#jogamp @ irc.freenode.net - 20130808 10:09:06 (UTC)
20130808 10:09:06 -jogamp- Previous @ http://jogamp.org/log/irc/jogamp_20130808050606.html
20130808 10:09:06 -NickServ- This nickname is registered. Please choose a different nickname, or identify via /msg NickServ identify <password>.
20130808 10:09:12 * jogamp (~jogamp@anon) has joined #jogamp
20130808 10:09:12 * Topic is 'http://jogamp.org | Hacking 3D Graphics, Multimedia and Processing across Devices'
20130808 10:09:12 * Set by rmk0 on 20130116 23:58:04
20130808 10:09:12 -NickServ- You are now identified for jogamp.
20130808 10:10:45 * sgothel (~sven@anon) has joined #jogamp
20130808 10:10:45 * ChanServ sets mode +v sgothel
20130808 10:16:06 <sgothel> Note: FRAND like royalties are far below 2.25% per device, a chip or phone. 2.25% was the magic number between Google and Moto.
20130808 10:16:21 <sgothel> http://antitrust.weil.com/articles/district-court-enters-the-frand-fray/
20130808 10:16:46 <sgothel> "Using these indicators, the court set a FRAND royalty rate of 0.555 cents per unit (with a range of 16.389 cents to 0.555 cents) for one Motorola SEP portfolio and a FRAND royalty rate of 3.471 cents per unit (with a range of 19.5 to 0.8 cents) for another Motorola SEP portfolio, resulting in approximately $1.8 million in annual royalty payments. This is far lower than Motorola’s proposed royalty rate of 2.25% of the end product pric
20130808 10:17:38 <sgothel> My conclusion .. setting up a royalty price tag of 60 cents .. is like to say "Don't even consider using our stuff!"
20130808 10:24:00 <sgothel> Besides .. they don't event use the device's pro-rated net costs, volume and surplus into account .. quite unprofessional for the embedded space IMHO.
20130808 10:37:13 * gouessej (546051e4@anon) has joined #jogamp
20130808 10:37:28 <gouessej> hi
20130808 10:37:34 <sgothel> ho
20130808 10:38:30 <gouessej> Sven, do you think JOGL and GlueGen can work by being loaded into 2 separate class loaders?
20130808 10:39:03 <sgothel> oh dear, all those complicated things in the morning :)
20130808 10:39:15 <sgothel> if they are daisy chained .. sure
20130808 10:39:30 <gouessej> What do you mean exactly?
20130808 10:39:35 <sgothel> myapp-cl [ jogl-cl [ gluegen-cl ] ]
20130808 10:39:41 <gouessej> yes
20130808 10:39:49 <sgothel> you can create a CL w/ a parent CL ..
20130808 10:39:56 <sgothel> sure
20130808 10:40:01 <gouessej> that's what I suggested in my latest post
20130808 10:40:06 <sgothel> OSGi .. right ?
20130808 10:40:09 <gouessej> yes
20130808 10:40:20 <sgothel> I agree ..
20130808 10:40:26 <gouessej> the class loader of GlueGen must be the parent of the class loader of JOGL
20130808 10:40:32 <sgothel> yup
20130808 10:41:03 <gouessej> it means some native libraries used by JOGL should be loaded in JOGL rather than in GlueGen ???
20130808 10:41:39 <sgothel> the daisy chain simply would allow to use the gluegen-cl w/o jogl .. if somewhat modularized
20130808 10:41:54 <sgothel> if used all-together .. it doesn't really matter
20130808 10:42:49 <sgothel> AFAIK Wade's extension to GlueGen's jar URL location is used for those purposed .. but I don't know really
20130808 10:43:02 <sgothel> so each CL can look into some special namespace / location
20130808 10:46:13 <gouessej> ok
20130808 10:46:40 <gouessej> Actually, it should already work as is
20130808 10:47:56 <gouessej> I have a good piece of news
20130808 10:48:01 <gouessej> about OSM
20130808 10:48:58 <gouessej> Last night, I succeeded in extracting a map, loading and viewing it in OSM2World. The conversion into OBJ works too :)
20130808 10:49:17 <sgothel> KUDOS!
20130808 10:50:40 <gouessej> My machine is too weak to manipulate large maps but it already works with half of the town in which I live close to Paris
20130808 10:51:13 <sgothel> yesterday we were thinking of having a Geiger Counter attached to a device .. are there phones w/ a Geiger Counter already ? folks could add their measurements via OSM :/
20130808 10:51:34 <sgothel> hmm .. so the map is not adapted to rendering performance .. yet ?
20130808 10:51:46 <gouessej> the problem isn't the rendering
20130808 10:51:50 <sgothel> i.e. reduced vertices etc .. ah
20130808 10:51:59 <sgothel> so the data itself does not scale ?
20130808 10:52:23 <sgothel> i.e. preprocessed to device capabilities ?
20130808 10:52:29 <gouessej> OSM2World keeps a lot of data in memory when I use it both for parsing and rendering
20130808 10:52:41 <gouessej> I will get something really sweet with Ardor3D
20130808 10:54:14 <gouessej> it would be even better with some indoor environments
20130808 10:54:20 <sgothel> not that familiar .. thought good scaling would apply to receive less data etc
20130808 10:54:44 <sgothel> Ardor3D .. doesn't that imply ffp ?
20130808 10:55:20 <gouessej> Ardor3D supports shaders too and I already made some changes to support the programmable pipeline too
20130808 10:55:43 <gouessej> a few parts don't support ES2
20130808 10:55:50 <sgothel> great! surely would be sweet for mobile devices as well ..
20130808 10:55:52 <sgothel> hmm
20130808 10:56:01 <gouessej> but anyway you can load the OBJ file with whatever you want
20130808 10:56:16 <sgothel> maybe switch to own 'engine' at some point in time
20130808 10:56:31 <sgothel> so OSM delivers meshes/tris via OBJs ?
20130808 10:56:40 <gouessej> Yes, feel free to do so
20130808 10:56:47 <sgothel> and granularity of data (resolution) can be chosen ?
20130808 10:57:01 <gouessej> OSM2World is able to export OSM maps into OBJ format
20130808 10:57:17 <gouessej> I'm not sure you can customize the granularity
20130808 10:57:31 <sgothel> and OSM maps are triangles/textures .. and 'fixed size' ?
20130808 10:57:31 <gouessej> but it is possible to modify the data in input
20130808 10:57:57 <sgothel> the custom granularity/res on the server side would allow devices to deal w/ data much better
20130808 10:58:04 <gouessej> OSM files are just plain XML files, they don't contain explicit meshes
20130808 10:58:52 <sgothel> so they incl. curves ? I mean .. they must contain something .. ?
20130808 10:59:02 <gouessej> they contain some nodes
20130808 10:59:16 <gouessej> those nodes are used to compute several polygons
20130808 11:00:28 <gouessej> I want to use Ardor3D in order to be able to use these maps with some other meshes in interactive worlds
20130808 11:00:45 <gouessej> not just for showing an empty map with no car, no people
20130808 11:01:26 <sgothel> yup
20130808 11:01:54 <gouessej> If you really want to show some stuff for a demo, just take the best engine for this purpose
20130808 11:02:03 <gouessej> yours if you want
20130808 11:03:07 <gouessej> I advise you to carefully choose your maps
20130808 11:03:20 <gouessej> I had to modify the loader to support some rotten eggs
20130808 11:05:40 <gouessej> Ardor3D still requires some work to support ES2
20130808 11:12:01 <sgothel> surely will be interesting working on our new Graph UI thingy .. I am sure your input will help alot, since you do have worked w/ many of those engines.
20130808 11:12:25 <sgothel> until then (or you start it now) .. jme3 .. ?
20130808 11:12:43 <gouessej> I already fixed some bugs in JME3 Sunday
20130808 11:12:52 <gouessej> especially the problem with DDS
20130808 11:13:32 <sgothel> oh .. a question which was raised .. your work (jme3, libgdx) .. is now marger and maintained by the projects ?
20130808 11:13:38 <sgothel> *merged*
20130808 11:14:05 <gouessej> For JME3 yes, for LibGDX It think it is not the case but Xerxes might have merged something
20130808 11:14:33 <gouessej> For Ardor3D, there is only one change not yet merged
20130808 11:14:40 <sgothel> KUDOS for jME3
20130808 11:14:40 <xranby> gouessej: libgdx is still in github merge request queue
20130808 11:14:49 <sgothel> libGDX .. they don't like it ?
20130808 11:14:59 <gouessej> I don't know
20130808 11:15:11 <sgothel> maybe an email could clarify ..
20130808 11:15:22 <xranby> let me give the link
20130808 11:15:27 <sgothel> often these github pull req. are ignored :)
20130808 11:15:44 <xranby> https://github.com/libgdx/libgdx/pull/234
20130808 11:16:06 <xranby> its not ignored entirely.. but the port had to be updated to match the upstream api
20130808 11:16:23 <xranby> also the authors requests a possible cleanup of the merge
20130808 11:16:39 <sgothel> sounds great actually .. i.e. they like to merge
20130808 11:17:26 <sgothel> sure .. needs a 'final commitment' for a proper 'handover'
20130808 11:17:39 <xranby> right now it needs a 2.0.2 release bump
20130808 11:18:29 <gouessej> I have to take care of JMonkeyEngine 3 too as JMonkeyEngine 3.1 will probably use the JOGL backend instead of the other one in desktop environments
20130808 11:18:52 <sgothel> hear .. hear ..
20130808 11:19:23 <xranby> it possible that i need to do a new merge branch without any added binarys to the git to address the @assets and @binaries request
20130808 11:19:50 <xranby> that would make the final merge easier
20130808 11:19:50 <xranby> and prevent their git to grow
20130808 11:20:29 <xranby> a merge also implies that the build system should work
20130808 11:21:36 <xranby> one todo is how to handle the libgdx natives
20130808 11:22:35 <gouessej> Do they want to keep applet support?
20130808 11:23:50 <xranby> i dont think it a deal breaker if applets are not yet working fully
20130808 11:23:56 <xranby> they do atleast compile and run
20130808 11:24:22 <sgothel> Don't see a reason why NApplet w/ NEWT shall not work ..
20130808 11:24:27 <xranby> i aknowledge that we still have issues with AWT applets
20130808 11:24:43 <sgothel> do we ? in JOGL ?
20130808 11:24:46 <xranby> currently the applet code is using AWT
20130808 11:24:49 <xranby> in libgdx
20130808 11:24:58 <sgothel> our NEWT/AWT applet .. I mean ..
20130808 11:25:00 <xranby> the libgdx-jogl- backend
20130808 11:25:23 <sgothel> i.e. our applet-tests all use NEWT .. via the NewtCanvasAWT .. etc
20130808 11:25:36 <sgothel> later .. w/ JWeb .. we can have this pure NEWT ofc
20130808 11:26:06 <xranby> sgothel: i and gouessej are talking about this issue specific to swing & applet interoperability using the jogl libgdx backend http://forum.jogamp.org/JOGL-2-0-OpenGL-OpenGL-ES-backend-for-LibGDX-td4027689i20.html
20130808 11:26:45 <sgothel> OSX offset .. I assume ?
20130808 11:26:52 <xranby> this was on linux
20130808 11:27:32 <sgothel> oh .. you mean something GL (GLCanvas for example) being using with a 'complicated' Swing component.
20130808 11:27:34 <sgothel> ?
20130808 11:27:37 * chakie (~chakie@anon) has joined #jogamp
20130808 11:28:09 <sgothel> i.e. scrolling etc .. where the inner GL component shall provide an image .. i.e. heavy-/lightweight mixing related
20130808 11:28:32 <xranby> sgothel: libgdx includes a lot of tests... two of the tests JoglAppletTest and SwingJoglTest expose this issue the rest of the tests runs fine
20130808 11:28:41 <sgothel> in those Swing cases .. well .. only GLJPanel will really work .. sadly, especially in Java7 and/or OSX
20130808 11:28:44 <xranby> i have not studied the details
20130808 11:30:21 <xranby> ok maybe we should update the backedn code to use GLJPanel for the libgdx jogl backend
20130808 11:30:32 <xranby> for AWT and Swing use
20130808 11:30:42 <sgothel> (Swing only)
20130808 11:31:03 <gouessej> chakie, do you succeed in using the text renderer?
20130808 11:32:37 <gouessej> Does someone plan to look at this anew soon? https://github.com/sgothel/jogl/pull/47
20130808 11:32:53 <xranby> we are currently using a NewtCanvasAWT in libgdx/backends/gdx-backend-jogl/src/com/badlogic/gdx/backends/jogl/JoglAWTCanvas.java
20130808 11:34:38 <sgothel> pull #47 - oh dear - they didn't clean it up .. they pledged to do so .. - hence no maintainer of that patch
20130808 11:35:49 <sgothel> @Xerxes: as long you use opaque things w/o require heavy- lightweight Swing mixing this is great. otherwise .. you may need to use GLJPanel.
20130808 11:36:28 <sgothel> I don't really see the point for libgdx to use swing .. well .. dunno
20130808 11:36:32 <gouessej> it still uses some classes that do the same than our shader utilities :s
20130808 11:37:23 <xranby> sgothel: all the main libgdx examples uses NEWT GLWindow direcly like the stuff you saw at fosedm and siggraph on the Pi
20130808 11:37:58 <sgothel> thats great then .. me was confused about mentioned Swing here ..
20130808 11:38:03 <xranby> hence i its a low priority for the merge
20130808 11:38:44 <xranby> swing and awt is only used for A) applets and B) putting a libgdx game inside a swing app
20130808 11:38:57 <gouessej> ok but nobody will use this backend if it is not merged
20130808 11:39:34 <gouessej> lol I forced the use of JOGL by accident some days ago in JMonkeyEngine, only one person complained
20130808 11:39:43 <sgothel> for applets .. we could use NEWT/AWT applet 'logic' JOGLNewtApplet1Run .. for Swing .. well :)
20130808 11:40:07 <sgothel> KUDOS :)
20130808 11:41:56 <gouessej> lol I had to revert my commit but anyway, it proved that switching to JOGL doesn't cause the end of the world :)
20130808 11:54:16 * gouessej (546051e4@anon) Quit (Quit: Page closed)
20130808 12:12:02 <xranby> sgothel: updated https://jogamp.org/bugzilla/show_bug.cgi?id=698#c7 with an attached screenshot
20130808 12:22:43 <xranby> gouessej: btw. openjfx claims to be fully opensourced now example how to cross compile it for the raspberry pi https://wiki.openjdk.java.net/display/OpenJFX/Building+OpenJFX#BuildingOpenJFX-CrossbuildingforARMhardfloat
20130808 12:24:45 <sgothel> @Xerxes: Great stuff .. - and above shows .. they are everything but 'cheap' :)
20130808 12:25:57 <sgothel> a subsequent logical question to Henrik might be: So when do you opensource ARM32 hotspot .. similar to Intel, .. CPU support ?
20130808 12:26:03 <sgothel> :]
20130808 12:36:59 * xranby (~xranby@anon) Quit (Ping timeout: 260 seconds)
20130808 12:39:26 * xranby (~xranby@anon) has joined #jogamp
20130808 12:39:43 <xranby> sgothel: fell free to join the forum and ask the question!
20130808 12:39:48 <xranby> henrik have answered the question last year
20130808 12:39:52 <xranby> https://blogs.oracle.com/henrik/entry/oracle_releases_jdk_for_linux
20130808 12:39:56 <xranby> "Is the Oracle JDK port to ARM available in OpenJDK? No, and we are not planning on open sourcing it at this point."
20130808 12:40:12 <xranby> Aug 14, 2012
20130808 12:41:09 <sgothel> thx - sure .. it's more a rhetorical one :) .. sure .. they still try to monetize it .. even though the 60 cents seem to proof the opposite for high scale deployment :)
20130808 12:42:10 <xranby> interesting compare to the FRAND legal cases
20130808 12:43:03 <xranby> http://jogamp.org/log/irc/jogamp_20130808100906.html#l9
20130808 12:44:19 <sgothel> .. at least, their tech is easy to avoid now :) .. hence no monopoly case
20130808 12:45:09 <xranby> are we done with bug 698? or do we need more confirmations?
20130808 12:46:27 <sgothel> I am not a lawyer :) precedence is missing. You think this statement is really a legal statement from Oracle ?
20130808 12:46:42 <xranby> yes that IS the question
20130808 12:46:46 <sgothel> Maybe we should ask Oracle officially to confirm ?
20130808 12:47:38 <sgothel> .. i.e. including derivations of OpenJDK ?
20130808 12:48:09 <xranby> sure, that is still a question
20130808 12:48:33 <xranby> ok leaving 698 open untill we know if we may ship derivations of OpenJDK
20130808 12:48:39 <sgothel> i.e. what _is_ OpenJDK
20130808 12:49:20 <sgothel> i.e. our JiGong .. openjdk8 .. push our profiles .. maybe patched somehow while complying w/ the GPLv2 ofc.
20130808 12:49:45 <sgothel> example: modify the profiles ... add/remove something .. etc
20130808 12:49:58 <xranby> its possible we may keep jigong GPLv2 + classpath exception
20130808 12:50:08 <sgothel> ofc!
20130808 12:50:22 <xranby> like it is stated in the OpenJDK 8 source
20130808 12:50:23 <sgothel> I mean .. the pledge of Oracle: It's OK!
20130808 12:51:07 <sgothel> i.e. Oracle's statement: we won't apply our FUD (patent grant not for mobile, we may sue you) on this ..
20130808 12:51:14 <sgothel> GPLv2 wise .. sure we are in the clear
20130808 12:52:14 <sgothel> Henrik hinted in that direction .. that it is OK, but maybe one could elaborate on this .. to make people and companies feel safe
20130808 14:40:33 * xranby (~xranby@anon) Quit (Quit: Leaving.)
20130808 19:35:08 <sgothel> http://lavabit.com/
20130809 09:04:24 * xranby (~xranby@anon) has joined #jogamp
20130809 09:44:30 <xranby> good morning: QOTD: "(11:42:37) mjw: xranby, BTW. I looked where I could donate to jogamp but didn't find anything.
20130809 09:44:30 <xranby> (11:42:41) mjw: you only take code? :)"
20130809 09:47:34 <sgothel> good morning .. me currently cleaning up the mediamplayer .. multithreaded etc
20130809 09:59:02 <xranby> mjw have some good suggestions on how to handle donations for jogamp similar to how icedtea currently do it w/o forming a legal body..
20130809 09:59:22 <xranby> i will send an email containing the discussion i had in #openjdk
20130809 09:59:37 <sgothel> which are ? so this is posted personally ?
20130809 10:00:07 <sgothel> I was thinking about all those crowd things .. but then .. why giving 'em 10% ?
20130809 10:00:08 <xranby> mjw, Mark Wielaard, runs the icedtea servers and is the person behind: http://mail.openjdk.java.net/pipermail/discuss/2010-January/001451.html
20130809 10:00:50 <xranby> all details by email shortly
20130809 10:01:07 <xranby> yes.. i agree letting flattr get 10% for all worlds donations is insane
20130809 10:02:08 <sgothel> or other crowdfunding resources ..
20130809 10:02:14 <sgothel> all similar
20130809 10:02:29 <sgothel> flattr as well ? oh .. didn't read the fineprint .. damn :)
20130809 10:03:02 <sgothel> fineprint & banks: http://rt.com/business/man-outsmarts-banks-wins-court-221/ :)
20130809 10:04:37 <xranby> :D
20130809 10:06:31 <xranby> ok email sent
20130809 10:06:40 * kermyt (~kermyt@anon) Quit (Ping timeout: 245 seconds)
20130809 10:12:21 <sgothel> dunno though whether the FSF expressed 'wish' of an implicit patent grant is valid .. but yes, it 'increases' confidence .. somewhat
20130809 10:12:29 <sgothel> .. reading your chat ..
20130809 10:13:32 <sgothel> archive: our bug reports ? :)
20130809 10:16:03 <xranby> mjw: suggested at creating a html copy of the texts.. like the old sun legal page that he have archived at his own server: http://icedtea.classpath.org/openjdk/java/faq.jsp.html
20130809 10:16:09 <sgothel> I let this sink .. (guess mjw is hinting to put jogamp into an umbrella org ..) .. this motivates me .. again .. to have us have our own foundation ..
20130809 10:16:34 <sgothel> yup .. maybe within ji-gong .. doc/legal ?
20130809 10:16:50 <sgothel> i.e. ji-gong git is more like a meta doc / build git for this purpose ..
20130809 10:16:59 <xranby> sure
20130809 10:18:18 <xranby> good, communication fuels discussions, and brings refined distilled ideas
20130809 10:18:36 <xranby> cheers
20130809 10:21:52 <sgothel> ok .. called notary again .. will get a call back .. uh
20130809 10:22:46 <sgothel> guess we need 'founders' for the foundation .. guess you are 'in' .. ? will give you details after chat w/ notary.
20130809 10:24:22 <sgothel> btw .. lavabit (good hearted US email provider) closed due to NSA requests, which he cannot disclose :/
20130809 10:26:45 <xranby> national interest loopholes undermining democracy :/
20130809 10:27:07 <sgothel> sure .. but the irony is: it also undermines their own economy :)
20130809 10:28:17 <sgothel> haven't heard from those disclosure letters from the EU yet .. hope the EU won't pick that up, usually they copy all things evil .. 8-]
20130809 10:29:13 <sgothel> https://sks-keyservers.net/status/generate_key_chart.php
20130809 10:38:33 * gouessej (546051e4@anon) has joined #jogamp
20130809 10:42:26 <gouessej> hi
20130809 10:42:37 <sgothel> hi ho
20130809 10:44:02 <gouessej> nice link about lavabit
20130809 10:44:37 <gouessej> The OBJ file is ready, feel free to use it whenever you want
20130809 10:45:27 <gouessej> http://svn.code.sf.net/p/tuer/code/pre_beta/obj/planet_2.413_48.853_f55ff5ae.osm.obj
20130809 10:46:13 <gouessej> Ardor3D minimized ther vertices count but it is only reflected in the .abin file
20130809 10:47:22 <gouessej> Do you think Dominik could help us to make something beautiful from this rudimentary model?
20130809 10:51:38 <sgothel> obj .. can you show us a snapshot ? no obj reader/app here at hand ..
20130809 10:52:02 <sgothel> Re Dominik .. sadly I have not heard from him for a while .. sure send him an email
20130809 10:53:13 <gouessej> I can send you a snapshot by email tonight
20130809 10:53:23 <gouessej> or now...
20130809 10:54:41 <gouessej> done
20130809 10:54:45 <gouessej> Lok at your emails
20130809 10:54:56 <gouessej> Look at the enclosed file osm.png
20130809 10:55:08 <gouessej> I took this capture in Blender 2.68a
20130809 10:55:28 <gouessej> Blender no longer works on my main machine :(
20130809 10:55:31 <sgothel> note to me: Julien can work w/ Blender :)
20130809 10:55:38 <gouessej> it requires SSE2 + OpenGL 2
20130809 10:56:08 <gouessej> lol I only can do a very few things with Blender
20130809 10:56:24 <gouessej> modify Python scripts, perform basic operations
20130809 10:56:35 <sgothel> it's a map alright .. nice - so your tool does the OSM -> OBJ ?
20130809 10:57:05 <gouessej> OSM2World (created by Tobias Knerr) performs this conversion
20130809 10:57:11 <sgothel> ah
20130809 10:57:32 <sgothel> and now you like to 'overlay' .. some other data .. to make it a game or .. whatever ?
20130809 10:57:33 <gouessej> I modified this tool because some of my polygons have no aread :s
20130809 10:57:53 <gouessej> I'm going to put such 3D maps into TUER
20130809 10:58:16 <gouessej> You can recognize Montreuil Gate close to Paris in the bottom left corner :)
20130809 10:58:31 <sgothel> hmm new chapter 'locate the economic hitman' :)
20130809 10:59:10 <gouessej> I assume that someone can still make something nice with this map for a demo, at FOSDEM or Siggraph
20130809 10:59:49 <gouessej> or maybe with a better map
20130809 10:59:53 <gouessej> LA or Paris
20130809 11:00:10 <sgothel> I would like to toy w/ it w/ our new data input stuff (camera, GPS, compass, ..) .. visualize things .. add photos ..
20130809 11:00:43 <sgothel> but the obj map .. well, maybe generated server side .. hmm dunno, we will see when playing w/ it
20130809 11:00:48 <gouessej> Is this format (OBJ) ok for you?
20130809 11:00:49 <sgothel> IMHO no need to have it static
20130809 11:01:34 <sgothel> I don't know .. me novice, .. I read that it's triangles right ? generated from those OSM polygons (no curves?!) .. hmm
20130809 11:01:43 <sgothel> (you told me yesterday)
20130809 11:02:11 <gouessej> you can modify the OBJ file, add some texture coordinates, use a texture for a particular face
20130809 11:02:15 <sgothel> so maybe we can setup a server side thingy .. to perform something there, if too heavy for the device
20130809 11:02:40 <sgothel> IMHO WW2 already does something like this w/ some OSM data server .. hmm
20130809 11:02:54 <gouessej> the map I gave you is well for a small device, especially the .abin
20130809 11:02:57 <sgothel> http://mapserver.org/
20130809 11:03:03 <sgothel> http://www.mapgears.com/en/blog/archive/2012-10-16-nasa-world-wind-mapserver
20130809 11:03:47 <sgothel> do you know this one ?
20130809 11:04:09 <gouessej> I didn't know this one
20130809 11:04:16 <gouessej> thanks for the links
20130809 11:05:33 <sgothel> so it is server based .. hmm, good - maybe we can utilize it to have the data generated dynamically scaled for device
20130809 11:06:34 <gouessej> Personally, I would use static meshes for all levels of details and load only the one I plan to use
20130809 11:06:46 <sgothel> sure .. game data
20130809 11:06:57 <sgothel> I was thinking of our JogMap application :)
20130809 11:07:27 <sgothel> or JogmAp :)
20130809 11:07:38 <gouessej> lol
20130809 11:09:33 <gouessej> In my humble opinion, adding photos is doable, the input stuff will obviously requires some more efforts
20130809 11:09:41 <gouessej> "require"
20130809 11:14:15 <gouessej> Are you sure you can move in 3D with MapServer? You get 2D images or SVG files in output
20130809 11:14:47 <sgothel> dunno .. got the hint via WW2 .. hence I thought they must have the 3D data
20130809 11:15:00 <sgothel> probably on of the other output formats
20130809 11:15:05 <sgothel> of a custom one ..
20130809 11:16:23 <gouessej> ok but in the documentation, they mention no 3D format
20130809 11:16:42 <gouessej> maybe WW does something with the data
20130809 11:18:08 <gouessej> the current meshes of my file are quite coarse. If you get too fine meshes, you can downscale them so that they fit into the constraints of your device
20130809 11:19:04 <sgothel> the idea of a scaling server .. is that you get an appropriate mesh depending on the viewport and device spec ..
20130809 11:19:28 <sgothel> viewport here in those globe coord. ofc
20130809 11:19:39 <gouessej> ok I see what you mean
20130809 11:20:24 <gouessej> but then, the file is transmitted once and the client installed on the device does the rest of the job, doesn't it?
20130809 11:21:37 <sgothel> oh .. cached yes .. depends on your use case ofc
20130809 11:21:55 <sgothel> for your game it might be fit to simply have a few resolutions available (mipmap style)
20130809 11:22:29 <sgothel> for JogmAp .. we might need to have it all dynamic .. while caching a few maps which are being used regularly ..
20130809 11:23:17 <gouessej> I will use some coarse maps with a few refined parts in my game. I will probably be unable to show big towns without space partitioning. I will have to look at jspatial
20130809 11:23:55 <gouessej> Wouldn't it be safer for a demo to have a few ready and scaled maps locally on the client?
20130809 11:24:11 <sgothel> cached .. sure
20130809 11:24:34 <sgothel> we did that for siggraph, i.e. cached all applets - that is why they showed up quite fast :)
20130809 11:26:04 <gouessej> If you know the syntax of the OBJ (+ MTL) file format, you can add photos without using a modeler.
20130809 11:27:52 <gouessej> Adding support of these new input stuff won't be trivial. Have you ever looked at my draft?
20130809 11:28:34 <sgothel> *hiding* *duck* .. - will do when tackling the other new input things
20130809 11:29:38 <gouessej> ok but I would feel better if you could just look at it quickly before I start implementing the NEWT Controller API
20130809 11:30:06 <gouessej> Compared to JInput, I'll just remove all things depending on DirectInput, I don't need them
20130809 11:30:19 <gouessej> Raw input API is ok
20130809 11:32:46 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=595 ?
20130809 11:32:58 <sgothel> https://jogamp.org/bugzilla/attachment.cgi?id=413&action=edit ?
20130809 11:34:03 <sgothel> newInputApi.odt ? this one ?
20130809 11:34:04 <gouessej> yes
20130809 11:34:40 <gouessej> it's a bit raw but it is understandable with the explanations I gave in the bug report
20130809 11:35:38 <sgothel> Controller == Joystick ?
20130809 11:36:33 <gouessej> yes, more or less
20130809 11:36:37 <sgothel> question arises: what is so different from a pointer ?
20130809 11:36:38 <gouessej> joypad
20130809 11:37:03 <gouessej> good question
20130809 11:37:07 <sgothel> i.e. pointer -> fixed coords, joystick/pad: amount in direction
20130809 11:37:42 <gouessej> I need to operate both on absolute and relative data in this API like in JInput
20130809 11:38:27 <gouessej> I made an effort to drive it optional, in case some people don't want to use it, it's not vital
20130809 11:40:04 <sgothel> looks like we could 'free' MouseEvent -> PointerEvent
20130809 11:40:27 <gouessej> Can you elaborate?
20130809 11:40:30 <sgothel> while the latter either uses a rel. vector move, or abs. new position
20130809 11:40:56 <sgothel> sure .. we would need to add more components, i.e. z axix (like we have for rotation)
20130809 11:41:14 <sgothel> the event could state that it is an abs. position ..
20130809 11:41:33 <gouessej> yes, I have to think about your suggestion
20130809 11:41:47 <sgothel> we could 'feed' in an origin to the consumer device (window) .. so an abs. pos could be derived
20130809 11:42:30 <sgothel> the ole MouseEvent .. we could still have it as deprecated (derived from PointerEvent) and remove it later
20130809 11:42:54 <sgothel> while at it .. 'Pointer' might be changed to a more generic name .. hmm
20130809 11:43:39 <sgothel> In the end, this would unify the event type, instead of exploding things ..
20130809 11:43:51 <gouessej> In that case, a relative movement could be derived in an absolute movement by using the consumer window
20130809 11:43:59 <sgothel> yup
20130809 11:44:22 <sgothel> well, currently events are being send to the consuming window
20130809 11:44:24 <gouessej> maybe we should look at existing standards too
20130809 11:44:33 <sgothel> how to determine this for joystick etc .. dunno, i.e. focused window ?
20130809 11:44:45 <gouessej> yes
20130809 11:44:49 <sgothel> great
20130809 11:45:01 <gouessej> for security reasons I will only forward events to the focused window
20130809 11:45:31 <gouessej> If some people wants to write key loggers, they can use JInput instead
20130809 11:45:39 <sgothel> :)
20130809 11:45:54 <sgothel> I like that .. generalize Mouse/Pointer event ..
20130809 11:46:09 <sgothel> generalize it's buttons .. it's axis .. etc
20130809 11:46:21 <sgothel> we started that .. but did not complete yet
20130809 11:46:21 <gouessej> Do you know a W3C norm about events of joypads?
20130809 11:46:28 <sgothel> nope ..
20130809 11:47:08 <xranby> the USB spec contains a quite detailed list of various input devices, all from joypads to golf clubs & magic carpets
20130809 11:47:19 <sgothel> hehe
20130809 11:47:22 <gouessej> Where is PointerEvent?
20130809 11:47:31 <sgothel> damn it .. I missed to add the W3C pointer spec reference ..
20130809 11:47:37 <sgothel> -> MouseEvent :)
20130809 11:47:48 <sgothel> it still has it's old name .. well
20130809 11:47:59 <xranby> was catout/jogamp operational when we did discuss this hmm?
20130809 11:48:29 <sgothel> http://www.w3.org/Submission/pointer-events/#pointerevent-interface
20130809 11:48:34 <gouessej> :)
20130809 11:48:36 <sgothel> (adding to source code .. ahem)
20130809 11:49:29 <sgothel> missing: tiltXY (need to add this as well)
20130809 11:49:43 <sgothel> then generalize for rel-pointers (joystick ..)
20130809 11:49:50 <sgothel> I guess .. then we are fine
20130809 11:50:02 <xranby> http://www.usb.org/developers/devclass_docs/Hut1_11.pdf <- HID Usage Tables incl. the magic carpet spec
20130809 11:50:07 <sgothel> sure .. the annoying rename MouseEvent -> PointerEvent for 2.1.0 :)
20130809 11:50:25 <sgothel> lol
20130809 11:50:39 <gouessej> Do you think it could replace all classes Controller*Event?
20130809 11:50:48 <xranby> page 42 :)
20130809 11:50:59 <sgothel> .. reading
20130809 11:51:26 <sgothel> well .. lets discuss data 1st .. then the user-events
20130809 11:51:38 <sgothel> so PointerEvent would hold:
20130809 11:52:26 <sgothel> per pointer: abs-axis-data[], rel-axis-move[], rel-axis-rot[], axix-tilt[], pressure
20130809 11:52:47 <sgothel> missing today: rel-axis-move[] and axix-tilt[]
20130809 11:53:33 <gouessej> there is just a weird thing
20130809 11:53:38 <sgothel> your events are pretty elaborated .. guess we can reduce them to bare minimum
20130809 11:54:14 <gouessej> sometimes, you just want the relative value and the absolute value has no meaning because it would/could go beyond the window
20130809 11:54:40 <sgothel> we have pointer-type as well, so application can make their choice
20130809 11:55:18 <gouessej> yes, you're right
20130809 11:55:19 <sgothel> well, we could still create an abs-pos beyond the window .. hmm .. details
20130809 11:55:36 <sgothel> but probably not, i.e. clip to window size
20130809 11:55:52 <sgothel> ensures same behavior w/ mouse for example
20130809 11:56:00 <gouessej> yes but it is tricky, you have a focused window who receives an event whereas its "pointer" is outside :s
20130809 11:56:09 <gouessej> ok
20130809 11:56:11 <sgothel> I know ..
20130809 11:56:23 <sgothel> we can vote on this :)
20130809 11:56:44 <sgothel> I don't care .. so either clipped or allow external abs coords .. as you wish
20130809 11:56:50 <gouessej> but the documentation has to remind that the absolute value is meaningless when the pointer type indicates that the pointer is not visible
20130809 11:57:10 <sgothel> again ?
20130809 11:57:23 <sgothel> 'not visible' ?
20130809 11:57:33 <sgothel> (touch / joystick are never visible)
20130809 11:57:42 <gouessej> not always
20130809 11:57:57 <gouessej> in some cases you may want to move the pointer with the joypad
20130809 11:58:17 <sgothel> the 'pointer cursor' you mean ?
20130809 11:58:25 <gouessej> yes sorry
20130809 11:58:47 <sgothel> hmm .. yes then it's like that .. dunno how to do it though :)
20130809 11:58:55 <sgothel> this could be done as an user app, warpPointer ..
20130809 11:59:11 <sgothel> or even internally .. assign pointer->cursor .. wow :)
20130809 11:59:28 <sgothel> guess we leave this up to later .. hehe
20130809 11:59:41 <sgothel> but keeping it in mind ..
20130809 11:59:46 <gouessej> the developper can forward a controller event as a mouse event
20130809 12:00:15 <sgothel> but to move the cursor .. you would need to use warpPointer ..
20130809 12:00:29 <sgothel> ok, so all this can be done explicitly .. no need to add in core API for now
20130809 12:00:52 <gouessej> yes, this particular aspect shouldn't be in the core
20130809 12:00:58 <gouessej> but what about the rest?
20130809 12:00:59 <sgothel> you really want to get those pointer-connection events ?
20130809 12:01:17 <sgothel> -> controllerAdded/Removed
20130809 12:01:18 <gouessej> not yet, there is no hurry but I want a nice doc
20130809 12:01:29 <gouessej> controllerAdded/removed yes
20130809 12:01:32 <gouessej> it is required
20130809 12:01:39 <gouessej> for video game consoles
20130809 12:01:49 <gouessej> and even on a computer
20130809 12:01:52 <sgothel> so based on data .. I guess we would be complete w/ enh. PointerEvent
20130809 12:02:16 <sgothel> sure .. we can add those transient events
20130809 12:03:01 <sgothel> indeed we would need a PointerContext per Pointer-ID which then keeps track of things (abs/rel coords .. etc)
20130809 12:03:07 <sgothel> and allows user to set origin .. etc ..
20130809 12:03:10 <gouessej> it is not the most important feature as nothing tells you when I don't use my mouse
20130809 12:03:20 <sgothel> i.e. now we need to do the user query thingy
20130809 12:04:17 <sgothel> We can use a statically shared pool for those PointerContexts ..
20130809 12:04:52 <sgothel> one context can be added when an event happens (mouse) if context does not exist yet
20130809 12:05:03 <sgothel> or when being plugged in .. queried devices at startup time ..
20130809 12:05:38 <gouessej> ok
20130809 12:05:45 <sgothel> probably turns out to be an InputDeviceContext .. for keyboard as well -> Xerxes's multi-seat thingy :)
20130809 12:06:38 <sgothel> so .. 'PointerDevice implements InputDevice' .. etc
20130809 12:06:50 <gouessej> Do you really think we can put some information about buttons, sliders, axis, pov and vector3 into PointerEvent/MouseEvent
20130809 12:06:51 <gouessej> ?
20130809 12:07:08 <sgothel> buttons .. we already have
20130809 12:07:16 <sgothel> sliders ?
20130809 12:07:24 <gouessej> ok then this API wouldn't be optional
20130809 12:07:51 <sgothel> what are sliders ?
20130809 12:08:26 <gouessej> it is in JInput and OOIS...
20130809 12:08:29 <sgothel> potentiometers ?
20130809 12:08:33 <gouessej> I have to remember what I meant
20130809 12:08:50 <sgothel> analogue resistor ?
20130809 12:09:05 <sgothel> that would be the rel. axis move ..
20130809 12:09:06 <gouessej> it is a kind of component
20130809 12:09:14 <gouessej> yes
20130809 12:09:27 <sgothel> a digital joystick would only do a '1' for a given direction
20130809 12:09:54 <sgothel> to get the abs-pos one would need to attach a plugin, i.e. considering speed/accel. :)
20130809 12:10:46 <sgothel> ( but we can postpone abs-pos to later for such devices ..)
20130809 12:11:02 <gouessej> slider is a kind of axis
20130809 12:11:08 <gouessej> different of pov
20130809 12:11:11 <sgothel> we have rotation as well ..
20130809 12:11:34 <sgothel> axisMoved == rotation
20130809 12:11:38 <gouessej> pov -> 9 directions, slider -> abscissa + ordinate
20130809 12:12:00 <sgothel> don't get it .. sorry
20130809 12:12:08 <sgothel> pov == point of view ?
20130809 12:12:13 <gouessej> I forgot to add a part of the document :s
20130809 12:13:13 <gouessej> pov is a component of a joypad, a tiny joystick
20130809 12:13:33 <gouessej> typical on Dual Shock and GameCube gamepads
20130809 12:15:12 <sgothel> hmm .. ok .. creating a new bug for enhancing/fix PointerEvent .. trying to put the above on it.
20130809 12:15:20 <sgothel> will post here when done ..
20130809 12:15:45 <gouessej> my main source of inspiration is this API: https://github.com/wgois/Object-oriented-Input-System--OIS-
20130809 12:17:08 <gouessej> sliding axis seem to be supported only under Win32, not very useful :s
20130809 12:20:57 <xranby> (14:07:00) sgothel: probably turns out to be an InputDeviceContext .. for keyboard as well -> Xerxes's multi-seat thingy :) <--- good memory! yes it would be preferably to have a way to differentsiate mouse pointers with n-mice attached to a system
20130809 12:21:15 <xranby> and n-keyboards
20130809 12:21:33 <xranby> and n-<insert your hid device here>
20130809 12:21:38 <xranby> magic carpets etc
20130809 12:24:40 <gouessej> Sven, I assume I will have to clarify the role of the events in this API
20130809 12:27:01 <gouessej> What means MouseEvent.PointerClass.Offscreen?
20130809 12:28:47 <gouessej> you don't have to touch the screen to use the Offscreen things, do you?
20130809 12:30:18 <sgothel> @Xerxes: no great idea gets lost :)
20130809 12:30:39 <sgothel> Offscreen pointer class is e.g. the mouse :)
20130809 12:31:14 <xranby> for multiseat the end user may want to know the star topology of attached devices
20130809 12:31:34 <sgothel> keyboard w/ GPS .. of course :)
20130809 12:31:37 <xranby> that is a mouse attached to a usb hub sitting on a keyboard may be percieved by the user to be one seat
20130809 12:31:48 <xranby> usb know the topology
20130809 12:32:00 <sgothel> holy moly ..
20130809 12:32:02 <gouessej> What would be the pointer type of a joypad?
20130809 12:32:15 <xranby> the hard part is to give the topology of attached devices to the app
20130809 12:32:25 <sgothel> Pi.USB -> splitter -> n * [monitor.usb - [ mouse, kbd ] ] :)
20130809 12:32:42 <sgothel> easy as that ? lol
20130809 12:32:53 <sgothel> then monitor device needs to get the usb-id :)
20130809 12:33:15 <sgothel> @Julien: Good question!
20130809 12:33:26 <xranby> plugable.com is one of the driving actors for multiseat
20130809 12:33:26 <sgothel> sure they are offscreen
20130809 12:33:32 <gouessej> PointerClass -> Offscreen like the mouse
20130809 12:33:33 <xranby> http://plugable.com/2009/11/16/setting-up-usb-multiseat-with-displaylink-on-linux-gdm-up-to-2-20
20130809 12:33:52 <gouessej> PointerType -> Undefined (temporarily)
20130809 12:34:38 <xranby> thus you may purcase these USB seat thingies from them
20130809 12:34:42 <xranby> and create a mess
20130809 12:35:08 <sgothel> JoyPad == Digital Joystick ?
20130809 12:35:08 <xranby> http://dxg49ziwjgkgt.cloudfront.net/wp-content/uploads/2009/10/2009-09-22-195134-fourseats.JPG
20130809 12:35:38 <sgothel> @Xerxes: Lets do this _after_ this change we discuss .. :)
20130809 12:35:50 <xranby> aknowledge
20130809 12:36:01 <sgothel> in the meantime .. you may add the USB query thingy .. hehe
20130809 12:36:21 <sgothel> (02:35:08 PM) sgothel: JoyPad == Digital Joystick ?
20130809 12:36:25 <xranby> i only have native jni code to do the query atm
20130809 12:36:53 <sgothel> i.e. keys for up, down .. (digital like the ole joysticks)
20130809 12:36:58 <sgothel> ?
20130809 12:37:04 <xranby> i have the query code to scan the usb device tree to find the medical devices we use at zafena
20130809 12:37:23 <sgothel> w/ whole topology like lsusb ?
20130809 12:37:24 <xranby> so sure this is on my table so to speak
20130809 12:37:29 <xranby> yes i use libusb
20130809 12:37:30 <sgothel> cool
20130809 12:37:42 <sgothel> linux: easy .. windows/osx ?
20130809 12:37:51 <xranby> i have code for windows as well
20130809 12:37:55 <xranby> using windows driver model
20130809 12:38:02 <sgothel> for sure that part would be very welcome !!!
20130809 12:38:15 <xranby> can scan and attach to hid windows devices without having to installe xtra drivers
20130809 12:38:21 <sgothel> then indeed we can do what you suggest .. auto-assigment of InputDevice -> monitor
20130809 12:38:46 <sgothel> so I add the auto-assignment of input .. hmm .. in a more generic way .. as a new bug entry! cool!
20130809 12:38:50 <xranby> this using jni to scan windows driver model without using libusb
20130809 12:39:01 <xranby> targeting the windows api directly
20130809 12:39:21 <xranby> osx.. no idea
20130809 12:39:24 <sgothel> if you can make this (windows/linux) create a graph in java (node-topo) .. awesome!
20130809 12:39:28 <xranby> i hope libusb can ahndle it
20130809 12:41:22 <xranby> the only trouble building it for windows is that you have to get the driver sdk from microsoft web website, hence the licence needs to be scrunitized
20130809 12:41:24 <gouessej> a joypad is more or less a gamepad, think about Dual Shock PS2 & 3, Gamecube
20130809 12:41:53 <gouessej> I think we should talk about that to the creator of JInput
20130809 12:42:26 <sgothel> sorry .. I don't know all these brand names ..
20130809 12:42:37 <sgothel> @Xerxes: which bugzilla account ? guidinna ?:
20130809 12:43:07 <xranby> on jogamp bugzilla use the gudinna email
20130809 12:43:54 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=812 you got it Sir :)
20130809 12:44:05 <sgothel> *filling bugzilla task for next siggraph :)
20130809 12:48:10 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=813
20130809 12:49:20 <gouessej> lol
20130809 12:50:28 <sgothel> https://jogamp.org/bugzilla/showdependencytree.cgi?id=814&hide_resolved=1
20130809 12:50:37 <sgothel> the web of bugs :)
20130809 12:51:18 <sgothel> I guess this is a good starting point
20130809 12:51:30 <sgothel> sure we will get more knowledge / use-cases as we go
20130809 12:51:49 <sgothel> we may also add new derived PointerEvent types .. maybe .. we will see
20130809 12:52:21 <xranby> JDM JogAmp developers online meeting is on! Goal: file blueprints for the ext 12month cycle
20130809 12:52:35 <sgothel> haha
20130809 12:52:49 <sgothel> closed: NOW :)
20130809 12:53:05 <sgothel> resources: exceeded :)
20130809 12:53:20 <sgothel> where is Mark ? @Mark .. :)
20130809 12:53:29 <sgothel> they all hide under the keyboard ..
20130809 12:53:33 <xranby> rmk0: ^
20130809 12:54:21 <sgothel> @Julien: Thank you pushing this .. now I know, I was procrastinating all this due to me realizing the amount of work required :)
20130809 12:55:12 <gouessej> I have to talk about that to the creator of JInput as he has a lot of experience in this topic and I will probably reuse some parts of his source code
20130809 12:55:32 <gouessej> I'd like to know his opinion
20130809 12:55:35 <rmk0> lo
20130809 12:55:57 <rmk0> working on lighting
20130809 12:56:01 <sgothel> Sure .. Bug 814 only gives one some vehicle to transport the data to the user .. that's all, it still must be delivered/created :)
20130809 12:56:10 <rmk0> http://waste.io7m.com/2013/08/09/proradii.png
20130809 12:56:13 <sgothel> @Mark: Light under the keyboard ? :)
20130809 12:56:23 <gouessej> Sven, should I add some information into the bug report 814?
20130809 12:56:31 <sgothel> Sure
20130809 12:56:45 <sgothel> 814 or the JInput .. where it fits best
20130809 12:56:57 <gouessej> Why do you want to use 2 pointer types for joypads?
20130809 12:57:01 <sgothel> i.e. detailed info about your devices and hence data requirements .. in 814
20130809 12:57:12 <sgothel> read the questionmark ? :)
20130809 12:57:21 <gouessej> I just want to get some advises from an expert, he is open minded enough to answer me
20130809 12:57:27 * rmk0 reads up
20130809 12:57:28 <sgothel> nice
20130809 12:57:43 <sgothel> @Mark: quite environmental :)
20130809 12:58:33 <sgothel> per pixel .. ? gauraud .. ?
20130809 12:58:40 <sgothel> (all those names .. forgot)
20130809 12:58:52 <rmk0> sgothel: this is per pixel forward rendering
20130809 12:59:01 <rmk0> implementing it so that i understand enough to implement deferred rendering
20130809 12:59:12 <rmk0> sgothel: what was the question i was supposed to be answering?
20130809 12:59:14 <rmk0> can't see it
20130809 12:59:15 <gouessej> rmk0, do you use GL_LINES?
20130809 12:59:24 <rmk0> gouessej: yeah
20130809 12:59:40 <gouessej> :s
20130809 13:00:01 <gouessej> it can cause serious decrease of performance when used with shaders :s
20130809 13:00:13 <rmk0> can it?
20130809 13:00:20 <rmk0> it's only for "debug" geometry
20130809 13:00:30 <gouessej> I had this problems some years ago
20130809 13:00:37 <gouessej> "problem"
20130809 13:01:01 <gouessej> well, as you wish, I know it is difficult to obtain the same effect with geometry shaders
20130809 13:02:01 <sgothel> @Mark: TRIANGLE* something .. AFAIK is best
20130809 13:02:06 <rmk0> the actual renderer core only knows about triangles
20130809 13:02:17 <sgothel> *reading my mind* :)
20130809 13:02:25 <rmk0> the use of lines is just in the sandbox to draw the axes and grid
20130809 13:02:51 <gouessej> some drivers badly support mixing GL_LINES and/or OpenGL picking with shaders
20130809 13:03:10 <rmk0> won't be touching picking in OpenGL
20130809 13:03:25 <gouessej> that's why I only use software picking or hardware accelerated picking but not with the fixed pipeline
20130809 13:04:11 <gouessej> no GL_SELECT
20130809 13:04:13 <gouessej> etc...
20130809 13:05:02 <gouessej> I don't know how to draw lines with a constant thickness without geometry shaders
20130809 13:05:04 <gouessej> :s
20130809 13:06:57 <sgothel> a texture!
20130809 13:07:55 <gouessej> Why would it solve my problem?
20130809 13:08:10 <sgothel> oh .. confused lines w/ points :)
20130809 13:08:40 <sgothel> fragment shader knows about the position in framebuffer/window ..
20130809 13:08:42 <gouessej> How can I obtain the same effect than glLineWidth ?
20130809 13:09:00 <sgothel> so you can use that for const. resol. depending thickness
20130809 13:09:13 <gouessej> in gpu gems 3, they say such an effect can't be achieved with a fragment shader
20130809 13:09:40 <sgothel> I used it in one shader .. one sec
20130809 13:09:49 <gouessej> Cool :)
20130809 13:10:20 <sgothel> ruler.fp
20130809 13:10:47 <sgothel> uses gl_FragCoord.xy and the monitor dpi value to render a grid w/ correct distance
20130809 13:11:30 <sgothel> so you could even render resol. independent thickness ..
20130809 13:11:54 <sgothel> sure, this shader does it for the grid distance, but you can use it for the line thickness itself (vice versa)
20130809 13:13:03 <sgothel> .. back to multithreaded media player now :)
20130809 13:13:04 <gouessej> ok I have to improve my skills in shader programming
20130809 13:13:30 <gouessej> how do you get the dpi value?
20130809 13:13:48 <sgothel> NEWT MonitorDevice .. :) .. see the whole demo code
20130809 13:14:03 <gouessej> ok thanks
20130809 13:14:49 <sgothel> jogl/src/test/com/jogamp/opengl/test/junit/jogl/glsl/TestRulerNEWT01.java
20130809 13:16:13 <gouessej> glLineWidth(x) is broken or unsupported on lots of OpenGL-ES devices
20130809 13:16:22 <gouessej> that's why I ask this question
20130809 13:16:34 <sgothel> makes sense, since this is a shader task by itself
20130809 13:17:09 <gouessej> it is even broken in some OpenGL drivers :( Sometimes you can only pass 1 :s
20130809 13:19:44 <sgothel> so to do the lines w/ such ruler, you need to either pass a 'thick' rectangle (as triangles) .. or make a whole screen sweep
20130809 13:19:50 <sgothel> then you need to calc the distance ofc ..
20130809 13:20:04 <sgothel> ruler.fp simple does the screen-sweep
20130809 13:20:33 <sgothel> the distance might be done via texture coordinates (we are getting complicated now .. )
20130809 13:20:53 <sgothel> i.e. let them to be scaled by the vertex shader .. hmm
20130809 13:21:05 <sgothel> or was it the other way around (brainstorming mode)
20130809 13:21:27 <sgothel> right, the tex-coords are sort of static .. and you can test whether your current fragment is within the line
20130809 13:21:38 <gouessej> ok
20130809 13:22:13 <sgothel> a more elaborated version is in graph w/ curves :)
20130809 13:23:02 <gouessej> rulerPixFreqV contains the dpi, doesn't it?
20130809 13:23:05 <sgothel> if you are done .. maybe you like to add this to some glsl utils in JOGL :)
20130809 13:23:41 <gouessej> I'm not the best guy for this task, you know I'm not very comfortable with shaders
20130809 13:23:59 <sgothel> res / dim * 10f
20130809 13:24:08 <sgothel> pixel/mm *10f
20130809 13:24:22 <sgothel> line 122 in TestRulerNEWT01
20130809 13:24:44 <sgothel> so .. dots per cm :)
20130809 13:25:11 <sgothel> ruler.fp renders a grid .. centimeter spacing
20130809 13:26:36 <sgothel> since I never really learned how to do things w/ FFP - I like shaders a lot, almost like requires you to know nothing :)
20130809 13:26:58 <sgothel> .. but a little math
20130809 13:28:08 <gouessej> I'll need a little math too to work around the limitations of glLineWidth with the FFP
20130809 13:29:20 <gouessej> ok I have to leave. I'll try to add some information into the bug report 814. Bye.
20130809 13:29:34 * gouessej (546051e4@anon) Quit (Quit: Page closed)
20130809 13:29:36 <sgothel> you can simply setup the rectangle vertices (as triangles) and also put this to texture coords using a texture-unit.
20130809 13:29:39 <sgothel> .. too late
20130809 13:40:44 <chakie> ah, i wanted to pong gouessej too
20130809 13:44:02 <sgothel> whats your issue ?
20130809 13:44:13 <sgothel> your discussion about Text ?
20130809 13:45:00 <xranby> chakie: hi jan!
20130809 13:45:13 <chakie> hi :)
20130809 13:45:31 <chakie> sgothel: yeah, just wanted to say hi
20130809 13:46:03 <xranby> https://github.com/sgothel/jogl/pull/47 <-- Cleanup and OpenGL 3 support for TextRenderer
20130809 13:46:03 <sgothel> hi Jan (incr. my mapping space for those nicks)
20130809 13:46:16 <sgothel> yes Xerxes ? :)
20130809 13:46:24 <sgothel> I answered this yesterday ..
20130809 13:46:29 <xranby> ok awesome
20130809 13:46:36 <sgothel> i.e. maintainer n/a
20130809 13:47:03 <sgothel> 2012 siggraph it was being requested .. to merge - I told him to clean it up, they said OK, but nothing happened
20130809 13:47:22 <chakie> i guess you're all more or less developers working on jogl here?
20130809 13:47:24 <sgothel> me/us: not really interested in awt text renderer using fixed resolution etc
20130809 13:47:39 <sgothel> you could say so :)
20130809 13:48:01 <chakie> i won't disturb you then
20130809 13:48:03 <sgothel> and what are you doing Jan ? do I know you already ? sorry ..
20130809 13:48:12 <xranby> chakie: we try to fit all world issues into bugreports :)
20130809 13:48:13 <sgothel> no no ..
20130809 13:48:28 <sgothel> do inspire us w/ your ideas!
20130809 13:48:53 <chakie> sgothel: i don't think you know me. i know your name from the fact that you work on jogl
20130809 13:48:53 <sgothel> .. and solve it later w/ next release 2.1 :)
20130809 13:49:42 <chakie> i try to do terrain rendering for mapping solutions
20130809 13:49:44 <sgothel> if you have ideas .. or like to change things, everybody is welcome
20130809 13:49:49 <sgothel> nice
20130809 13:50:13 <sgothel> so what do you use as a source ?
20130809 13:50:25 <chakie> i do some work for a couple of friends that have old mapping stuff done with java
20130809 13:50:27 <sgothel> 'life' data .. we discussed this earlies .. ?
20130809 13:50:54 <sgothel> i.e. OSM or mapserver.org .. or WW
20130809 13:51:01 <chakie> all custom
20130809 13:51:05 <sgothel> oh
20130809 13:51:13 <sgothel> so .. closed stuff ?
20130809 13:51:19 <chakie> they have all kinds of mapping data that they use in their current 2d engine
20130809 13:51:25 <chakie> yes, totally
20130809 13:51:42 <chakie> but the 2d engine is a normal awt thing
20130809 13:51:54 <sgothel> oh well, but ok - if you guys like to open source modules .. maybe some day we like to do our own JogmAp :)
20130809 13:51:58 <chakie> won't scale too nicely with these new 4k displays
20130809 13:52:02 <sgothel> hehe
20130809 13:52:20 <chakie> so opengl is being investigated
20130809 13:52:21 <xranby> chakie: professional "geo"-data rendering is also a topic worked on by Eclesia
20130809 13:52:22 <sgothel> sure .. we prefer something ES2 and curvy
20130809 13:52:46 <sgothel> Eclesia ? .. the PD dude ?
20130809 13:52:51 <xranby> yes
20130809 13:53:01 <sgothel> ah .. memory still works :)
20130809 13:53:06 <chakie> es2 and desktop opengl are my targets
20130809 13:53:29 <xranby> also a geotools maintainer
20130809 13:53:46 <chakie> right now i'm mostly testing things with dummy data to see how well it would run
20130809 13:53:57 <chakie> as well as build up some foundations
20130809 13:54:14 <sgothel> all the mapping stuff .. definitely something fun we could toy w/ :)
20130809 13:54:22 <chakie> yes
20130809 13:54:39 <chakie> i don't think opening parts will be totally impossible
20130809 13:54:54 <sgothel> @Jan: since it's closed source .. well, nothing we can do about it .. but you may join some open thingy in parallel (or convince your company to change their minds :)
20130809 13:54:55 <chakie> as open source gives a lot of good will in certain commercial areas too
20130809 13:54:56 <xranby> and like you probably noticed, gouessej is working on some map data importers for his revolution game
20130809 13:55:33 <sgothel> and so on .. and so forth .. ok, now really cont. on the media player :)
20130809 13:58:11 <chakie> i mainly wanted to add some altitude/long/lat and tile loading debug data to my view, thus my adventures into rendering text
20130809 13:58:42 <chakie> i'll stick with logging it for now, can't get it to work :)
20130809 13:58:59 <xranby> chakie: did the new graph based ES2 text rendering work on your devices?
20130809 13:59:23 <chakie> yes
20130809 13:59:30 <chakie> or only tested on osx for now
20130809 13:59:44 <chakie> or, well
20130809 13:59:51 <chakie> the web demo works
20130809 13:59:57 <chakie> in my code, not so much
20130809 14:00:26 <sgothel> .. API will be changed a bit .. have some bug entries about it ..
20130809 14:00:42 <sgothel> https://jogamp.org/bugzilla/showdependencytree.cgi?id=803&hide_resolved=0
20130809 14:02:56 <chakie> the API isn't too complex, so small changes don't matter to me at all
20130809 14:03:39 <chakie> i use the class somehow wrong and have yet to see any text in my own code
20130809 14:03:47 <xranby> chakie: please brainstorm so that we may capture your desired use cases for the graph ui package
20130809 14:04:02 <xranby> this is the extended JogAmp developers meeting JDM graph ui session :)
20130809 14:04:13 <sgothel> :)
20130809 14:04:30 <chakie> my use case (for now) is to simply have normal 2d HUD like text
20130809 14:05:14 <chakie> later at some point i'll need to draw map labels too, which could mean various transformations
20130809 14:05:21 <sgothel> probably simply a PMV miss, or some GLSL / other state mess
20130809 14:08:08 <chakie> possibly
20130809 14:08:35 <chakie> my own stuff renders nicely though
20130809 14:09:09 <chakie> now that i think about it, i need to check the init() method of TextRenderer
20130809 14:09:29 <chakie> in case there's some extra state that gets set and it assumes "sticks"
20130809 14:09:34 <chakie> like GL_BLEND
20130809 14:10:22 <sgothel> maybe .. (not blend .. but others maybe, if so .. please tell us)
20130809 14:10:28 <sgothel> or PMV is 'off the map' :)
20130809 14:11:21 <sgothel> test: add a simple things w/ the text-PMV .. .. to show something simple
20130809 14:12:42 <chakie> the code i use does reset the matrices before rendering
20130809 14:13:16 <chakie> and set up an orto projection
20130809 14:24:36 <chakie> anyway, i'll make do without text for now and check again at some later point
20130809 14:33:01 <xranby> sgothel: do i have access to any of the jogamp debian builders?
20130809 14:33:23 <sgothel> Sylvestre .. does this AFAIK
20130809 14:33:45 <xranby> i need access to a proper debian dev machine where i may prepare and test a new "avian" JVM debian package
20130809 14:33:52 <xranby> need to practise my debian fu
20130809 14:34:10 <sgothel> "a proper debian dev machine" ?
20130809 14:34:20 <sgothel> define ?
20130809 14:34:27 <sgothel> one of ours ?
20130809 14:34:30 <xranby> building avian debian package on a ubuntu machine runs into odd issues
20130809 14:34:45 <xranby> building ubuntu packages on ubuntu work ok
20130809 14:35:16 <xranby> yes i am asking if i have access to any of the nodes or if i should look elsewhere
20130809 14:35:44 <sgothel> I can set up an account .. and means to access it directly .. hmm
20130809 14:35:55 <sgothel> ssh would work for you I guess ?
20130809 14:35:56 <xranby> or / if i should setup my own
20130809 14:36:11 <xranby> yes
20130809 14:36:23 <sgothel> jogamp02 is debian7
20130809 14:41:38 <xranby> sgothel: thank you.. ok i will try access from the jogamp.org server when i have access to my keychain later
20130809 15:08:12 <sgothel> xranby@ac100-armhf <- should work ..
20130809 15:08:47 <xranby> yes.. thats where my key is
20130809 15:08:52 <sgothel> ssh -v -p 1023 xranby@jau.dyndns.org
20130809 15:08:54 <xranby> i have the chromebook in by backpack
20130809 15:09:23 <sgothel> tested from a remote machine .. working
20130809 15:09:26 <xranby> thank you for setting this up!
20130809 15:11:27 <xranby> For Ji Gong: there is now a 0.7 avian release http://oss.readytalk.com/avian/avian-0.7.tar.bz2
20130809 15:11:42 <xranby> http://oss.readytalk.com/avian/status.html
20130809 15:12:07 <sgothel> you need anything special to install ?
20130809 15:12:21 <sgothel> you don't need X11 right ?
20130809 15:12:32 <xranby> right
20130809 15:12:35 <sgothel> to use .. I mean .. good!
20130809 15:13:56 <xranby> basically i need to be able to checkout the debian avian package install the apt-get build-dep jvm-7-avian-jre
20130809 15:14:11 <xranby> apt-get source jvm-7-avian-jre
20130809 15:14:31 <xranby> fix the source to use the above 0.7 tar.bz2
20130809 15:14:38 <sgothel> Unable to find a source package for jvm-7-avian-jre
20130809 15:15:02 <sgothel> apt-get build-dep jvm-7-avian-jre
20130809 15:15:04 <xranby> hmm the source package is avian it seems
20130809 15:15:05 <xranby> http://packages.debian.org/unstable/main/jvm-7-avian-jre
20130809 15:16:54 <sgothel> n/a in wheezy, debian7
20130809 15:17:32 <xranby> ok.. hmm well maybe i should simply setup a chroot then on one of my machines
20130809 15:17:43 <xranby> compared to messing with the builder
20130809 15:18:22 <sgothel> yup .. however, was good to set the direct ssh channel to build nodes up ..
20130809 17:05:34 * gouessej (52794f26@anon) has joined #jogamp
20130809 17:05:40 <gouessej> hi
20130809 17:05:43 <gouessej> again
20130809 17:05:57 <gouessej> I've just updated the bug report 814
20130809 17:21:42 * doby (~doby@anon) has joined #jogamp
20130809 17:28:03 <doby> I'm attempting to use GlueGen for the first time and am running in to an odd error. I am on RedHat Enterprise Linux 6.4 and using Sun/Oracle Java 1.7.0_25. The .h file I am running against has a #include <stdio.h> so I added -I/usr/include which results in this error: Caused by: java.io.IOException: Expected token '40' but got '__need_FILE' at file /usr/include/stdio.h, line 23
20130809 17:28:35 <doby> if I add --debug to the arg list I get a NullPointerException
20130809 17:28:52 <doby> any ideas what to try?
20130809 17:39:31 <gouessej> hi doby
20130809 17:39:37 <doby> hi
20130809 17:40:29 <gouessej> How have you tried to build GlueGen?
20130809 17:41:45 <gouessej> I'm a bit surprised, stdio is very basic
20130809 17:42:04 <gouessej> If it isn't found, you can almost compile nothing in your environment :s
20130809 17:43:07 <doby> its definitely there
20130809 17:43:17 <doby> and, yes, that was my thought exactly
20130809 17:43:28 <doby> I did build GlueGen from source
20130809 17:43:36 <doby> I got it from GitHub...
20130809 17:43:55 <doby> Here: https://github.com/sgothel/gluegen
20130809 17:44:40 <gouessej> I think that the problem doesn't come from GlueGen itself in your case
20130809 17:44:42 <doby> I am attempting to get it working using the command line with the following command /usr/java/jdk1.7.0_25/bin/java -jar gluegen.jar -I/usr/include ~/cci/cci.h
20130809 17:45:07 <gouessej> Why do you do that???
20130809 17:45:16 <doby> its command line or ant, right?
20130809 17:45:52 <gouessej> You should use the Ant script in the make/ directory
20130809 17:46:52 <doby> Maybe I'm missing something, I followed the instructions here:
20130809 17:46:56 <doby> http://jogamp.org/gluegen/doc/manual/
20130809 17:47:36 <doby> After building, I am attempting to use it as a executable jar
20130809 17:47:53 <doby> "GlueGen can be run either as an executable jar file (java -jar gluegen.jar; note that antlr.jar must be in the same directory as gluegen.jar in order for this invocation to work) or from within Ant "
20130809 17:47:59 <doby> ...
20130809 17:48:09 <doby> which is followed by a description of the command line switches
20130809 17:48:24 <gouessej> some parts seem obsolete :s
20130809 17:49:10 <doby> ok, what is the recommended method?
20130809 17:49:37 <gouessej> ok just wait for a minute, I read this document anew
20130809 17:50:03 <doby> thank you... can't say I always keep my docs up to date either ;)
20130809 17:50:17 <gouessej> You succeed in building GlueGen but you fail in using it with your own header, don't you?
20130809 17:50:31 <doby> yes
20130809 17:51:39 * DeSpeckle1 (~uname@anon) has joined #jogamp
20130809 17:51:53 <gouessej> You need to adapt this line: java -cp gluegen.jar:antlr.jar com.sun.gluegen.GlueGen \ -I. -Ecom.sun.gluegen.JavaEmitter -Cfunction.cfg function.h
20130809 17:52:03 <gouessej> some packages have been renamed
20130809 17:52:05 <gouessej> :s
20130809 17:52:14 <doby> ok thank you i will give that a shot
20130809 17:52:44 <gouessej> If something doesn't work, please post the whole stack trace or error message
20130809 17:52:57 <doby> ok
20130809 18:05:17 * kermyt (~kermyt@anon) has joined #jogamp
20130809 18:16:04 * gouessej (52794f26@anon) has left #jogamp
20130809 18:30:39 * DeSpeckle1 (~uname@anon) has left #jogamp
20130809 21:08:28 <sgothel> back .. @Julien: pretty verbose .. very good (814) thx
20130809 21:08:52 <sgothel> maybe you can describe all those things like 'axis events, sliding events, ... etc'
20130809 21:09:11 <sgothel> or better: add a link to documentation ..
20130809 21:09:27 <sgothel> https://jogamp.org/bugzilla/show_bug.cgi?id=814#c1
20130809 21:09:43 <sgothel> No Idea: - axis events (float) [analog] - sliding axis events (float, like 2 axis: abscissa, ordinate) [analog] - hat switch or POV (point-of-view) hat events (9 directions) [analog] - vector3 events (float, like 3 axis: abscissa, ordinate, applicate) [analog]
20130809 22:28:25 * xranby (~xranby@anon) Quit (Read error: Operation timed out)
20130809 22:45:03 * xranby (~xranby@anon) has joined #jogamp
20130809 23:21:33 * xranby (~xranby@anon) Quit (Ping timeout: 276 seconds)
20130809 23:27:19 * xranby (~xranby@anon) has joined #jogamp
20130810 01:09:57 * kermyt (~kermyt@anon) Quit (Ping timeout: 248 seconds)
20130810 01:13:05 * kermyt (~kermyt@anon) has joined #jogamp
20130810 05:06:07 -jogamp- Continue @ http://jogamp.org/log/irc/jogamp_20130810050607.html