Re: [G3D] "Shader Compilation Failed" due to macros in UniversalSurface_*.vrt files being out of sync with UniversalSurface.cpp, fix included

[Michael Mara]

That's what I get for not checking my patches on all platforms daily! All set, your patch was exactly the right one, it's checked in to svn now.

-Michael

On Thu, Jan 31, 2013 at 4:58 AM, Mike Roberts <mikerob...@gmail.com> wrote:

Hi there,

Please allow me to introduce myself. My name is Mike Roberts. I'm a first-year PhD student in the graphics group at Stanford. I'd like to use G3D for a new project, and I'm in the process of getting my build set up on Mid 2009 Macbook Pro with an NVIDIA 9600M running Mac OSX 10.8.2. I realize that I am in somewhat uncharted waters here (64-bit new Mac OS, somewhat older GPU, etc). I am posting my experiences here in an attempt to give back to the open-source community, and to help improve G3D.

I had problems with the 8.0.1 binaries. In this case, the pre-compiled viewer would hang my display when I tried to exit the app, thereby forcing me to reboot. I was also having problems with the released 9.0 beta 4 binaries and sources. In this case, the viewer would crash in UserInput::endEvents() when exiting. This happened when I built from the 9.0 beta 4 sources, as well as when I used the pre-compiled viewer. These problems motivated me to get the latest sources from SVN. As far as I know, I am on r2812, which is the most up-to-date possible version of the code I can get.

Anyway, I built the r2812 code from scratch by doing:

>> ./buildg3d update

For what it's worth, I set my G3D9DATA environment variable to g3d-code/G3D9/build/data. While I'm on the subject, is it intended behavior that the enet.lib libraries and headers are not copied into g3d-code/G3D9/build like all the other headers and libraries? I needed to adjust my INCLUDE and LIBRARY environment variables especially for enet because buildg3d did not copy these libraries and headers into g3d-code/G3D9/build. No big deal either way, but was this intended behavior?

Anyway, back to the main bug. I compiled the starter sample app by doing:

>> icompile

I tried to run it (from the terminal only - double clicking on starterd in finder hangs), and I got an error message about undefined macros in the UniversalSurface_*.vrt files (screenshot attached).

As far as I can tell, this error happens because UniversalSurface_*.vrt files refer to macros that are not always set in UniversalSurface.cpp. I made a couple of very minor tweaks to UniversalSurface.cpp, which allow the starter app to run as expected. I'm not sure if my workaround is the correct solution, but at least it helps to understand the problem. Here is my workaround:

dn0a21025a:G3D9 mike$ svn diff

Index: GLG3D.lib/source/UniversalSurface.cpp

===================================================================

--- GLG3D.lib/source/UniversalSurface.cpp (revision 2813)

+++ GLG3D.lib/source/UniversalSurface.cpp (working copy)

@@ -129,6 +129,8 @@

         if (thisSurfaceNeedsAlphaTest) {

             args.setUniform(SYMBOL_lambertianMap, lambertian);

             args.setMacro(SYMBOL_HAS_ALPHA, 1);

+        } else {

+            args.setMacro(SYMBOL_HAS_ALPHA, 0);

         }

         if (thisSurfaceNeedsAlphaTest && isParallaxMapped) {

             args.setUniform(SYMBOL_normalBumpMap, bumpMap->normalBumpMap()->texture());

@@ -764,6 +766,7 @@

 G3D_DECLARE_SYMBOL(g3d_PackedTangent);

 G3D_DECLARE_SYMBOL(g3d_BoneIndices);

 G3D_DECLARE_SYMBOL(g3d_BoneWeights);

+G3D_DECLARE_SYMBOL(USE_BONES);

 void UniversalSurface::GPUGeom::setShaderArgs(Args& args) const {

     debugAssert(normal.valid());

     debugAssert(index.valid());

@@ -789,6 +792,8 @@

         args.setStream(SYMBOL_g3d_BoneWeights, boneWeights);

     }

 

+    args.setMacro(SYMBOL_USE_BONES, 0);

+

     args.setIndexStream(index);

     args.geometryInput = primitive;

 }

dn0a21025a:G3D9 mike$ 

Cheers,

Mike

--
You received this message because you are subscribed to the Google Groups "G3D Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to g3d-users+...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Morgan McGuire]

I also committed (but didn't yet test) the OS X enet.lib patch.

-m

Prof. Morgan McGuire
Computer Science Department
Williams College
http://cs.williams.edu/~morgan