- sfx_openal.cpp uses isnormal(), not supported on VC++.I replaced by detecting with NaN ("factor != factor"), I don't see why denormalized floating points number could be a problem?
- shadow_importance.cpp: roundf() isn't supported too on VS2012 (it's C++11 standard though, according to cplusplus.com
). so I did glVertex2s((short)((xpos * 32767.f)+0.5f), (short)((ypos * 32767.f)+0.5f) );
- it crashed on Intel GPU on Windows. This is not the same one I used up to now, so I had to update the drivers again. In case we detect an Intel driver (and probably only on Windows), maybe we should point the user to this webpage: http://www.intel.com/p/en_US/support/detect/graphics
-> it's a Java tool that detects which driver you need. Considering how messy the Intel drivers web pages are, I think this is useful (at least it got me the good one, which I failed to find using the search tool).
- I'm at the point where shader "PASSFAR" failed to load ("GLSL: failed to retrieve uniform location information"). The error message doesn't give that much info...(which file? which uniform? have to look at the code...). I'll look at that.
Just a note, if someone else has to use git on Windows: Windows dynamically links with the DLLs that are in the current folder if any, so git.exe tries to use libcurl.dll that comes from the dependencies package, and which doesn't have HTTPS support...just for info.