Since the SpawnedThings array is still available when polyobjects are spawned it makes no sense to create an expensive linked list in P_SpawnMapThing.
This can be done far better by scanning through the array again and collect all matching items in a second array.
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.
This code was written when the window wasn't resizable and didn't actually manage to restore it before. With today's changes this design flaw caused totally incorrect results.
Like Linux and macOS this will only support borderless fullscreen in the active desktop resolution now, which is what modern systems need.
The list of discrete resolutions has been removed as it makes no sense anymore with a fixed video mode - all the other scaling options remain active, though.
At least on faster NVidia hardware, setting this to false and gl_finishbeforeswap to true gives a better experience because it reduces screen tearing - but the same setting will reduce frame rate quite dramatically on Intel and can cause bad stalls on some older GPUs when rendering camera textures.
I did not consider that this is an init-only option. So changing the CVAR may not affect game behavior at all. Instead its value must be moved to some globally accessible variable on startup that never gets changed again.
At least one version of Windows SDK (10.0.17134.0) has broken _pgmptr/_get_pgmptr()
It points to an empty string for multi-byte character set applications
GetModuleFileName() is now used instead regardless of compiler/toolchain
Added extra guard against unexpected program paths to avoid crashes
https://forum.zdoom.org/viewtopic.php?t=60598
They were already deactivated because with 5 render modes present this was destructive, having only 2 options available.
Since the render mode can now be changed on the fly this isn't as critical anymore as it once was.
- 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 does not work with a setup where the same backend is driving both renderers.
Most of this is now routed through 'screen', and the decision between renderers has to be made inside the actual render functions.
The software renderer is still driven by a thin opaque interface to keep it mostly an isolated module.
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.
In its current form this is quite useless. What's really needed is to require a lock on the RenderBuffer for the 3D scene, but since this is not needed for the 2D stuff anymore it can be done far simpler.
This was a bad idea from the start and really only made sense with DirectDraw.
These days a FrameBuffer represents an abstract hardware canvas that shares nothing with a software canvas so having these classes linked together makes things needlessly complicated.
The software render buffer is now a canvas object owned by the FrameBuffer.
Note that this commit deactivates a few things in the software renderer, but from the looks of it none of those will be needed anymore if we set OpenGL 2 as minimum target.
This is a necessary first step for simplifying the texture handling in order to refactor it.
# Conflicts:
# src/gl/system/gl_swframebuffer.cpp
# src/textures/textures.h
# src/win32/fb_d3d9.cpp
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.
These cannot be done with the regular textures so there needs to be an option to create more than one native texture per FTexture. For completeness' sake there is also the option now to create a paletted version of a texture if the regular one is true color. This fixes a long standing problem that translations were not applied to non-paletted textures.
The native byte order converters were defined as macros which hid some issues due to lack of type checks.
Additionally the ???Long variants taking 'long' variables were removed, because longs are not always 32 bits so this could be destructive.
As a result of this, removed several DWORDs from struct definitions in i_crash.cpp.
* do not use the global temp directory. Instead create one in the AppData folder.
* removed lots of unneeded code from tmpfileplus.
* use C++ strings in there.
