We checked out a version of Chromium from the cvs repository a few days ago. We're trying to get vtk4.0 running with Chromium. It built without a hitch. But, our crserver's keep crashing. It seems that the dispatch table is being corrupted somehow. The server is trying to execute an ActiveTextureARB (something like that, the code is at work) call even though we don't have any of these in vtk. We suspect that someone is clobbering the stream.
So, has anyone run with vtk4.0 or even 3.x?
Bill
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I ran into a similar problem. I believe the state differencer is generating the calls to ActiveTextureARB on the servers. Do you have an environment variable named CR_DEBUG set? If so, you can turn the ActiveTextureARB error into a warning by changing the crError call in state_tracker/state_error.c into a crWarning call.
I have been running ParaView as a single process and using Chromium's tilesort for a while. There are a few problems (turning on the colorbar and legend can crash recent Chromiums) but otherwise quite nice.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I'm using a CVS snapshot of the trunk from 3/14. I've also (in the last half hour) got a checkout from yesterday off of the multithread-1-branch working by commenting out the XOpenDisplay() call in tilesortspu_context.c. I even have it working in parallel, but I've only tested it with 6 spheres being drawn from different processes. :-)
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
We tried the 3/14 and multithread branch. Still no luck. We are building on Windows (a mix of NT and Windows2k). The multithread branch would not build.
If we comment out the ActiveTextureARB code in the packer, we get a little further. We still get a crash when our 2 tiles are horizontal and we're rendering polydata data with color scalars per vertex. (VTK/Graphics/Testing/Tcl/combStreamers.tcl)
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
We checked out a version of Chromium from the cvs repository a few days ago. We're trying to get vtk4.0 running with Chromium. It built without a hitch. But, our crserver's keep crashing. It seems that the dispatch table is being corrupted somehow. The server is trying to execute an ActiveTextureARB (something like that, the code is at work) call even though we don't have any of these in vtk. We suspect that someone is clobbering the stream.
So, has anyone run with vtk4.0 or even 3.x?
Bill
I ran into a similar problem. I believe the state differencer is generating the calls to ActiveTextureARB on the servers. Do you have an environment variable named CR_DEBUG set? If so, you can turn the ActiveTextureARB error into a warning by changing the crError call in state_tracker/state_error.c into a crWarning call.
I have been running ParaView as a single process and using Chromium's tilesort for a while. There are a few problems (turning on the colorbar and legend can crash recent Chromiums) but otherwise quite nice.
I'm not getting an error, the crserver is crashing. What version of Chromium are you using?
I'm using a CVS snapshot of the trunk from 3/14. I've also (in the last half hour) got a checkout from yesterday off of the multithread-1-branch working by commenting out the XOpenDisplay() call in tilesortspu_context.c. I even have it working in parallel, but I've only tested it with 6 spheres being drawn from different processes. :-)
We tried the 3/14 and multithread branch. Still no luck. We are building on Windows (a mix of NT and Windows2k). The multithread branch would not build.
If we comment out the ActiveTextureARB code in the packer, we get a little further. We still get a crash when our 2 tiles are horizontal and we're rendering polydata data with color scalars per vertex. (VTK/Graphics/Testing/Tcl/combStreamers.tcl)