(e.g. the linking book loading animation). Nobody ever noticed because the debug build was completely broken in MSVC7, and H-uru/Plasma does it completely differently.
Rename WRITE_RESTRICTED due to a collision with windows headers.
Remove PY_UNICODE_TYPE from config header. unicodeobject.h should define this correctly.
Add special signal handling for MSVC8 and up.
Include stddef.h to avoid a redefinition problem with offsetof(s,m).
Remove wchar_t typedef. This is a native type is MSVC10 and should be appropriately defined with the inclusion of stddef.h in earlier versions.
The previous implementation had the following bugs:
- DST was truncated to end of October
- DST start was off by one week in years where March 1st is a Monday
- DST start was off by one second (1:59:59 -> 2:00:00 -> 3:00:01 instead of the correct 1:59:59 -> 3:00:00 -> 3:00:01)
Tested against tzdata 2009g.
The default value of the "dst" argument must be 0, because -1 leads to incorrect results when the given time falls into local DST. However, a "dst" argument makes no sense on a method that deals with GMT anyway, so remove it entirely.
The bounds type test was performed on stale bounding data from the previous
frame resulting in the avatar being drawn without textures for a frame after
switching from 1st to 3rd person camera.
_PyTuple_Resize may destroy the original tuple and return a new one through the pointer argument. When it does that and we don?t put the new one back into the map, we end up with a stale pointer to a destroyed object in the map, which is likely to blow up one way or another next time it?s accessed.
Untested because, as far as I can see, this code isn?t actually used currently. All uses of this method deal with fixed-size SDL variables. Resizable variables are not used at all in age SDL, only in non-age SDL (animation, avatar, clothing etc.), and I?m not even sure if those are even accessible using this Python API.