| Age | Commit message (Collapse) | Author |
|---|
|
Still need to add support for ResultActions and Controls that directly
use OSystem::copyRectToScreen()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
After redoing the math, I now know that the algorithm requires for(x) { for(y) {}}
|
|
on a signed int
|
|
it from mutateImage
|
|
|
|
from the actual image dimensions
|
|
|
|
Instead of specifying a subRect. This also reverts the code that rendered to a
backbuffer and then did a full backbuffer warp. Wrapping is done in the RenderTable
|
|
a backbuffer
This makes wrapped warping much easier as well as allowing changeLocation offsets to
work properly
|
|
panorama warping
|
|
|
|
|
|
|
|
|
|
subRect coords
|
|
The general thought is int is faster than int16 or byte. So if you can afford the
space, use it over int16 or byte. Also, only use int32 when you specifically need
the 32 bits.
|
|
Control options
|
|
|
|
|
|
This is to match Marisa Chan's implementation. Images at this FoV also look better.
|
|
|
|
commands for them
|
|
This allows the panorama/tilt options to mutated without having to generate the table multiple times.
|
|
Also remove templating
|
|
|
|
|
|
absolute postions.
This allows the destRectangle to be offset within the RenderTable
|
|
The width and height were swapped when passed to the class constructor.
|
|
transposed
|
|
Aka, creating the lookup tables, storing the current render state, and eventually doing
the image mutations.
|
