-
Is there going to be a 1.6.4 release? I'm curious how this is propagated to the trunk (e.g. a mass patch from 1.6.3 at some point, or reapply individual patches as they are "tested"?)
2009-08-10 15:58:38 UTC in OGRE (O-O Graphics Rendering Engine)
-
The GLX/OSX/WIN32/... directories in OgreMain/include assume that this 'include' is on the header search path post-installation. Adding a simple '../' to their #includes will remove this assumption and allow a basic Ogre installation to function out-of-box if installed to a standard system path (e.g. /usr/local)
Plugins and other optional components may still have search path assumptions...
2009-07-09 14:00:19 UTC in OGRE (O-O Graphics Rendering Engine)
-
Hmm, as a minor optimization, is it necessary to calculate all three traces? I don't have a strong proof, but I think choosing the case based on the biggest trace element will also give the biggest trace... if the second biggest is negative you'd want to subtract it, and if it's positive then you'd want to stick with using the biggest for the positive value, so either way the biggest total...
2009-06-20 16:03:58 UTC in mjbWorld
-
> 1) danger of division by zero or square root of a negative number.
> 2) accuracy of dividing by (and square rooting) a very small number.
I think because we pick the biggest diagonal element in the 'else', and subtract the others, we can never get a negative trace in the 'else' cases. Assuming the following proof holds, then because of the +1 in the sqrt calls, the divisors won't...
2009-06-19 01:24:20 UTC in mjbWorld
-
Hi Martin,
I was looking at the matrixToQuaternion page, and I think the "Better code" and "C++ code" sections may have a problem with numerical stability as Tr + 1 may approach 0 in some cases. The test for greater-than-epsilon helps, but a better solution may be found in the implementation from the Wild Magic package by Dave Eberly...
2009-06-17 18:09:51 UTC in mjbWorld
-
What you have is the current directory of gravity from the accelerometer. Let's assume "home" position is if gravity is straight down along the z vector. (Really this is arbitrary, you could record the actual home position to account for however you have it mounted).
So the axis of rotation between these two vectors is the cross product. This gives you a vector which is normal to...
2009-06-17 17:58:56 UTC in mjbWorld
-
Might want to ask ark42 why he introduced FREEIMAGE_COLORORDER in the first place... seems equivalent to FREEIMAGE_BIGENDIAN... just wanted a more descriptive name for what was being tested?.
2009-04-20 19:21:50 UTC in FreeImage
-
Ah, thanks for tracking that down.
I don't think you should be testing __APPLE__ at all as an indication of byte order. Intel macs are little endian, PowerPC macs are big endian.
So for instance, I tested that this gives an accurate report:
#include
#include
int main(int argc, char** argv) {
#if defined(FREEIMAGE_BIGENDIAN)
printf("big-end detected\n");
#else...
2009-04-20 17:35:02 UTC in FreeImage
-
Ogre3D is seeing an issue with FreeImage swapping the red and blue channels.
This problem was apparently introduced in 3.10 (3.9.3 worked), and persists through 3.12.
http://www.ogre3d.org/forums/viewtopic.php?f=2&t=49326.
2009-04-16 18:38:12 UTC in FreeImage
-
I'm curious if it's possible to direct-command the NXT via wired USB instead of bluetooth. The page for this software at http://mac.softpedia.com/get/Math-Scientific/LegOSX-Direct.shtml seems to indicate this is available, but I don't see code for it in the source. I can't figure out how to access the device myself either... doesn't seem to show up as a /dev/tty device like bluetooth does. :-/.
2008-12-28 06:49:05 UTC in LegOSX-Direct
