did not mark eax as an early-clobber register, so GCC might decide to pass the memory address
in eax, and it would get clobbered by the inline assembly before fetching the value to divide
by.
But rather than fix it by adding another '&', I have opted to mark it as in/out and do the
zeroing outside the inline assembly, so GCC has maximum flexibility for scheduling the code.
SVN r3242 (trunk)
GETPALOOKUP. The end result is that there is a minimum distance around you where light
amplification stops and it gets no brighter. Should this scale with visibility? I can't say.
So, yeah, it turns out all these years ago, I made this out to be harder than it really is.
SVN r3224 (trunk)
also decided to compile some other shaders slightly differently, too.)
- Fixed: The InGameColormap had been designed without taking alpha into consideration.
As the least likely parameter to be used, desaturation has been moved into a constant
register to make room for the alpha parameter to live in the vertex's color value.
SVN r3208 (trunk)
ignored the Y locations of patches drawn on two-sided midtextures and always drew them at the
top of the texture. Added a compatibility flag.
SVN r3205 (trunk)
(e.g. invulnerability). This should only be necessary for SM1.4 cards with buggy drivers. I
doubt the problem is very widespread. The one piece of SM1.4 hardware I obtained specifically
for testing purposes has no problems using shaders for these effects.
SVN r3200 (trunk)
- Fixed: When DDrawFB::Lock() has to recreate resources, it left the LockCount at 0. This causes
problems if something else locks it before it is unlocked, because the second locker will
think it is the first. This happens in R_RenderViewToCanvas(). See DDrawFB::PaletteChanged()
for the most common reason why Lock() would need to recreate resources.
- Fixed: DDrawFB::CreateSurfacesComplex() had debugging cruft left in that skipped all but the
last attempts.
- Fixed logging of video debug info to a file to not multiply define dbg.
SVN r3195 (trunk)
sides of missing textures, this can take a very long time, because each missing textures causes
a scan of every single line (for the sake of packed sidedefs), and each output line also requires
an update of the hidden RichEdit logging control.
SVN r3192 (trunk)
non-zero variables to load. This was fine when all variables defaulted to zero, but map variables
have been able to default to something else for some time now. This meant that if all map
variables were zero when the game was saved, they would be restored to their default values
when the game was loaded.
SVN r3190 (trunk)
of the field of vision, since it reduces the chances of duplicate sky columns output next to
each other compared to a straight tangent-to-angle calculation.
SVN r3189 (trunk)
