiple GLContext, which itself shares a common master, multithreaded might be an issue -> crash
20141007 18:25:13 <sgothel> I attempted to lock the ShareSet we provide via GLContextImpl, which works - but is too much / invasive
20141007 18:25:24 <sgothel> (taking away all the multithreading fun)
20141007 18:26:12 <rmk0> sounds deeply horrible
20141007 18:26:12 <sgothel> OSX not always crashes .. only if user app actually access the shared object concurrently ..
20141007 18:26:40 <sgothel> for example, our GLMediaPlayerImpl does not crash - due to proper use of RingBuffer*
20141007 18:27:10 <sgothel> yup .. horrible regression in OSX 10.9.5
20141007 18:27:54 <sgothel> so what I do now is to add a Quirk for known renderer .. but won't handle it in our API (-> too complex/invasive)
20141007 18:28:25 <sgothel> as a test, I decorated user synchronization in GearsES2's draw methods .. no crash
20141007 18:29:09 <rmk0> OS X is one big regression
20141007 18:29:14 <rmk0> not sure which version the mac here has
20141007 18:29:22 <rmk0> sits in a corner unloved, kicked occasionally
20141007 18:29:35 <sgothel> we (or the user) .. could add higher level GL-Objects adding synchronization for shared usage though ..
20141007 18:29:53 <rmk0> oh ... 10.9.5!
20141007 18:29:57 <sgothel> yeah :)
20141007 18:30:21 <sgothel> I wondered .. how come unit test failure went up from 10 -> ~24 or so .. :)
20141007 18:30:26 <sgothel> _after_ the upgrade
20141007 18:31:19 <rmk0> can't get this trace to replay... think it's a bug in apitrace
20141007 18:32:09 <Eclesia> rmk0: :(
20141007 18:32:26 <rmk0> i'm curious as to why there's an increasing number of opengl calls each frame
20141007 18:32:41 <Eclesia> incrising ?
20141007 18:32:45 <rmk0> steadily goes up from 844 to 7788 over ~100 frames
20141007 18:33:28 <Eclesia> I might have a leak somewhere ... checking
20141007 18:34:32 <rmk0> seems to be mostly glVertex3f calls, drawing lines
20141007 18:34:47 <Eclesia> ha I know
20141007 18:35:00 <Eclesia> my bad it's the camera debug
20141007 18:35:07 <Eclesia> I didn't remove it
20141007 18:35:21 <Eclesia> the white lines on the screenshot
20141007 18:36:12 <Eclesia> I'll make a new dump without those, maybe it will be better
20141007 18:37:26 <rmk0> annoying that it won't replay the trace... am basically stuck looking at raw opengl calls
20141007 18:37:30 <rmk0> will try on another machine
20141007 18:38:58 <Eclesia> rmk0: new trace uploaded, same adresse, it's much smaller and always the same number of calls per frame
20141007 18:39:21 <rmk0> got it
20141007 18:39:55 <Eclesia> works ?
20141007 18:40:13 <rmk0> it still doesn't replay, but it's easier to read
20141007 18:40:34 <rmk0> the call that causes it is inside jogl, is nothing you can do
20141007 18:40:36 <Eclesia> I use opengl 4, your pc handle it ?
20141007 18:40:55 <rmk0> i don't have anything opengl 4 here
20141007 18:41:13 <Eclesia> glsl is marked 400
20141007 18:41:41 <rmk0> it's the actual context creation it doesn't like
20141007 18:41:55 <rmk0> will tell the apitrace people... it should still be able to run to some extent
20141007 18:42:18 <rmk0> i'm seeing calls like glViewport(0,0,512,512)
20141007 18:42:24 <rmk0> i'm guessing that means "i'm rendering a shadow map now"
20141007 18:42:53 <Eclesia> yes
20141007 18:43:30 * xranby (~xranby@anon) Quit (Ping timeout: 244 seconds)
20141007 18:44:22 <sgothel> 'GLRendererQuirk.SharedObjectSyncRequired' comprehensible for Bug 1088 ?
20141007 18:45:41 <rmk0> sgothel: think so
20141007 18:46:00 <rmk0> Eclesia: out of curiosity... what happens if you cull front faces while rendering the shadow map, instead of back faces?
20141007 18:46:36 <Eclesia> testing
20141007 18:46:50 <rmk0> it's not clear why blending, dithering, multisampling, etc, is enabled when producing the map, either
20141007 18:48:21 <Eclesia> result is slightly different, but still a hole in the middle
20141007 18:49:44 <Eclesia> rmk0: I enable multisample whithout thinking, is that bad if the fbo does not contain any ?
20141007 18:50:38 <rmk0> not sure... i'm working in deferred rendering land, so i don't get any hardware multisampling
20141007 18:50:53 <rmk0> i think it's weird to see it enabled when producing a depth-only image, but it may be harmless
20141007 18:51:49 <rmk0> am pretty sure blending should be disabled
20141007 18:52:13 <rmk0> not exactly sure what opengl does in the case of a color attachment not having an alpha channel
20141007 18:52:45 <Eclesia> testing whitout all those
20141007 18:53:20 * xranby (~xranby@anon) has joined #jogamp
20141007 18:57:04 <Eclesia> rmk0: found the bug
20141007 18:57:14 <Eclesia> my color material which enable : gl.glEnablei(GL4.GL_BLEND,0);
20141007 18:57:15 <Eclesia> gl.glBlendFunci(0, GL4.GL_ONE, GL4.GL_ONE_MINUS_SRC_ALPHA);
20141007 18:57:31 <rmk0> right
20141007 18:57:44 <rmk0> am curious as to what it was doing to the final values
20141007 18:58:22 <Eclesia> can't tell, the value at this coordinate wasn't event between the 2 box distance
20141007 18:58:39 <Eclesia> I think he interpreted the float as a RGBA and blended them
20141007 18:58:51 * rmk0 eyes opengl spec
20141007 18:58:59 <Eclesia> 'blended' is that correct ?
20141007 18:59:07 <rmk0> sounds it
20141007 19:01:02 <Eclesia> rmk0: thanks for the help, it might not be of any use for you but I started POP/SMTP protocol in the lib if you need them ^^ or do prefer a >o< ? :D
20141007 19:01:11 <Eclesia> do you*
20141007 19:01:34 <rmk0> hehe, no charge
20141007 19:01:49 <rmk0> try to avoid SMTP for everyone's sake
20141007 19:01:53 <rmk0> and sugar!
20141007 19:02:25 <rmk0> still trying to work out what the opengl spec says should happen when confronted with a single-channel floating point format
20141007 19:02:45 <sgothel> *blending states* .. tell me about it .. :)
20141007 19:02:56 <Eclesia> héhé, I'm doing smtp because it's simple. I'm not an expert in protocol, so implementing them slowly, echo, smtp, pop, imap, ftp ,...
20141007 19:03:37 <Eclesia> okay, time to eat, see you later
20141007 19:03:40 <sgothel> re SMTP: I guess w/o SSL/TLS its a wide open spam bot free to use, but there are so many already :)
20141007 19:04:03 <sgothel> *SSL/TLS auth* that is .. ofc
20141007 19:04:33 <sgothel> blending: writing to a target .. blends .. yup
20141007 19:07:21 <rmk0> ah...
20141007 19:07:23 <rmk0> looks like
20141007 19:07:57 <rmk0> a single RED channel will be converted to (R, 0, 0, 1)
20141007 19:08:08 <rmk0> and it won't be clamped to [0.0, 1.0] if it's floating point
20141007 19:08:32 <sgothel> i.e. for 2D blending .. I had (hindsight ofc) disable depth-testing .. Graph-UI .. took me a while :)
20141007 19:09:01 <rmk0> hehe, you drawing everything onto the same plane?
20141007 19:09:28 <sgothel> no, but forcing overdraw .. to see the transparent stuff behind
20141007 19:09:41 <sgothel> i.e. depth culling off
20141007 19:10:18 <rmk0> right
20141007 19:10:31 * void256 (~void@anon) has joined #jogamp
20141007 19:38:01 * hija (~hija@anon) has joined #jogamp
20141007 19:59:23 * Eclesia is back
20141007 19:59:39 <Eclesia> Is there a magic function : gl.reset ?
20141007 19:59:47 <rmk0> nope
20141007 20:00:37 <Eclesia> damn you kronos group :w
20141007 20:03:51 <sgothel> you could use shared context, since no states are shared there, but the buffer objects only
20141007 20:05:19 <sgothel> however, it is not uncommon to reflect all GL states in higher level object to allow reset when 'leaving' desired state .. yeah, PIA
20141007 20:05:37 <Eclesia> pia ?
20141007 20:06:34 <sgothel> Pain In the Ass :)
20141007 20:06:50 * Eclesia will remember this one ^^
20141007 20:07:30 <Eclesia> okay, with shadows starting to work, the game pipeline is nearly finished :)
20141007 20:07:39 <Eclesia> makes a nice class: https://bitbucket.org/Eclesia/unlicense/src/tip/engine/engine-opengl/src/main/java/un/engine/opengl/phase/GamePhases.java?at=default
20141007 20:07:46 <rmk0> which algorithm are you using for shadow mapping?
20141007 20:08:10 <Eclesia> lightperspective shadow maps, basic
20141007 20:08:15 <Eclesia> level 1 :D
20141007 20:08:32 <Eclesia> only spotlight, I need to do pointlight too
20141007 20:09:36 <rmk0> http://http.developer.nvidia.com/GPUGems3/gpugems3_ch08.html
20141007 20:09:40 <rmk0> would recommend those
20141007 20:09:54 <rmk0> they seem the least fucked up and most predictable of all the "soft" shadow algorithms
20141007 20:10:15 <rmk0> i didn't bother to implement the summed area part, just the plain variance shadows with all the improvements they listed
20141007 20:11:07 <Eclesia> it's still a bit to complicate for me yet :)
20141007 20:11:25 <sgothel> nice nice .. so I know where to find my shadows (eclesia, mark) .. now :)
20141007 20:11:57 <Eclesia> go ahead ;)
20141007 20:14:23 <Eclesia> rmk0: of course if you want to give me the shaders I won't complain ;)
20141007 20:16:10 <rmk0> they're not written in GLSL... /o\
20141007 20:16:25 <Eclesia> directX ?
20141007 20:16:36 <rmk0> parasol
20141007 20:16:39 <sgothel> some weirdo texture operations ?
20141007 20:16:44 <sgothel> ah ..
20141007 20:16:49 <Eclesia> ???
20141007 20:17:01 <sgothel> probably can parasol -> GLSL
20141007 20:17:21 <rmk0> it does eject GLSL at compile time, but the result isn't exactly pretty
20141007 20:17:34 <rmk0> mangled names to respect GLSL's insane naming "rules"
20141007 20:17:35 <rmk0> etc
20141007 20:18:05 <rmk0> http://waste.io7m.com/2014/10/05/Transform.p
20141007 20:18:08 <rmk0> http://waste.io7m.com/2014/10/05/ShadowVariance.p
20141007 20:19:56 <Eclesia> funny syntax, looks like a mix of VisualBasic and C
20141007 20:20:16 <rmk0> hehe
20141007 20:20:59 <Eclesia> let me guess, the tool to convert parasol syntax in glsl is called ... hm ... Towel ?
20141007 20:21:21 <rmk0> hehe, no, parasol's the language and it (only) has a GLSL backend
20141007 20:21:27 <rmk0> lost patience with GLSL very, very quickly
20141007 20:21:34 <rmk0> needed something with modules and stronger typing
20141007 20:22:12 * rmk0 attempts to extract glsl
20141007 20:22:23 <sgothel> yup - was just saying, looks so pretty :)
20141007 20:24:24 <rmk0> http://waste.io7m.com/2014/10/07/variance.txt
20141007 20:24:32 <rmk0> that's the GLSL the compiler produces... like i said, not too pretty
20141007 20:24:44 <rmk0> is more or less exactly what the article says, though
20141007 20:25:31 <rmk0> you pass p_com_io7m_r1_core_ShadowVariance_factor() a little config struct, a floating point RG texture consisting of depth values in the red channel and depth squared in the green, and an eye space position p
20141007 20:26:29 <rmk0> it'll return a value in the range [0.0, 1.0]
20141007 20:27:06 <rmk0> 0.0 if the object is fully shadowed, 1.0 if it isn't... you'd just multiply that by the current light term
20141007 20:27:30 <rmk0> the values in the config struct are tunable on a per-scene basis... they give the meanings in the article
20141007 20:27:59 <rmk0> the basic idea is that ordinary shadow mapping is a binary thing, either a fragment is in shadow or it isn't
20141007 20:28:13 <Eclesia> the code seems very short compared to the article
20141007 20:28:34 <rmk0> the article's quite a detailed description
20141007 20:28:48 <rmk0> explains how the original variance shadow map algorithm fails, and suggests improvements
20141007 20:29:03 <rmk0> and then in the last section, they add these summed area tables for some sort of penumbra thing
20141007 20:29:07 <rmk0> i didn't bother with that, seems unnecessary
20141007 20:30:13 <rmk0> .. i tell a lie, you pass it a clip-space position, not eye-space
20141007 20:30:22 <rmk0> hehe, been a while since i looked at this code
20141007 20:31:16 <Eclesia> you need only the shadow map for this technique ?
20141007 20:31:49 <rmk0> you need a single "depth variance map"... in your case where you're rendering depth to an R texture, you'd instead render (R, R * R)
20141007 20:32:00 <rmk0> to an RG texture, obviously
20141007 20:32:28 <Eclesia> I could do the R*R in the fragment shader directly no ?
20141007 20:32:46 <rmk0> nope, because the technique relies on being able to filter maps before use
20141007 20:33:31 <rmk0> the article explains it better, but the general idea is to get the shadow map data into a form that's linear, so that it can be interpolated
20141007 20:33:41 <rmk0> you can apply a box blur to it, etc, and get softer and softer shadows
20141007 20:34:17 <rmk0> as i started saying... traditional shadow mapping is a very "binary" thing
20141007 20:34:31 <rmk0> a fragment is in shadow or it isn't, based on a comparison between the depth value in the map and the fragment's depth value
20141007 20:34:54 <rmk0> variance shadow mapping is more "how probable is it that this fragment is in shadow?"
20141007 20:35:38 <Eclesia> I'll read the article at lunch break tomorrow
20141007 20:35:47 <rmk0> it's pretty painless
20141007 20:36:52 <Eclesia> we'll see. I'm not as good as you are in 3d ^^
20141007 20:37:22 <rmk0> i'm not great... i've just been doing it non-stop for three years
20141007 20:37:35 <rmk0> you tend to get used to things!
20141007 20:41:23 * hija (~hija@anon) Quit (Quit: hija)
20141007 20:41:59 <rmk0> managed to miss harvey when he appeared...
20141007 20:42:05 <rmk0> got to nag him about that blender stuff
20141007 20:47:25 * hija (~hija@anon) has joined #jogamp
20141007 20:48:57 * monsieur_max (~maxime@anon) Quit (Quit: Leaving.)
20141007 20:49:32 * hija (~hija@anon) Quit (Client Quit)
20141007 20:49:39 * bbbruce (~bx@anon) Quit (Read error: Connection reset by peer)
20141007 20:49:45 * bbbruce_ (~bx@anon) has joined #jogamp
20141007 20:52:07 * hija (~hija@anon) has joined #jogamp
20141007 20:54:00 <rmk0> http://waste.io7m.com/2014/10/06/dxhr/dxhr.html
20141007 20:54:07 <sgothel> @Mark: Nice - and it still looks pretty!
20141007 20:54:15 <rmk0> yep
20141007 20:54:47 <rmk0> ^^ the above might be mildly interesting... spent yesterday running deus ex human revolution in API trace to see how they do their rendering
20141007 20:55:03 <sgothel> hmm
20141007 20:56:20 <sgothel> excellent teaser .. very much earmark apitrace :)
20141007 20:56:32 <sgothel> you say .. their output can be imported .. hmm
20141007 20:56:45 <rmk0> yeah, although it didn't work too well earlier, as you may've seen
20141007 20:56:49 <void256> I liked deus ex human revolution a lot =D
20141007 20:56:55 <sgothel> wonder whether we would be able to produce such output ourselves ..
20141007 20:57:04 <void256> and the api trace is great O_O
20141007 20:57:13 <rmk0> it's fun to step through a commercial game and see how they've done things
20141007 20:57:24 <rmk0> there's very little technical information on the game available, and no modding scene whatsoever
20141007 20:57:33 <sgothel> yup .. games, apps ..
20141007 20:57:37 * Eclesia (~eclesia@anon) Quit (Ping timeout: 260 seconds)
20141007 20:57:53 <rmk0> that was direct3d 9
20141007 20:58:06 <rmk0> supposedly it'll use direct3d 11 on newer hardware... not got any to try
20141007 20:58:11 <sgothel> I have to admit, I haven't any modern game since .. err .. 2007 or so .. and then only to see the *magic* :)
20141007 20:58:24 <rmk0> yeah... they're pretty bland in general
20141007 20:59:36 <rmk0> am going to try tracing half life 2 next
20141007 20:59:49 <rmk0> they use a pretty traditional forward renderer, but i'd like to see how they implemented refraction
20141007 20:59:54 <sgothel> that is great .. really, like seeing how to achieve certain effects
20141007 21:00:09 * Eclesia (~eclesia@anon) has joined #jogamp
20141007 21:00:20 <rmk0> apitrace is better for opengl than directx, currently
20141007 21:00:35 <rmk0> it's not currently able to tell you what format textures are, for example
20141007 21:00:56 <void256> deus ex looked pretty to my eyes
20141007 21:01:06 <rmk0> one of the things i wanted to see was what texture format deus ex uses to store normals, but all i get out of apitrace is RENDER_TARGET_0,UNKNOWN
20141007 21:01:11 <rmk0> yeah, deus ex looked great
20141007 21:01:18 <rmk0> their artists were extremely clever
20141007 21:01:38 <void256> I have that digital artbook
20141007 21:01:38 <sgothel> hmm
20141007 21:01:48 <rmk0> so many little touches... cleverly animated normal maps to simulate tarpaulin flapping in the wind
20141007 21:01:56 <void256> which was great but a bit thin :|
20141007 21:02:01 <void256> yes
20141007 21:02:05 <void256> it was greatly designed
20141007 21:02:11 <void256> as was half life 2 actually!
20141007 21:02:15 <rmk0> yep
20141007 21:02:16 <void256> (for its time)
20141007 21:02:23 <void256> I have that artbook too :)
20141007 21:02:34 <void256> they spent a lot of time in designing stuff
20141007 21:02:46 <sgothel> so those books explain how to impl. such effects ?
20141007 21:03:11 <sgothel> and detail the psycho impact .. like space perception .. and the like ?
20141007 21:03:56 <rmk0> think they're higher level than that, if they're what i'm thinking of
20141007 21:04:06 <rmk0> explaining the psychological effect of lighting and so on, on the player
20141007 21:04:19 <sgothel> but not going low .. (GL) ?
20141007 21:04:32 <rmk0> don't think the audience is programmers
20141007 21:04:39 <sgothel> however .. that is also great to know
20141007 21:04:53 <sgothel> in other projects .. the team had an artist .. so he knew .. :)
20141007 21:05:01 <rmk0> they use this great cyan lighting in half life 2 for the antagonists
20141007 21:05:35 <rmk0> they tend to use existing human buildings but install brutal looking metal contraptions
20141007 21:05:44 <void256> not about effects ... more like real design
20141007 21:05:52 <rmk0> almost every area they've adapted has this nice cold lighting
20141007 21:06:12 <sgothel> 'nice cold lighting' :)
20141007 21:06:30 <sgothel> winter is coming
20141007 21:06:34 <rmk0> http://gamerlimit.com/wp-content/uploads/2009/03/half-life-2-combine.jpg
20141007 21:06:40 <void256> what does a east european country looks like after a long time being invaded by aliens
20141007 21:06:44 <void256> that sort of stuff :)
20141007 21:07:17 <void256> its the same with gears of war
20141007 21:07:17 <sgothel> well .. sadly, one only need to watch the news
20141007 21:07:24 <void256> make something pretty and then destroy it
20141007 21:07:25 <void256> =)
20141007 21:08:10 <sgothel> soon there will be blooming flowers, blue skies and sunshine rendered - where there is none
20141007 21:08:21 <void256> they usually invest a lot of time into character and environment design and it really shows ... its not like some 3d models slaped together =)
20141007 21:08:42 <void256> everything has a purpose
20141007 21:08:44 <sgothel> yup, looks like - sure
20141007 21:08:45 <void256> and such things
20141007 21:08:47 <void256> :>
20141007 21:08:51 <rmk0> http://www.electricblueskies.com/wp/wp-content/uploads/2012/07/55-Half-Life-2-1080p-Wallpaper-Garrys-Mod-CITY-17-COMBINE-ARCHITECTURE-WALLS.jpg
20141007 21:08:55 <rmk0> ♥
20141007 21:08:58 <void256> EXACTLY
20141007 21:09:07 <void256> and the lighting is ace too!
20141007 21:09:21 <rmk0> is all static, ray-traced
20141007 21:09:31 <rmk0> it looks nice, but it's awful to work with
20141007 21:09:36 <void256> you have like that alien technology penetrating the real world :>
20141007 21:09:43 <rmk0> "i've just edited this map, now i need to recompile the lighting for the next eight hours"
20141007 21:09:49 <void256> :>
20141007 21:09:59 <sgothel> yup .. back to the books .. i.e. how to realize design w/ the limitation of crooked 3D (which is GL)
20141007 21:10:57 <sgothel> so apitrace will help here, great
20141007 21:13:24 <sgothel> Q: each of the frames is a snapshot of the target color renderbuffer ?
20141007 21:13:41 <sgothel> default framebuffer .. or on of many ?
20141007 21:14:11 <sgothel> i.e. how does apitrace trigger a non-default framebuffer op has finished ?
20141007 21:14:13 <rmk0> you can sort of assume that each group of pictures is a separate framebuffer
20141007 21:14:35 <rmk0> in opengl, you see the calls to glBindFramebuffer(..., 0);
20141007 21:14:45 <rmk0> in direct3d, it's something else... SetRenderTarget
20141007 21:15:07 <Eclesia> Ho right, made a demo on the game phases, all effects active :) http://jsorel.developpez.com/temp/test.ogv
20141007 21:15:07 <rmk0> in the gui, the right pane shows the current render target... can inspect the color and depth attachments and so on
20141007 21:15:10 <sgothel> yeah .. but that is not required, especially when you use many FBOs .. i.e. 1, 2, 3, .. may only reach default 0 at the very end
20141007 21:15:38 <rmk0> sgothel: yep... usually you'll see a glBlitFramebuffer call at the end, or the directx equivalent
20141007 21:15:47 <Eclesia> it's not deus ex or halflife, but I'm proud of my work ^^
20141007 21:15:49 <sgothel> is MSAA .. only
20141007 21:15:52 <sgothel> *if*
20141007 21:16:11 <rmk0> i mean it's literally a list of opengl calls without any of the program logic... apitrace just tracks the state changes on the client side and keeps a copy of all shaders and textures
20141007 21:16:25 <sgothel> sure ..
20141007 21:16:27 * rmk0 downloads test.ogv
20141007 21:17:03 <sgothel> eclesia: nice 'dance' .. sharp / blurry scene .. nice
20141007 21:17:31 <rmk0> looks good!
20141007 21:17:40 <sgothel> interpolating animation, time-frame .. (forgot the name)
20141007 21:17:51 <sgothel> could be more smooth ?
20141007 21:17:53 <Eclesia> kinematics ?
20141007 21:18:16 <sgothel> so you apply constraints ..
20141007 21:18:35 <sgothel> inverse kinematics ?
20141007 21:18:42 <Eclesia> for legs yes
20141007 21:18:48 <void256> bones?
20141007 21:18:54 <Eclesia> joints
20141007 21:19:15 <sgothel> the sony engine .. what was it .. jbullet or the like ?
20141007 21:19:36 <sgothel> @Mark: can one add apitrace instrumentation ?
20141007 21:19:39 <Eclesia> nope, my physics engine isn't working correctly yet
20141007 21:19:54 <sgothel> i.e. auto snapshot .. etc .. special logging in stream
20141007 21:20:07 <rmk0> sgothel: it's got no controllable API at all, as far as i know
20141007 21:20:25 <sgothel> an APITRACE_EXT would be nice :)
20141007 21:20:26 <rmk0> you can just about tell it where to dump a trace file with an environment variable
20141007 21:20:42 <rmk0> other than that, it's just a replacement libGL.so that writes all calls and data to a file
20141007 21:20:46 <sgothel> hmm .. maybe we can add something ..
20141007 21:21:20 <sgothel> like an extension .. some functions accessible the usual way
20141007 21:21:39 <rmk0> what is it you'd want to do?
20141007 21:22:20 <sgothel> a determined trigger for saving framebuffers .. associated w/ point in trace log etc - proper filename (buffer name, frame-count, etc)
20141007 21:22:34 <sgothel> adding commentary in log file
20141007 21:22:47 <sgothel> (yup debug API could be used for that maybe)
20141007 21:23:14 <rmk0> i should mention... the contents of framebuffers don't appear in the saved traces
20141007 21:23:19 <rmk0> not sure if that was clear or not
20141007 21:23:31 <sgothel> you created them .. online ?
20141007 21:23:35 <rmk0> it's when the "glretrace" program replays the trace that it can get access to the framebuffers
20141007 21:23:47 <sgothel> ok
20141007 21:23:49 <rmk0> because it's literally just making the same opengl calls that the original program did
20141007 21:24:24 <rmk0> but all the source material like textures, shaders, etc... all of that is present in the trace (it'd have to be, otherwise how would glretrace replay it...)
20141007 21:25:19 <sgothel> so you create the snapshot w/ glretrace .. or done auto-magically ?
20141007 21:25:32 <sgothel> (i.e. the trigger question)
20141007 21:25:56 <rmk0> glretrace is the program that does the hard work, there's a gui on top of that called qapitrace where you can actually inspect state like the current framebuffer and so on
20141007 21:25:59 <sgothel> hence the extension, could be triggered by the program itself :)
20141007 21:26:16 <sgothel> yup .. understood that
20141007 21:26:26 <rmk0> it's... probably easiest if you try running it
20141007 21:26:29 <sgothel> :)
20141007 21:26:38 <rmk0> you get presented with a list of frames, inside those frames are all the opengl calls
20141007 21:26:56 <rmk0> double click one of the calls, and an opengl window will pop up and the program will replay all calls up to the one you clicked on
20141007 21:27:07 <sgothel> noticed the new lack of 'no alternative' in UIs ? like 'we inform you that you have to obey - or we do things -> OK' :-/
20141007 21:27:12 <rmk0> a pane on the right opens, and shows the currently bound textures, shaders, etc
20141007 21:27:34 <sgothel> nice
20141007 21:27:39 <rmk0> not really... don't tend to really see new UIs
20141007 21:27:48 <sgothel> websites and the like :)
20141007 21:27:57 <rmk0> not really... don't tend to really see new web sites /o\
20141007 21:28:01 <sgothel> lol
20141007 21:28:19 <rmk0> i'm just happy when a graphics paper written more than a year ago is still accessible!
20141007 21:28:44 <sgothel> right, wget is already assumed as a hackers tool :)
20141007 21:28:55 * rmk0 sniggers
20141007 21:29:31 <sgothel> well, thank you - you made my day. now me finalizing the OSX bugs.
20141007 21:29:40 <rmk0> \o/
20141007 21:29:54 <rmk0> i might do some articles on this stuff
20141007 21:30:00 <rmk0> has been more interesting to people than i'd expected
20141007 21:30:02 <sgothel> awesome!
20141007 21:30:02 <rmk0> not just here
20141007 21:30:19 <rmk0> is obviously not something people regularly see
20141007 21:30:38 <sgothel> maybe wiki entry in JogAmp .. which then also links to your site w/ your analysis ..
20141007 21:30:47 <rmk0> ok
20141007 21:30:51 <sgothel> great!
20141007 21:31:14 <rmk0> http://fabiansanglard.net
20141007 21:31:20 <rmk0> he does a lot of it
20141007 21:31:35 <rmk0> think he only covers programs that have sources available, though
20141007 21:31:35 <sgothel> Server not found
20141007 21:31:41 <rmk0> woops
20141007 21:31:45 <rmk0> http://fabiensanglard.net
20141007 21:31:56 <rmk0> i'm typing out URIs because copy/paste has broken in virtualbox
20141007 21:32:01 * rmk0 kicks it
20141007 21:32:30 <rmk0> the quake and doom 3 reviews are particularly good
20141007 21:35:14 <Eclesia> good night all and thanks again :)
20141007 21:35:29 <rmk0> byee!
20141007 21:35:30 * Eclesia (~eclesia@anon) Quit (Quit: Leaving.)
20141007 21:35:31 <sgothel> good night, a pleasure, as always
20141007 21:35:38 <sgothel> opps .. too late :)
20141007 21:43:26 <void256> fabien rocks
20141007 21:43:28 <void256> :>
20141007 21:43:52 <void256> I liked his strike commander reverse engineering
20141007 21:44:08 <sgothel> http://www.iquilezles.org/www/index.htm <- another good resource, just for the record
20141007 21:45:15 <void256> *bookmarked* :>
20141007 22:05:59 * void256 (~void@anon) Quit (Remote host closed the connection)
20141007 23:00:33 * hija (~hija@anon) Quit (Quit: hija)
20141007 23:01:23 * hija (~hija@anon) has joined #jogamp
20141007 23:21:05 * hharrison (~chatzilla@anon) has joined #jogamp
20141007 23:21:45 <hharrison> Finally got a release to show off some of Jaamsim's rendering capabilities.
20141007 23:21:46 <hharrison> https://www.youtube.com/watch?v=NpoJ80xGWSQ&list=PLhKFtFk5I4upypes73KeR1e4PbAlsI9Ny
20141007 23:22:13 <hharrison> Note this video was recorded using the built in vp8 encoder we ship in jaamsim
20141007 23:27:26 * hija (~hija@anon) Quit (Quit: hija)
20141007 23:29:57 * hija (~hija@anon) has joined #jogamp
20141007 23:40:44 <sgothel> ohhh ... :)
20141007 23:41:15 <sgothel> camera sweep is a bit stuttering ..
20141007 23:41:20 <hharrison> yes it is
20141007 23:41:28 <sgothel> nice zoom out view ..
20141007 23:41:44 <sgothel> nice machines
20141007 23:41:58 <hharrison> Finally got a client to say it was OK to use their assets
20141007 23:42:07 <sgothel> I can imagine :)
20141007 23:42:16 <hharrison> It's been a long wait
20141007 23:42:43 <sgothel> I would rework fluent animation .. to proof the speakers .. fluent animation .. ahem .. *duck*
20141007 23:42:54 <sgothel> probably the encoder ?
20141007 23:43:21 <sgothel> link video to jaamsim link on JOGL ?
20141007 23:43:43 <sgothel> Harry in a suite, sweet :)
20141007 23:44:33 <sgothel> (no Harvey in there though .. the followup video)
20141007 23:44:37 <hharrison> Yeah, we're in the process of wiring animations in more fully, that will be a new video soonish
20141007 23:44:53 <hharrison> Get the buckets on the machines turning....etc
20141007 23:44:58 <sgothel> but I think .. the renderer does it fluent .. no ?
20141007 23:45:12 <hharrison> The stutter is from the camera 'script' we used
20141007 23:45:16 <sgothel> ah ..
20141007 23:45:26 <hharrison> Also something that is a bit WIP
20141007 23:45:33 <sgothel> well, I use ffmpeg/libav here for screen-cast
20141007 23:45:47 <sgothel> all fluent .. w/ v-sync .. etc
20141007 23:46:01 <hharrison> Yes, but we get frame-perfect capture this way (it's built into the simulation engine)
20141007 23:46:20 <hharrison> So the model is in a known state
20141007 23:46:50 <sgothel> science/tech vs .. general user perception :)
20141007 23:46:53 <hharrison> And we can do it all offscreen at whatever resolution we desire
20141007 23:47:01 <sgothel> nice
20141007 23:47:13 <hharrison> I think so ;-)
20141007 23:47:28 <sgothel> just picking on details - hey, looks great!
20141007 23:47:44 <hharrison> Of course! Don't be afraid to pick
20141007 23:48:22 <sgothel> make a patch to jogl/www/index.html for video(s)
20141007 23:48:32 <hharrison> sure
20141007 23:50:47 <sgothel> So thats the alternative to port (?) to ship stuff from Northern BC to customers ? (Pipeline being still a dream)
20141007 23:50:55 <sgothel> *alternative port*
20141007 23:53:41 <sgothel> http://www.co.whatcom.wa.us/pds/plan/current/gpt-ssa/ <- nope
20141007 23:54:06 <hharrison> that;s the one
20141007 23:54:29 <hharrison> Just happens to be in the neighbourhood, not a pipeline alternative
20141008 00:26:19 * hharrison (~chatzilla@anon) Quit (Quit: ChatZilla 0.9.90.1 [Firefox 33.0/20140929180120])
20141008 01:52:25 * [Mike] (~Mike]@anon) Quit ()
20141008 02:53:25 * hija (~hija@anon) Quit (Quit: hija)
20141008 04:51:37 * rmk0-alt (~rmk0@anon) Quit (Ping timeout: 244 seconds)
20141008 04:52:35 * rmk0-alt (~rmk0@anon) has joined #jogamp
20141008 05:05:54 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20141008050554.html