This argument is now non optional, but this never hits the same bad access.
I voted to simplify the code here since the argument is never used optionally in our codebase.
For some reason AFAICT mesa reports a feature as enabled even when its
extension isn't supported. The Vulkan specification says nothing aboutd
this so this is technically more of a workaround, but it works.
Previously, only forward basis distance from the camera was used.
This means that unnecessarily high LOD levels were used for objects located to the side of the camera.
The distance from the camera origin is now used, independently of direction.
Replace all TODO uses of `#warning` by proper TODO comments, and will open
matching bug reports to keep track of them.
We don't have a great track record fixing TODOs, but I'd wager we're even
worse for fixing these "TODO #warning" so we should prohibit this usage.
- Outright disable spammy warnings due to past or present GCC bugs:
* `-Wno-strict-overflow` for GCC 7.
* `-Wno-type-limits` for GCC before 11 (regressed in 9/10, might work in
earlier releases but at this stage we don't care).
* `-Wno-return-type` for GCC 12/13 (regression, still not fixed).
- Enable extra warnings conditionally when broken on earlier GCC:
* `-Wnoexcept` was removed due to an upstream regression in GCC 9, could
be re-enabled (but commented out for now as we actually have `-Wnoexcept`
warnings to fix.
* `-Wlogical-op` was broken on our variadic templates before GCC 11, now
seems fine.
Mainly:
- Make `max_descriptors_per_pool` project setting Vulkan-specific.
- Use a common, render driver agnostic magic FourCC for shader binary data.
- Downgrade spirv_reflect to Vulkan-only dependency.
- Add a `RENDER_DRIVER_*` macro to GLSL shader code for per-driver customizations.
This removes the countless small UBO writes we had before
and replaces them with a single large write per render pass.
This results in much faster rendering on low-end devices
but improves speed on all devices.
End users would get spammed with messages of varying verbosity due to the
mess that thirdparty layers/extensions and drivers seem to leave in their
wake, making the Windows registry a bottomless pit of broken layer JSON.
I'm all for helping end users clean up mess in their registry / system paths
for Vulkan ICDs, layers and extensions, but the way this is done by
VK_EXT_debug_utils is just horrible - and the way for them to fix it (manual
edit of system files) is also not a good thing to recommend.
Closes countless issues where users think Godot is broken because it reports
weird errors.
- `LIBC_FILEIO_ENABLED` wasn't defined anywhere, even in _other platforms_.
- `NO_NETWORK` is also never defined. It probably isn't enough anyway to
disable network APIs in the current codebase.
- `UNIX_SOCKET_UNAVAILABLE` is never defined in this code but used by some
other platforms, clarify that.
- `NO_STATVFS` can be removed as Android supports it since API level 19,
which is our current min SDK level. It's also only used for
`DirAccessUnix::get_space_left()` which is anyway overridden by
`DirAccessJAndroid::get_space_left()` so it shouldn't make a difference.
* Fixed documentation for `DirAccess.get_space_left()`.
- `NO_FCNTL` is likely also a remnant of early Android days, in current NDK
r23 it seems to be available. Also cleaned up unused `fcntl.h` includes.
- `NO_ALLOCA` is never defined, and we use alloca in many places now.
`CreateDirectoryW()` chokes on absolute paths that contain `..`
example: "C:\\workspace\\..\\games\\assets"
Simplifying the path before creating the dir fixes this.
- `_DEBUG` is MSVC specific so it didn't make much sense to define for
Android and iOS builds.
- iOS was the only platform to define `DEBUG`. We don't use it anywhere
outside thirdparty code, which we usually don't intend to debug, so it
seems better to be consistent with other platforms.
- Consistently define `NDEBUG` to disable assert behavior in both `release`
and `release_debug` targets. This used to be set for `release` for all
platforms, and `release_debug` for Android and iOS only.
- Due to the above, I removed the only use we made of `assert()` in Godot
code, which was only implemented for Unix anyway, should have been
`DEV_ENABLED`, and is in PoolAllocator which we don't actually use.
- The denoise and recast modules keep defining `NDEBUG` even for the `debug`
target as we don't want OIDN and Embree asserting all over the place.
The flag INSTANCE_DATA_FLAG_MULTIMESH is used for both multimesh and particles instances, this commit adds a new INSTANCE_DATA_FLAG_PARTICLES flag to discriminate between them.
This flag will also be used in the future to properly support TAA in particles.
Instead of updating all viewports, then blitting all viewports
to the backbuffer, then swapping all buffers, we run through
all viewports and render, blit, and swap backbuffer before
going to the next viewport.
Fixes include using proper depth buffer format in 3D (this had previously been fixed already but the changes were lost in a rebase), Remove unused lighting and shadowing code in 2D, and update 2D UBOs using glBufferSubData so that they remain the appropriate size.
Using this command:
```
find -name "thirdparty" -prune -o -name "*.h" -exec sed -i {} -e '/return /! s/\t\([A-Za-z0-9_]* \*[A-Za-z0-9_]*\)\;/\t\1 = nullptr;/g' \;
```
And then reviewing the changes manually to discard the ones that don't
seem correct/safe/good (notably changes to `core` unions).
This allows light sources to be specified in physical light units in addition to the regular energy multiplier. In order to avoid loss of precision at high values, brightness values are premultiplied by an exposure normalization value.
In support of Physical Light Units this PR also renames CameraEffects to CameraAttributes.
Per-light energy gives more control to the user on the final result of
volumetric fog. Specific lights can be fully excluded from volumetric fog
by setting their volumetric fog energy to 0, which improves performance
slightly. This can also be used to prevent short-lived dynamic effects
from poorly interacting with volumetric fog, as it's updated over several
frames by default unless temporal reprojection is disabled.
Volumetric fog shadows now obey Light3D's Shadow Opacity property as well.
The shadow fog fade property was removed as it had little visible impact
on the final scene's rendering.
At this time, it works best in the Vulkan Renderers as they support using multiple samplers with the same texture.
In GLES3 this feature really only allows you to use the screen texture without mipmaps if you want to save the cost of generating them.
This can be used to make shadows translucent for a specific light.
The light distance fade system also uses this to smoothly fade the shadow
when the light fade transition distance is greater than 0.
Adds a FramebufferCache singletion that operates the same way as UniformSetCache.
Allows creating framebuffers on the fly (and keep them cached if re-requested) such as:
```C++
RID fb = FramebufferCache::get_singleton()->get_cache(texture1,texture2);
```
`shader_uniform` is now consistenly used across both per-shader
and per-instance shader uniform methods. This makes methods easier
to find in the class reference when looking for them.
Mipmap LOD bias can be useful to improve the appearance of distant
textures without increasing anisotropic filtering (or in situations
where anisotropic filtering is not effective).
`fsr_mipmap_bias` was renamed to `texture_mipmap_bias` accordingly.
The property hint now allows for greater precision as well.
- Validate format conservatively. (This is to have VRS images created regardless whether VRS attachments are supported, which avoids errors in places where the code assumes such images were created on low-spec GPUs.)
- Create a non-layered default VRS image, which is what Vulkan (and D3D12, by the way) expect.
Implement built-in classes Vector4, Vector4i and Projection.
* Two versions of Vector4 (float and integer).
* A Projection class, which is a 4x4 matrix specialized in projection types.
These types have been requested for a long time, but given they were very corner case they were not added before.
Because in Godot 4, reimplementing parts of the rendering engine is now possible, access to these types (heavily used by the rendering code) becomes a necessity.
**Q**: Why Projection and not Matrix4?
**A**: Godot does not use Matrix2, Matrix3, Matrix4x3, etc. naming convention because, within the engine, these types always have a *purpose*. As such, Godot names them: Transform2D, Transform3D or Basis. In this case, this 4x4 matrix is _always_ used as a _Projection_, hence the naming.
* Moved preprocessor to Shader and ShaderInclude
* Clean up RenderingServer side
* Preprocessor is separate from parser now, but it emits tokens with include location hints.
* Improved ShaderEditor validation code
* Added include file code completion
* Added notification for all files affected by a broken include.
Now the `linuxbsd` platform can be built headlessly (e.g. without X11
development libraries).
I also cleaned up some weird (old?) usages of the `env` variable which
seem to make no difference and are used nowhere else.
- Check block decoration in addition to type decoration to be sure to find `readonly` decorators
- Verify uniforms have same writability across all shader stages in Vulkan RD
- Include Godot version and commit hash in shader cache key
- Reject files when format doesn't match, even if it's lower, since we don't have backwards compatibility here
`rendering/quality/shadows` is now `rendering/quality/positional_shadow`
to explicitly denote that the settings only affect positional light shadows,
not directional light shadows.
Shadow atlas settings now contain the word "atlas" for easier searching.
Soft shadow quality settings were renamed to contain the word "filter".
This makes the settings appear when searching for "filter" in the
project settings dialog, like in Godot 3.x.
This was done by refactoring directory and file access handling for the Android platform so that any general filesystem access type go through the Android layer.
This allows us to validate whether the access is unrestricted, or whether it falls under scoped storage and thus act appropriately.
In far most cases it seems like it's going to message about bogus manifests
in the Windows registry which point to JSON files which have since been
uninstalled, but without clearing the registry.
This happens with bogus Vulkan overlays from Twitch, Epic Online Services,
NVIDIA Nsight Systems, OBS Studio, Rockstar Games... fix your mess folks.
Fixes#56089.
- Initialize queue indices to values meaning 'unset'
- Remove unused parameters & members
- Make texture update access flags consistent with texture copy
- Fix style and pass type of some parameters
- Synchronize setup-draw in flush with a semaphore
- Add no current list validation to draw_list_begin_splits()
- Update texture usage flags on destination of copy
- Fix misuse of Vulkan flag
Clean up and do fixes to hash functions and newly introduced murmur3 hashes in #61934
* Clean up usage of murmur3
* Fixed usages of binary murmur3 on floats (this is invalid)
* Changed DJB2 to use xor (which seems to be better)
This has several benefits:
- Transparency sorting issues inherent to alpha blending no longer occur.
- Alpha hash materials can now cast shadows (also works with
GeometryInstance3D Transparency's property for alpha hash materials).
- Higher performance.
Initial TAA support based on the implementation in Spartan Engine.
Motion vectors are correctly generated for camera and mesh movement, but there is no support for other things like particles or skeleton deformations.
* Map is unnecessary and inefficient in almost every case.
* Replaced by the new HashMap.
* Renamed Map to RBMap and Set to RBSet for cases that still make sense
(order matters) but use is discouraged.
There were very few cases where replacing by HashMap was undesired because
keeping the key order was intended.
I tried to keep those (as RBMap) as much as possible, but might have missed
some. Review appreciated!
Adds a new, cleaned up, HashMap implementation.
* Uses Robin Hood Hashing (https://en.wikipedia.org/wiki/Hash_table#Robin_Hood_hashing).
* Keeps elements in a double linked list for simpler, ordered, iteration.
* Allows keeping iterators for later use in removal (Unlike Map<>, it does not do much
for performance vs keeping the key, but helps replace old code).
* Uses a more modern C++ iterator API, deprecates the old one.
* Supports custom allocator (in case there is a wish to use a paged one).
This class aims to unify all the associative template usage and replace it by this one:
* Map<> (whereas key order does not matter, which is 99% of cases)
* HashMap<>
* OrderedHashMap<>
* OAHashMap<>
fixed and simplified gl_manager_windows
swap buffers now called for all windows
fixed missing pixel format setting in additional windows
this makes them work in OpenGL contexts
changed verbose error printing to write once
this error message happens very frequently while opengl3 is not finished
removed dead code no longer needed after changes
fixed comments that were misinformation
window messages during window creation now handled
these were previously discarded
messages now tunnel the required context
changed failure to create opengl3 window on windows to be more fatal
marked a problem with pen code
conditional compilation of vulkan and opengl3 on windows fixed
windows debug builds now show messages on debug console also
rendering driver selection box now shows only compiled drivers
marked some problematic code
thanks to akien-mga for patiently rewriting my style mistakes
* Changed to use the same stages as extensions.
* Makes the initialization more coherent, helping solve problems due to lack of stages.
* Makes it easier to port between module and extension.
* removed the DRIVER initialization level (no longer needed).
Adds the is_process_running function to the native OS class and exposes it to script.
This is implemented on Windows and Unix platforms. A stub is provided for other platforms that do not support this function.
Documentation is updated to reflect new API function.
These typedefs don't save much typing compared to the full `Ref<Resource>`
and `Ref<RefCounted>`, yet they sometimes introduce confusion among
new contributors.
This method can be used to get the graphics API version currently in
use (such as Vulkan). It can be used by projects for troubleshooting
or statistical purposes.
Didn't commit all the changes where it wants to initialize a struct
with `{}`. Should be reviewed in a separate PR.
Option `IgnoreArrays` enabled for now to be conservative, can be
disabled to see if it proposes more useful changes.
Also fixed manually a handful of other missing initializations / moved
some from constructors.
Convert method signature parameters to const where it is possible
# Conflicts:
# drivers/gles3/rasterizer_canvas_gles3.cpp
# drivers/gles3/rasterizer_canvas_gles3.h
# editor/plugins/animation_state_machine_editor.cpp
# editor/plugins/animation_state_machine_editor.h
Add "generate_mipmap" font import option.
Add some missing features to the Sprite3D.
Move BiDi override code from Control to TextServer.
Add functions to access TextServer font cache textures.
Add MSDF related flags and shader to the standard material.
Change standard material cache to use HashMap instead of Vector.
This commit removes a lot of enum values related to the color render pass in favor of a new flag-bases approach. This means instead of hard-coding all the possible option combinations into enums, we can write our logic by checking a bit-mask.
The changes in rendering_device_vulkan.cpp add support for unused attachments. That means RenderingDeviceVulkan::framebuffer_create() can take null RIDs in the attachments vector, which will result in VK_ATTACHMENT_UNUSED entries in the render pass.
This is used in this same PR to establish fixed locations for the color pass attachments (only color and separate specular so far, but TAA will add motion vectors as well). This way the attachment locations in the shader can stay the same regardless of which attachments are actually used.
Right now all the combinations of flags are generated, but we will need to add a way to limit the amount of combinations in the future.
3 options are available:
- Light and Sky (default)
- Light Only (new)
- Sky Only (equivalent to `use_in_sky_only = true`)
Co-authored by: clayjohn <claynjohn@gmail.com>
This has been superseded by GDExtension so this code is no longer useful
nor usable.
There's still some GDNative-related stuff in platform export code which
needs to be adapted for GDExtension (e.g. to include GDExtension libraries
in exports).
* Changed syntax usage for RD::Uniform to create faster with a single RID
* Converted render pass setup to use this in clustered renderer to test.
This is the first step into creating a proper uniform set cache system to simplify large parts of the codebase.
This can be used to fade lights and their shadows in the distance,
similar to Decal nodes. This can bring significant performance
improvements, especially for lights with shadows enabled and when
using higher-than-default shadow quality settings.
While lights can be smoothly faded out over distance, shadows are
currently "all or nothing" since per-light shadow color is no longer
customizable in the Vulkan renderer. This may result in noticeable
pop-in when leaving the shadow cutoff distance, but depending on the
scene, it may not always be that noticeable.
This updates VMA and instead of using the custom small pool approach from 4e6c9d3ae9, lazily creates pools for the relevant memory type indices, which doesn't require patching VMA.
Also, patches already merged upstream or not needed any longer are removed.
* Adds optional vec4 USERDATA1 .. USERDATA6 to particles, allowing to store custom data.
* This data is allocated on demand, so shaders that do not use it do not cost more.
Implemented via `BCryptGenRandom` on Windows.
Implemented via `getentropy` syscall when available.
Implemented via `/dev/urandom` device as a fallback.
The `/dev/urandom` fallback can be disabled via the `NO_URANDOM` build
flag.
Note: The HTML5 version relies on emscripten file system urandom
device which itself uses the Crypto API when available or the plain
old not crypto-safe `Math.random()` otherwise.
Restore get_entropy.
Using codespell 2.2-dev from current git.
Added `misc/scripts/codespell.sh` to make it easier to run it once in a
while and update the skip and ignore lists.
16-bit shadow atlases are already the default in the project settings,
but low-level methods used 24-bit shadows by default.
This makes low-level methods more consistent with the default project
settings to avoid accidental performance issues when users change
the shadow size at run-time.
First, we should not insert into cache if the hostname resolution has
failed (as it might be a temporary internet issue), second, the async
resolver should also properly insert into cache.
Took the chance to remove some duplicate code with critical section in
it at the cost of little performance when calling the blocking
resolve_hostname function.
