Age | Commit message (Collapse) | Author |
|
|
|
|
|
|
|
It seems it is possible that some objects or subsequent click fields
'pointed' to by next are outside of the initialized range and are
expected to be zeroed. This would explain the non-deterministic freezes
I experienced.
|
|
Besides the addition of inverse sections, the 'fullscreen' parameter was
removed as it was used only for testing purposes in the beginning.
|
|
This code section was converted from the original source, where the
sections are directly drawn to the screen instead of buffered.
If a section > 128 is set as parameter, the function takes the
dimensions of section - 128 and draws this region of section 0 to the
screen, thus restoring it.
It would not make sense loading part of section 0 in seperate Surfaces
especially since kMaxSection is smaller than 128 the else branch is
never entered so removing it doesn't change the programs behavior.
|
|
|
|
|
|
Substitute size_t with ScummVM defined uint type
|
|
|
|
While the 640x480 images are loaded correctly, they are displayed
cropped to 320x240, the default resolution.
|
|
|
|
Sections with an x2 value of 0 cannot span any area and are thus invalid
|
|
|
|
Colors on VGA are 6bit per color channel, that is why the colors
were so dark in earlier commits. Shifting the colors left by 2 gives us
an approximated value for CLUT8.
|
|
Palette color 255 doesn't seem to be used
|
|
|
|
Mistakenly, during image loading the system palette was completely
substituted by the image palette although the first 16 colors are
supposed to be preserved.
|
|
|
|
|
|
|
|
There has been many other changes like copy/pasting room definitions,
skeleton for image decoder, ..
Getting supernova mod format support seems to be a good next step.
|