- replace swapvalues with std::swap globally.
- added some additions to utility code from Raze, mainly to reduce file content differences.
- reduced some unused utilities
The main catch is now in D_DoomMain, only calling platform specific functions to handle the output for the error.
As a nice side effect, -norun can now be done without an exception, just by exiting D_DoomMain with a special exit code.
Now there is only one single entry point for both, instead of previously 2 entry and 4 exit points.
This also eliminates the explicit shutdown of ZMusic. Timidity++'s two buffers have been put in containers that self-destruct on shutdown and calling dumb_exit is not necessary because the only feature requiring it is not used by any code in the music library.
Instead of trying a homegrown way to avoid recursive exceptions, let's do it with the defined procedure C++ has for this case: call std::terminate.
This allowed removing some old hackery inherited from Boom and will now hopefully allow sanitizing the exit procedure to the point that it can be done without depending on exit handlers.
Since this needs to do cursor positioning calculations it's the one spot in the entire engine where UTF-8 would simply be to messy, especially when having to deal with double wide characters.
With localization for non-Latin languages on the support list the multibyte API doesn't cut it anymore. It neither can handle system text output outside the local code page nor can an ANSI window receive text input outside its own code page.
Similar problems exist for file names. With the multibyte API it is impossible to handle any file containing characters outside the active local code page.
So as of now, everything that may pass along some Unicode text will use the Unicode API with some text conversion functions. The only places where calls to the multibyte API were left are those where known string literals are passed or where the information is not used for anything but comparing it to other return values from the same API.
Previously it tried to copy all patches of composite sub-images directly onto the main image.
This caused massive complications throughout the entire true color texture code and made any attempt of caching the source data for composition next to impossible because the entire composition process operated on the raw data read from the texture and not some cacheable image. While this may cause more pixel data to be processed, this will be easily offset by being able to reuse patches for multiple textures, once a caching system is in place, which even for the IWADs happens quite frequently.
Removing the now unneeded arguments from the implementation also makes things a lot easier to handle.
As a bonus this already fixes several bugs caused by the botched texture scaling implementation the original texture manager came with.
System cursors are currently disabled because they rely on functionality that needs to be moved to different classes.
- with renderers freely switchable, some shortcuts in the 3D floor code had to be removed, because now the hardware renderer can get FF_THISINSIDE-flagged 3D floors.
- changed handling of attenuated lights in the legacy renderer to be adjusted when being rendered instead of when being spawned. For the software renderer the light needs to retain its original values.
This was originally invented to fix the sprite offsets for the hardware renderer.
Changed it so that it doesn't override the original offsets but acts as a second set.
A new CVAR has been added to allow controlling the behavior per renderer.
This was done mainly to reduce the amount of occurences of the word FTexture but it immediately helped detect two small and mostly harmless bugs that were found due to the stricter type checks.
This fixes two issues:
* timer related texture animations are not being recreated multiple times if a scene renders multiple viewpoints (e.g. camera textures or portals.)
* interpolation is smoother when maps have a high think time of multiple milliseconds. A good map to see the difference would be ZDCMP2 which has a think time of 4-5 milliseconds. With the timer taken in real time after the thinkers have run and VSync on this resulted in alternating time slices of 11 and 21 ms between frame interpolations instead of an even 16 as should be done for smooth 60 fps because roughly every second frame was offset by those 5 ms.
This one was particularly nasty because Windows also defines a DWORD, but in Windows it is an unsigned long, not an unsigned int so changing types caused type conflicts and not all could be removed.
Those referring to the Windows type have to be kept, fortunately they are mostly in the Win32 directory, with a handful of exceptions elsewhere.
