Cooki3d has been released !

OpenGL Development JOGL Development
#1

Hi all Jogl users,

Cooki3d is a cross platform game engine based on OpenSceneGraph(1.2) and SDL under dual licence…
Makes it possible to run a powerful 3d multimedia application (support wav, mp3, xml…) on a web navigator or using a specific GUI (SDL, Producer, Qt4, Java/jogl).
Cooki3d support fireFox and iExplorer under Linux and windowsXP/2000/98 via applets, you can test the online examples at now!.
Cooki3d has also its java framework to create, build and deploy easily cooki3d applications on the web or localy.

http://www.cooki3d.org

Thank you in advance for your advices and returns…

Best Regards,
Bertrand Greslier.

#2

I’ve had some difficulty running the web-based demos which is probably due to my non-standard development JRE. In general, however, I don’t understand the motivation for combining all of these libraries; it seems to me that there is significant overlap between OSG, SDL and JOGL. My personal opinion would be to focus on writing pure Java libraries on top of JOGL, for example a scene graph which provides the features you want; with this approach your resulting code will run on more platforms than you currently support.

#3

Hi Ken,
thank you for your opinion, (excuse me in advance my english)

it seems to me nobody like or is interessted by this solution in this forum :’(. Perhaps because it is not 100% Java?

I’ve had some difficulty running the web-based demos which is probably due to my non-standard development JRE.

It is very simple to install a standard JRE and use it, perhaps you should have one to also compare your own code behaviors…
it could be a bug in cooki3d with your configuration, I hope somebody has succed to test demos.

I don’t understand the motivation for combining all of these libraries

these libraries OpenSceneGraph,SDL are “standard” in open source and C/C++ devpts since many years and have a big community to work on.
It support already all fomats and all functionnality in 3d devpts. I think I use the best in this language.
And I think also a 100% java solution is not yet the best with performances today… it is certainly not the most important because java improve…
But it is especially that the resources available in 3d devpts are yet more numerous in C/C++ for somebody (I am alone, it is the work of only one person)
who don’t want rewrite all exist C/C++ ressources in java all his life.

with this approach your resulting code will run on more platforms than you currently support
It is a first version, it support Linux and Windows, all the code is cross platform, if I compile one other plaforms, it will work as the java jvm works.

Best Regards,
Bertrand Greslier.

#4

Hi,

Can’t get your applet working :

java.lang.NullPointerException
	at sun.applet.AppletPanel.findAppletJDKLevel(Unknown Source)
	at sun.applet.AppletPanel.createApplet(Unknown Source)
	at sun.plugin.AppletViewer.createApplet(Unknown Source)
	at sun.applet.AppletPanel.runLoader(Unknown Source)
	at sun.applet.AppletPanel.run(Unknown Source)
	at java.lang.Thread.run(Unknown Source)

I’m using jre 1.6.0-rc-b103.

And for the full java ontop of jogl I quite agree with ken, but it’s a matter of taste I think… I worked a lot with C/C++ and I’m quite happy to work nearly only with java now…

#5

I use the (old) tag
I haven’t test the jdk 1.6, it is a release version?
I use 1.4 compilance and a 1.5.0_09-b01 jvm to compile.
I don’t know the solution to resolv this problem, perhaps I must change the tag?
Anybody can light me?

And for the full java ontop of jogl I quite agree with ken, but it’s a matter of taste I think… I worked a lot with C/C++ and I’m quite happy to work nearly only with java now…
I haven’t say I love or I prefer C++ and I haven’t say C++ is better than java, I have give my arguments of my choice for this specific solution in this specific domain , it is all.

#6

Well it’s a release candidate version… and it do improve a lot of things.
I can’t ligth you on applet stuff but some gurus on this forum are experts in this field, they may help you on that.

I understand and I was given my opinion to, not a big deal in any case.

#7

After showing a “loading…” for a couple of minutes it crashed my browser (ff 2.0) it also added 30MB worth of files in my home dir after pressing a link saying “…[476K]”. Not quite sure I like that. Think the initiative is quite ok, but not working well at the moment.
WinXP, Java 1.5

#8

BTW, bertrandgreslier, your Personal Message Inbox is full.

I’ll try to look into the NullPointerException in the JRE later today. Maybe one of your jars is corrupted or something went wrong during the download process. Have you tried running this from a “clean” machine where you haven’t run it before?

#9

works for me, I didn’t tried all stuff but shadow sample is very interresting, nice work.

but I don’t understand why there is so much data loaded ?!

#10

the example is [476Ko],
the cooki3d libs (OSG…) are installed the first time you test an example. It is bigger…

if you have a crash, please send me directly this tree files
the hs_err_pid$(YOUR_PID).log file on your desktop
=> C:\Windows\temp\cookiLogs$(YOUR_PID).txt
=> C:\Documents and Settings/bertrand/Application Data/Sun/Java/Deployment/log/$(YOUR_JVM).trace
and tell me your graphic card. thanks

#11

My personal message box (java.dev) is full with only one message ! I dont understand !

I test on two machine windows and three linux but I clean manualy when I want begin from zero…

#12

the cooki3d libs (OSG…) are installed the first time you test an example. It is a big set of dll
but next version will be more light.

#13

I’ve looked into why your applets aren’t working properly apparently with JDK 6 and later and as far as I can tell it’s due to a problem with your web server. It looks to me like it’s occasionally dropping HTTP connections to download some of the associated jars. Either that or the jar files are somehow corrupt or in a non-standard format and not openable by the normal code path of the Java Plug-In once they’re cached on the local disk. I’ve placed some logs and the associated instrumented source code here so you can see what is going on and compare it to the output from the Gears applet on the jogl-demos web site. You should test your applets with JDK 6 and enable logging on your server to see why it’s apparently closing connections to the Java client on the other side. Good luck.

#14

Thank you for this analyse Ken, you have find the bug with the jar passed into the of the applet ? or you talk some additionnal download I make just after?

(If it is the first option archive="…,…,…")
It not my web server, it is the free.fr provider server and I don’t think they will change their server connection policy for the jdk6.
I know it it is limited as two connection at same time for one user and the timeout is very very small,
when I use http connection, I must always redo an url.connect() before do an operation.
And I verify for each execution than my donwload is the same on client and on server…
I am sorry but perhaps it is new bug introducting by the jdk6.

I have other most important bugs (some windows/nvidia computers) for the moment… but I will use your zip as soon I have the time
thank you.

#15

I have “re-think” of that and finally it could be a possibility for the user to have the option to choice a java implementation.
Also it could be a good idea for java games and OSG than it exists a 100% java impletation of these scene graph.
All the api is good and could be transform in java code, the languages are very near. But do t do this with hands is very hard, perhaps tools exist.
I think at perhaps use the Doxygen parser…
What do think about this? do you know good tools to do this?

#16

This is an interesting problem area and a few people I know have given it some thought. From what I hear the fastest thing to do might be to use lcc to parse the program and then do some transformations on the IR. You can consider that some “well-behaved” C/C++ programs might be easily transformable into Java classes and methods, where others which rely on casting real objects to arbitrary data types might be convertible using some slower but guaranteed correct methodology such as representing the storage for objects using byte arrays.