SDL3 uses ID when dealing with monitors, unlike SDL2 which uses Index
for the same thing. This problem was already fixed in multiple places
by use of preprocessor branches, so I did the very same thing.
Please, notice that this is a pretty bad solution to this problem,
and I only did it to keep it consistent with the rest of the code.
The more about why it's not correct is mentioned here:
https://github.com/raysan5/raylib/issues/5256#issuecomment-3429156919
Hopefully, someone will refactor it someday :)
Fixes: https://github.com/raysan5/raylib/issues/5256
* add base of rlsw.h
* implement state support
Also replace the triangle rasterization functions with macros that generate specific functions for each state of the rendering system.
Also, add the OpenGL definitions in order to add a binding for rlgl.
* branchless float saturation
* apply perspective correction to colors
* impl line clipping and rasterization
+ tweak function names
* impl face culling
* impl color blending
* fixes and tweaks
* add clear buffer bitmasks
* small optimizations / tweaks
* review ndc to screen projection
* avoid to recalculate MVP when its not needed + tweaks
* review the loading and management of textures
to be closer to the OpenGL API
* texture sampling optimization
* review get pixel functions
+ review unorm/float conversion
* add several buffer format support
Several depth and color formats have been added for the framebuffer.
8-bit, 16-bit, and 24-bit formats are now available for depth.
RGB 8-bit (332), RGB 16-bit (565), and RGB 24-bit (888) formats are now available for color.
Alpha support is no longer present for the framebuffer at the moment, but it can easily be restored by adding the formats and reinterpolating the alpha in the areas that do not perform color blending.
Additionally, this commit brings performance improvements.
* tweaks
* impl line width
* impl points + point size
* fix and improve polygon clipping functions
* impl polygone modes
* add some not planned functions
- `glDepthMask`
- `glColorMask`
* framebuffer resizing + handle init failure
* add quick notes about line clipping algorithms used
* start to impl scissor test + review line clipping
The support for the scissor test has been implemented for clearing as well as for triangle clipping.
The implementation for lines and points is still missing.
I also removed the 2D clipping of lines that used the Cohen-Sutherland algorithm, opting instead to always use the Liang-Barsky algorithm in all cases.
This simplifies the implementation, and the 2D version would have caused issues when interpolating vertices in the future if we want to implement additional features.
* review scissor clear
* review `swScissor`
* impl line scissor clipping
* round screen coordinate (line rasterization)
* impl point scissor clipping
* remove unused defs
* add getter functions
* gl binding
* add `glHint` and `glShadeModel` macros (not implmented)
* binding tweaks
* impl copy framebuffer function + glReadPixels
* review `swCopyFramebuffer`
* update rlgl.h
* update rlgl.h
* texture copy support
* fix typo..
* add get error function
* def sw alloc macros
* reimpl get color buffer func
just in case
* remove normal interpolation
* review texture wrap
* fix ndc projection (viewport/scissor)
* impl framebuffer blit function
* reduce matrix compuations and memory usage
* swBegin tweaks
* preventing a possible division by zero
* remove useless scissor related data
* review color blending system
* greatly improve float saturation
* tweak lerp vertex function
* use opitmized fract function in sw_texture_map
* tweak framebuffer functions for better readability
* optimized copy/blit functions for each dst format
* review framebuffer filling functions
* impl specific quad rendering func
* use of a single global vertex buffer
* fix 'sw_poly_point_render'
* added `SW_RESTRICT` and redesigned `sw_lerp_vertex_PNCTH`
* tweak the pipeline flow regarding the face culling
avoids misprediction, improves vectorization if possible
* new rendering path for axis aligned quads
* oops, translating some comments
* use of `restrict` for blending function parameters
* update rlgl.h
* adding `GRAPHICS_API_OPENGL_11_SOFTWARE` in `DrawMesh`
* add `RL_OPENGL_11_SOFTWARE` enum
* temp tweak
* build fixes
* fix DrawMesh for GL 1.1
* update swClose
* review texture format + fix copy
* set minimum req vertices to 3 (quads)
* check swInit
* review pixelformat
* tweaks
* fix animNormals (DrawMesh)
* fallback color/texcoord (swDrawArrays)
* review swMultMatrixf
* fix texture pool alloc..
* review triangle scanlines
increment all data
* fix `sw_quad_sort_cw`
* impl sdl platform
* rm def
* increase max clipped polygon vertices
* improve triangle rasterization along Y axis
improved robustness against numerical errors
incremental interpolation along Y
simplified function, fewer jumps
* review current vertex data
+ increase max clipped polygon vertices (for extreme cases)
* fix and improve polygon clipping
Sets the vertex count to zero when the polygon is invalid
Stops clipping when the vertex count drops below 3
* fix gradient calculation
* cache texture size minus one + comments
* tweaks
* BGRA copy support
* adding software backend option (cmake)
* update Makefile
* fix face culling
* excluse some exemple with the software backend
* review SW_CLAMP case in sw_texture_map
* review sw_saturate
* review line raster
* fix sw_quad_is_aligned
* review sw_raster_quad_axis_aligned
* tweaks
* codepoint fix (?)
* fix var name...
* rcore_drm software renderering
* cleanup and tweaks
* adding support for `GL_POINT_SIZE` and `GL_LINE_WIDTH` get
* fix sampling issue
* fix swBlendFunc
---------
Co-authored-by: Ray <raysan5@gmail.com>
* gate with OPENGL_11 prototype and definition for rlSetPointSize and rlGetPointSize
* gate with OPENGL_11 prototype and definition for rlSetPointSize and rlGetPointSize
* more coverage for fixing expected texcoords, normals, and colors
* fix guard styling
* fix guard issue, and remove fallback unneccessary else clause (default shader ill work)
* opengl3.3 and es2 need the color array allocated in order to allow for updates later (unlike opengl11)
Currently, scaling doesn't work correctly on macOS (see #5185). This commit
works around this issue by disabling SCALE_FRAMEBUFFER on macOS when
`FLAG_WINDOW_HIGHDPI` is unset.
* fix definition to use SDL3
* [cmake] fix definition to use SDL3
* [build][cmake] prefer USING_VERSION_SDL3 over PLATFORM_DESKTOP_SDL3
* [build][cmake] when libs are added externally, consumer needs to import them too
The following code would crash the previous version when calling MemFree:
// 53 * A
const char maliciousBase64Input[] = "AAAAAAAAAAAAAAAAAAAAAAAA"
"AAAAAAAAAAAAAAAAAAAAAAAAAAAAA";
int decodedSize = 0;
unsigned char *decodedData = DecodeDataBase64(
maliciousBase64Input, &decodedSize);
if (decodedData) {
MemFree(decodedData);
}
The reason is a lack of array bound checks in the decoding loop, which
corrupted here the heap (though this is platform dependent).
Adding the bound checks here prevents the memory corruption.
Tested with encoding random data of sizes 0-1023 and comparing it
with the decoded result.
* add const qualifier to the first arg of TextJoin
* Update raylib_api.* by CI
* Update raylib.h
---------
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Ray <raysan5@gmail.com>
This redesign is a big improvement on font loading time and memory requirements. It only loads glyphs available on font from requested codepoints and only processes those glyphs for packaging. When processing +10K codepoints (CJK), the loading time improves considerably.
Previously stated it was the near plane width in orthographic, in actuality it's the height. The wording has been fixed, with added clarity that it's in world units so no conversion is needed.
* win32 backend
* [rcore][win32] use SwapBuffers instead of wglSwapLayerBuffers
I don't understand OpenGL well enough to know the difference between
SwapBuffers and wglSwapLayerBuffers but the former seems to double
my FPS (from 2000 to about 4000 in core_vr_simulator).
* [rcore][win32] stop lying to the OS about when our window is updated
Instead of calling BeginPaint/EndPaint in WM_PAINT which signals to the
OS that our window content is updated, now when we encounter the WM_PAINT
message instead we return back to the raylib application which will
trigger it to render a new frame. We've replaced the call to BeginPaint
and EndPaint with a call to ValidateRect in SwapBuffers, which, means
we're now correctly telling the OS when our window content is actually
up-to-date.
Note that this doesn't fix the window content not being updated during
a window resize/move beacuse thos have their own message loop which
doesn't return early when it's time to paint.
* [rcore][win32] fallback to finding functions from opengl32.dll
* [rcore][win32] fixes from review/for gcc
* [rcore][win32] incorporate style conventions
* [rcore][win32] workaround unused function error SetupFramebuffer
* [rcore][win32] re-enable sanitize flags check for MSAA_4X
* [rcore][win32] more style changes and remove old DPI cases
Added some more missing spaces after conditional statements. Also
made unsupported MSAA_4X an assert instead of an abort and also
removed dpi-aware cases for older OS's. More changes would be needed
to support those OS versions, namely, removing the dependency
on shcore.
* [rcore][win32] fixes for compling with w64devkit without -DUNICODE
* Update build.zig
* Update build.zig
---------
Co-authored-by: Ray <raysan5@gmail.com>
Added CMake support for SDL3.
Now supports including SDL2 or SDL3 as a subdirectory within the project. The system will first check for SDL3, then SDL2. If neither is found, it will attempt find_package(SDL3), followed by find_package(SDL2). If all these checks fail, the process will terminate with an error.
- make sure that src/external/rl_gputex.h uses it's own macros
for printing warnings and allocating memory
- add few additional macros in order to decouple rl_gputex from headers
that are being included by Raylib (e.g. *_MEMSET, *_NULL and so on)
- make sure that rl_gputex uses RL_*ALLOC, TRACELOG and RLAPI macros
when (and only when) being included by src/rtextures.c
- replace LOG() and RL_GPUTEX_SHOW_LOG_INFO macros with: RL_GPUTEX_WARN
and RL_GPUTEX_SHOW_WARN_INFO (this is a breaking change, but it was
broken and unusable anyway, see:
https://github.com/raysan5/raylib/issues/5039#issuecomment-3065732596 )
- fixes issue all issues mentioned in:
https://github.com/raysan5/raylib/issues/5039#issuecomment-3065732596
- remove logging upon successfully saving a file; Raylib does this
anyway and we want rl_gputex to only print WARNings upon failures;
see: https://discord.com/channels/426912293134270465/541710686468702250/1394406734306480352
- add additional condition when saving a file; now checks for both
fwrite() and fclose() failures; before it would warn about both,
but only return result based on fclose()
- add some notes about the current state of compiling rl_gputex
without depending on Raylib (spoiler: it's still broken...)
- bump rl_gputex version to 1.1 since this is a potential breaking
change (only for people attempting to use it without Raylib)
Fixes: https://github.com/raysan5/raylib/issues/5039
Reference: https://discord.com/channels/426912293134270465/541710686468702250/1394403611852931255
I did this with the intention of fixing
https://github.com/raysan5/raylib/issues/5027
but it's hard to tell if it will help in that case or not.
Anyway, bumping the header with mappings should be done once in a while.
I haven't found any Raylib script for doing this, so here's how I did it
step-by-step with CMake rule provided by GLFW (I'm on Linux, but this
should be similar on any other POSIX system):
```sh
$ git clone https://github.com/raysan5/raylib.git
$ cd raylib/src/external/glfw/
$ cmake -B ./build -DGLFW_BUILD_EXAMPLES=OFF -DGLFW_BUILD_TESTS=OFF -DGLFW_BUILD_DOCS=OFF -G "Unix Makefiles"
$ make -C ./build update_mappings
$ rm -fr ./build
```
then I checked, if any mappings were deleted (without updating) with
following shell one-liner:
```sh
# you must be at 'raylib/src/external/glfw/' when executing this
$ for i in $(git diff ./src/mappings.h | sed -n -e 's/^-"\([0-9a-z]*\),.*$/\1/p'); do grep <./src/mappings.h "$i" > /dev/null || echo "$i"; done;
```
which gave me the following IDs:
```txt
0300000066f700000500000000000000
030000000d0f00001600000000007803
03000000efbe0000edfe000000000000
03000000786901006e70000000000000
050000005e040000e002000030110000
03000000790000000600000010010000
030000000d0f00000d00000000010000
030000000d0f0000d800000072056800
03000000c0160000dc27000001010000
030000008916000000fd000024010000
```
I have checked all those IDs with `git log -p` against
https://github.com/mdqinc/SDL_GameControllerDB
and it seems that all of them were deleted by upstream (I was afraid that
maybe some mappings were added manually by a Raylib contributor and then
overwritten by my patch, but they weren't; All good!)
Fixes a regression since 5.5, where `ScanDirectoryFilesRecursively` no
longer does the "recursively" part due to `path` being `static`.
The issue was once already fixed in
5530a3ceb8
but recently made it back it in.
The original implementation created/destroyed framebuffers (FBs) per-frame, leading to kernel overhead and screen tearing. This commit replaces it with a different approach using:
- Asynchronous `drmModePageFlip()` with vblank sync
- Framebuffer caching to reduce repeated FB creation/removal operations
- Proper resource management through BO callbacks and buffer release synchronization
- Added error handling for busy displays, cache overflows, and flip failures
- Event-driven cleanup via page_flip_handler to prevent GPU/scanout conflicts
Co-authored-by: rob-bits
Examining the code shows that the rectangle is drawn winding counterclockwise, starting
with the top left. Therefore the colors used should be in the order: topLeft, bottomLeft,
bottomRight, topRight.
However, the variables actually being used are topLeft, bottomLeft, topRight,
bottomRight. I was confused by this as I was getting striping where I didn't expect any.
Put another way, the last two parameters are misnamed.
This diff swaps the parameter names and their usages. The result is that no runtime
behaviour changes: the same parameter order yields the same visual result both before and
after this change, but the parameter names now correctly reflect what they are actually
used for.
You can actually see this in the implementation of DrawRectangleGradientV, which
(correctly) passes top, bottom, bottom, top to DrawRectangleGradientEx.
I believe it makes sense to only do this when there are no known
touch points. I am not even sure if this should be done at all.
See https://github.com/raysan5/raylib/issues/4872 for more information.
Restore window currently says it sets the window state to:
"not minimized/maximized"
However, if a window is maximized and then minimized, it's typical that
it would restore back to being maximized, which is what seems to happen
from my testing. I've reworded the description to better reflect
this behavior.
raylib app crashing when started and a gamepad is already connected to the PC (even if the gamepad is not used in the app). I only tested this with a gamepad that has a layout which is not recognized. Using SDL3 as backend.
InitWindow() prints CWD during initialization,
but ChangeDirectory() does not, which is quite confusing when you start
messing with CWD. Now said function should log similar message.
Prevents unnecessary work and division by zero (when segments=0) in DrawCircleSector/DrawCircleSectorLines when startAngle equals endAngle, matching existing behavior in DrawRing/DrawRingLines.
This may not be the correct approach, however this appears to work. The idea is that before modifying `CORE.Window.flags` when first creating the window we keep a copy of the flags in order to call `SetWindowState` after initialization has completed, which should behave as if `MaximizeWindow` or `MinimizeWindow`, or conceptually any other flag modifying function were called after `InitWindow`.
This pull request only performs this for the windows platform, modify as needed in the switch statement at the end for others.
Undefine DEBUG to avoid external redefinition
warnings/conflicts. This is probably a common
definition for many external build systems'
debug configurations.
This ensures raylib will not emit
a warning about the DEBUG definition being
redefined in external build systems.
* [rcore] fix crash in InitWindow, due to unchecked result of InitPlatform
Check the result of InitPlatform(), if it isn't 0, report the Error
and return. This prevent crashes and allows for gracefully aborting
or recovering by checking IsWindowReady().
Partially-fixes: https://github.com/raysan5/raylib/issues/4801
* [rcore] style: store the result of InitPlatform() before checking it
Small style change that doesn't impact how the code behaves here.
Variable 'result' is not used anywhere else in this block,
it's just for compliance with Raylib's coding conventions.
Requested in PR: https://github.com/raysan5/raylib/pull/4803#discussion_r1976502788
* [rcore] use LOG_WARNING when InitPlatform() fails
...as this is preferred over LOG_ERROR.
Requested-by: https://github.com/raysan5/raylib/pull/4803#discussion_r1976651724
In the case of a failure within miniaudio on the function: ma_convert_frames, the dynamic memory allocated for the `data` variable will leak on the early return.
Raylib allows for providing custom allocators via macros.
GLFW supports this too, but via function pointers.
Make sure that GLFW uses those Raylib macros, by wrapping them in
function calls and setting them up inside of InitPlatform().
This is possible because of glfwInitAllocator() and GLFWallocator.
Fixes: https://github.com/raysan5/raylib/issues/4776
Relates-to: https://github.com/raysan5/raylib/issues/4751
I was really wondering what is going on here :D I believe this code
tried initially to out-cast 'const' specifier but this is not needed here
at all. Currently, it just confuses whoever reads this so I changed it.
The old code would also trigger -Wcast-qual warning on some compilers.
Currently, a warning about _GNU_SOURCE being redefined is emitted when
compiling rglfw.c
In file included from rglfw.c:99:
external/glfw/src/posix_poll.c:27:9: warning: "_GNU_SOURCE" redefined
27 | #define _GNU_SOURCE
| ^~~~~~~~~~~
<command-line>: note: this is the location of the previous definition
This can be avoided by not defining _GNU_SOURCE on the command line for
this file.
Defining feature test macros in source code is not really good practice
so this should probably reviewed in glfw itself, at least to maybe check
#ifdef _GNU_SOURCE first. But for now this change will suffice.
Fixes#4725
When asking Makefile to create SHARED library for WEB
$ make TARGET_PLATFORM=PLATFORM_WEB RAYLIB_LIBTYPE=SHARED
it would instead silently create STATIC library
thus not fulfilling the request as expected
This commit adds an error in this case and stops further execution.
This is not consistent with Cmake, because Cmake throws the warning and
does not stop, but Cmake can easily recover from this case and people
probably does not even notice it. However, Makefile is something that
you have to handle yourself and you have to recover from any issues so
having an error and aborting with exit code 1 is more expected.
Otherwise people may spend a lot of time debugging Makefile in order to
understand what's even going on.
Fixes: https://github.com/raysan5/raylib/issues/4717
define STBIR_NO_SIMD when __TINYC__ is defined so stb_image_resize2 will
not include *mmintrin which are not supported by all compilers.
There are similar checks for __TINYC__ already elswere in raylib
and they are also mostly there to disable SIMD headers.
Additionally, move similar check for stb_image, to be a little bit
deeper. Before it was defining STBI_NO_SIMD without including stb_image
It was also clashing with note, causing said note to make no sense.
Fixes: https://github.com/raysan5/raylib/discussions/2994
Reference: https://github.com/nothings/stb/issues/1738
This change is replacing the hardcoded "#canvas" element references in
rcore_web to allow using canvas elements that use different names
(which is necessary when using multiple canvas elements on one page).
Also adding a cursor hiding example to mouse example.
Lately got some compilation `errors` related, it seems GCC 14 interprets some `const`-missmatch as errors instead of warnings (as previous versions).
But in any case, I don't see why an user won't be able to operate directly over of those returned buffers; `const` adds a restriction (for security reasons?) that in my opinion is useless.
From an expert on compilers (w64devkit creator), here there are some notes I agree with:
```
No const. It serves no practical role in optimization, and I cannot recall an instance where it caught, or would have caught, a mistake. I held out for awhile as prototype documentation, but on reflection I found that good parameter names were sufficient. Dropping const has made me noticeably more productive by reducing cognitive load and eliminating visual clutter. I now believe its inclusion in C was a costly mistake.
(One small exception: I still like it as a hint to place static tables in read-only memory closer to the code. I’ll cast away the const if needed. This is only of minor importance.)
```
Ref: https://nullprogram.com/blog/2023/10/08/
This optimization works in the following ways:
1. Reduces calls to malloc to 1. Instead of needing an extra array, we
can just swap the top half with the bottom half of the one array.
2. Unroll the inner for loop and remove a condition. Unrolling loops
buys some performance wins, but the real goal was to remove the if check
and just set the alpha channel to 255.
On my hidpi arm64 laptop, I saw ~60% improvement in performance in my
debug build (29 FPS vs 47 FPS). When optimized, the gains were roughly
10% (75 FPS vs 83%).
Signed-off-by: K. Adam Christensen <pope@shifteleven.com>
* Add implementation for CORE.Window.eventWaiting on PLATFORM_DESKTOP_SDL
* Optimize GetFrameTime() reset
* Optimize FLAG_WINDOW_ALWAYS_RUN and GetFrameTime() reset for PLATFORM_DESKTOP_GLFW
* Enable FLAG_WINDOW_ALWAYS_RUN by default on PLATFORM_DESKTOP_GLFW
* Revert enabling FLAG_WINDOW_ALWAYS_RUN by default on PLATFORM_DESKTOP_GLFW
* Add implementation for FLAG_WINDOW_ALWAYS_RUN on PLATFORM_DESKTOP_SDL
* Add reset for GetFrameTime()
* Update rmodels.c
resolves segfault with missing bone weights or bone IDs
* Update rmodels.c segfault with animation and missing normals/animnormals
* correct place.
- Updating bones only once instead for each mesh.
- Updating only one `model.meshes[].boneMatrices` and then using deep copy for other meshes instead of calculating for each bone in each mesh.
**Other points:**
- Makes it a clean base/template/reference for bone updation functions. Because if using this as template then some calculations done in one mesh can affect bones in other mesh in next iteration(doubles the effect in for next mesh).
Signed-off-by: Kirandeep-Singh-Khehra <kirandeepsinghkhehra@gmail.com>
* Make sure ShaderUniformDataType matches rlShaderUniformDataType
* Update raylib_api.* by CI
---------
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: Ray <raysan5@gmail.com>
* Improve cmake config file generation
This allows for finding raylib on a non-standard location (eg, not in /usr/lib{,64})
* Only have glfw as private if using internal glfw and not static
* increased vertex data precision to match blender exports
* modified memory allocation to better fit higher precision data
* now accounting for newline characters during allocation
* now accounting for potentially higher vertex counts
* removed unnecessary final newline
The nextShapeEnd integer is a pointer in the OBJ data structures.
The faceVertIndex is a vertex index counter for the mesh we are
about to create. Both integers are not compatible, which causes
the code to finish the meshes too early, thus writing the OBJ data
incompletely into the target meshes.
It wasn't noticed because for the last mesh, it process all remaining
data, causing the last mesh to contain all remaining triangles.
This would have been noticed if the OBJ meshes used different textures
for each mesh.
* [rcore] add 'GetClipboardImage' for windows
* [rcore] GetClipboardImage removed some unneeded defines
* [rcore] PLATFORM_SDL: create a compatility layer for SDL3
* external: add win32_clipboard.h header only lib
* [rcore] using win32_clipboard on platforms rlfw and rgfw
* [rcore] fix warnings in SDL3 compatibility layer
* Makefile: Allow specifying SDL_LIBRARIES to link, this helps with SDL3
* Makefile: examples makefile now compile others/rlgl_standalone only when TARGET_PLATFORM is PLATFORM_DESKTOP_GFLW
* Makefile: examples makefile now compile others/rlgl_standalone only when TARGET_PLATFORM is PLATFORM_DESKTOP_GFLW
* [rcore]: PLATFORM_SDL: improve clipboard data retrieval
* external: remove unused function from win32_clipboard.h
* Makefile: allow for extra flags necessary when compiling for SDL3
* [rcore]: fix string typo
* [rcore]: Properly handle NULL dpi passing. As is allowed in SDL2
* external: fix arch finding on win32_clipboard.h to allow compilation on msvc cmake CI
* [rcore]: PLATFORM_SDL: Treat monitor as an ID in SDL3 as opposed to an index as in SDL2
* [rcore]: typo
When creating a new sphere mesh with high number of slices/rings the top and bottom parts of the generated sphere are removed. This happens because the triangles in those parts, due to high resolution, end up being very small and are removed as part of the 'par_shapes' library's optimization. Adding par_shapes_set_epsilon_degenerate_sphere(0.0); before generating the sphere mesh sets the threshold for removal of small triangles is removed and the sphere is returned to raylib correctly.
* Remove the panorama cubemap layout, it was not implemented.
Left a todo in the code for some aspiring developer to finish.
* Update raylib_api.* by CI
---------
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
* Update raylib_api.* by CI
* Add math operators for C++ to raymath
* better #define for disabling C++ operators
---------
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
* Updates SetWindowState() and ClearWindowState() to handle FLAG_WINDOW_MAXIMIZED for PLATFORM_WEB
* Update MaximizeWindow() and RestoreWindow() to set/unset the FLAG_WINDOW_MAXIMIZED
* (rcore_desktop_rgfw.c) fix errors when compiling with mingw
* define WideCharToMultiByte
* update RGFW
* move stdcall def to windows only
* fix raw cursor input
