This stuff is software rendering only so should be made optional anyway.
But aside from that it was the last non-sound data, being loaded into the cache.
With this gone the sound cache can be redone as a dedicated object and the atrocious cache1d implementation be tossed out.
Models are currently non-functional and need to reroute their texture handling to the global texture manager instead of handling everything themselves.
Voxels also need a bit of work to make their texture management more automated.
Sorry, but having a globally writable pointer to every texture is just insane and makes any functional management impossible.
This is merely a preparation for adding a real texture manager. That cannot be done if any code can write over the data at will. For that, it now has to make the texture writable first or create a writable empty texture.
These will have to do some texture management bookkeeping so directly changing the values is problematic.
This required changing the parameter interface in polymost.cpp because a few places hacked around with the global state to pass parameters to subfunctions.
The only difference was that bloodhack came from NBlood and playing_blood was added for unhandled differences between the backends of EDuke32 and NBlood.
- also removed the legacy fog modes because their implementation did not mix well with the texture system - they are also not really useful to begin with.
The last fog mode will most likely also be removed once true color shading is working but that one is less of an issue.
The EDuke32 and RedNukem frontends are working, Blood isn't yet.
Notes:
many of the CMake variables and its output still refer to zdoom. Before changing that I wanted to make sure to be able to commit something that works.
support code for Windows XP has been entirely removed. On Windows this will only target Vista and up.
the crc32.h header had to be renamed to deconflict from zlib.
several Windows API calls were changed to call the A-versions directly. Weirdly enough there were places that defined their parameters as T types but in a non-working way.
removed some remaining editor files and support for the native software rendering only Windows backend.
in a few simple cases, replaced 'char' with 'uint8_t'. The code as-is depends on chars being unsigned which is non-portable. This needs to be carefully reviewed.
Using that higher precision, interpolate at a higher granularity.
Further, truncate the target interpolation time to vertical blank boundaries to avoid producing temporal artifacts.
Fix issues caused by interpolation calculations being handled differently in multiple places (and fix cases where smoothratios were being thrown away only to be redone without checking all proper conditions).
Ensure ClockTicks changes do not break other targets (EKenBuild, VoidSW), but note any interpolation there is not similarly updated.
git-svn-id: https://svn.eduke32.com/eduke32@8050 1a8010ca-5511-0410-912e-c29ae57300e0
# Conflicts:
# source/build/src/build.cpp
# source/build/src/sdlayer.cpp
I don't know that this is any faster, but there's something to be said for standardization and consistency. I will be making most of this stuff use bitmap_set/test/clear() soon.
git-svn-id: https://svn.eduke32.com/eduke32@7876 1a8010ca-5511-0410-912e-c29ae57300e0
# Conflicts:
# source/build/src/build.cpp
# source/build/src/polymost.cpp
This was removed a long time ago because Duke3D and Shadow Warrior don't use it.
git-svn-id: https://svn.eduke32.com/eduke32@7873 1a8010ca-5511-0410-912e-c29ae57300e0
# Conflicts:
# source/build/src/polymer.cpp
# source/build/src/polymost.cpp
and additionally modifying G_SetupCamTile() to call EVENT_DISPLAYROOMSCAMERATILE after calling renderSetTarget()
git-svn-id: https://svn.eduke32.com/eduke32@7852 1a8010ca-5511-0410-912e-c29ae57300e0
These updatesector counterparts are more performant for tasks that require determining whether nearby sectors exist at a point as they will not exhaustively search the map.
This also leads to more correct behaviour when sectors are disjoint, as queries will breadth-first search neighbouring sectors rather than risk accidentally stumbling first across accomodating disjoint sectors.
git-svn-id: https://svn.eduke32.com/eduke32@7805 1a8010ca-5511-0410-912e-c29ae57300e0
# Conflicts:
# source/build/include/build.h
Additionally, raise minimum number of stencil buffer bits to accomodate future use of the stencil buffer.
git-svn-id: https://svn.eduke32.com/eduke32@7736 1a8010ca-5511-0410-912e-c29ae57300e0
# Conflicts:
# source/build/include/build.h
# source/build/src/polymer.cpp
# source/build/src/polymost.cpp
# source/kenbuild/src/game.cpp
It's garbage and the "functionality" it provided is something that will be able to be achieved in a better way in a future commit.
git-svn-id: https://svn.eduke32.com/eduke32@7708 1a8010ca-5511-0410-912e-c29ae57300e0
This was necessary because everything is already allocated with the Xmalloc() functions, but a future commit will make blocks allocated with those functions no longer compatible with the system implementation of free(), which Bfree() wraps.
git-svn-id: https://svn.eduke32.com/eduke32@7705 1a8010ca-5511-0410-912e-c29ae57300e0
# Conflicts:
# source/build/src/build.cpp
# source/build/src/mdsprite.cpp
# source/build/src/polymer.cpp
# source/build/src/polymost.cpp
# source/build/src/texcache.cpp
# source/build/src/voxmodel.cpp
Some of the variable initialization ordering may look counterintuitive at first, but it's intended to avoid read-after-write memory access penalties.
git-svn-id: https://svn.eduke32.com/eduke32@7701 1a8010ca-5511-0410-912e-c29ae57300e0
This may also fix a minor performance regression introduced in Polymer sometime after we switched to building as C++, caused by confusion between STRUCT_TRACKERS_ENABLED and UNTRACKED_STRUCTS. UNTRACKED_STRUCTS is now named UNTRACT_STRUCTS__ to dissuade use outside of buildtypes.h.
git-svn-id: https://svn.eduke32.com/eduke32@7695 1a8010ca-5511-0410-912e-c29ae57300e0
# Conflicts:
# source/build/src/polymer.cpp
