Hello all,
I have not had any contact with Kim Bale, who had started and was
heading the osgOcean project, since at least a few months, despite a few
e-mails from me and questions from others in that time. His last
activity on osgOcean was in May of this year (see
http://code.google.com/p/osgocean/updates/list). I am committer on the
project and have made modifications with Kim's consent in the past, but
now I find I might have to "take the reins" in Kim's stead if I don't
hear from him in the near future.
First of all, would anyone have heard from Kim lately? I have no other
means to communicate with him than the e-mail address I was using all
this time, and I'm wondering if something may have happened on a more
personal level (other than just changing jobs/interests which surely may
happen). So if anyone has a way to reach Kim, just to let us know
everything's all right, I'd appreciate it.
Secondly, I have recently reintegrated some changes I'd made locally
over a few months into osgOcean trunk, and was considering making a
release soon in order to keep the project moving forward. If anyone is
using osgOcean 1.0.1 and would like to help me out a bit, could you just
download the SVN trunk and test it out in your apps? The URL is:
http://osgocean.googlecode.com/svn/trunk/
The changes I've made will let apps decide whether they use the inline
shader source or customized shader files they would distribute with
their apps, and also let apps decide when and how fast the ocean surface
updates itself. More details:
Shaders: Previously, osgOcean hard-coded a value for each class that
used shaders, called USE_LOCAL_SHADERS, and this value decided if
osgOcean used the inline shader source or loaded the shader from file.
If an app wanted to customize one of the ocean shaders, for example the
ocean surface shader, and it had compiled osgOcean with
USE_LOCAL_SHADERS=1 in the FFTOceanSurface.cpp file, then it either had
to traverse to find the ocean surface node and override the shader
there, or recompile osgOcean with USE_LOCAL_SHADERS=0. This was all too
complicated and made integration difficult.
Now, osgOcean combines both ways. It will first try to load the shader
from file (and all shader files have been prefixed with "osgOcean_" to
prevent it from loading the wrong one if the user had a shader called
"heightfield" for example...) and if the shader is not found in the OSG
data paths, then it will use the inline shader source. This is more
flexible, so apps can decide not to distribute any shader files if they
don't want to, and if they want to customize one shader they just have
to distribute that one and make sure the OSG data path is set to find it.
Eventually I'd like to let the user specify their shader filename or
source directly to have maximum flexibility, but I'll have to think
about how to design/expose this in a good way.
Ocean surface updating: We run osgOcean with viewports spanning screens
on multiple computers, and we have dynamic objects that follow the ocean
surface. Both these factors meant we needed to have a way to control the
ocean surface update from some external source. FFTOceanSurface now has
a public OceanUpdateCallback that users can override. Using this, we
broadcast a simulation time from one master to all graphics slaves, and
both the dynamics and graphics (including the ocean surface) use this
value to update themselves. This makes the ocean surface synchronized
everywhere.
There were also a few small bug fixes made, which will be in the release
notes when I decide to make a release.
So, anyone interested in doing a bit of testing? :-)
Finally, if I still can't get any sign from Kim in the next few weeks,
I'll go over the active discussions in the issue tracker on the
googlecode site and see if I can respond to or close any of them. I
wasn't checking that at all in the past, thinking Kim was much better
placed to respond (since most changes I've made were integration/class
design changes and not the actual rendering/shaders) but now I'll need
to get more involved. And another thing is the submission of river
rendering that Quintijn Hendrickx made a while back, I'll be looking to
integrate that in the near future.
Thanks,
J-S
--
______________________________________________________
Jean-Sebastien Guay [email protected]
http://www.cm-labs.com/
http://whitestar02.webhop.org/
_______________________________________________
osg-users mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
