add a central parse_modular_table() function which anything can use
fix up weapon_expl so that it can properly handle modular tables and LOD count changes
add support for for a fireball TBM (handled a little different than a normal TBM is since it only changes rather than adds)
use #define's for path lengths when possible so it's easier to move between functions
fix huge Cfile_stack[] issue (how the hell did that get through ✌️ QA?)
add Int3() check on cfopen() so it's easier to know if it get's called before cfile is ready to use
move path separators to pstypes.h
fix possible string overruns when setting up CFILE roots
make sure we don't try to init current directory again thinking it's a CD-ROM
add the list of VP roots to debug log, this will undoubtedly be useful
when -nosound is use go ahead and set -nomusic too to both checks are correct
add list of cmdline options to debug log
fix possible overwrite issues with get_version_string() and remove '(fs2_open)' from string plus change OGL->OpenGL, D3D->Direct3D
split up WinMain() and main() so it doesn't resemble ifdef hell
rename WinMainSub() to game_main() and move anything that should have been in WinMain() to WinMain()
add -cache_bitmaps and have bitmap caching between levels off by default
when -cache_bitmaps is used then use C-BMP for top-right memory listing, and just BMP otherwise
- very minor performance boost from not doing stupid things :)
- minor change to 3d shockwave sizing to better approximate 2d effect movements
- for shields, Gobal_tris was only holding half as many as the game can/will use, buffer is now set to full size to avoid possible rendering issues
- removed extra tcache_set on OGL spec map code, not sure how that slipped in
always use bm_set_render_target() rather than the gr_ version so that the graphics state is set properly
save the original gamma ramp on OGL init so that it can be restored on exit
more non-standard res fixing
- I think everything should default to resize now (much easier than having to figure that crap out)
- new mouse_get_pos_unscaled() function to return 1024x768/640x480 relative values so we don't have to do it later
- lots of little cleanups which fix several strange offset/size problems
- fix gr_resize/unsize_screen_pos() so that it won't wrap on int (took too long to track this down)