Compare commits

...

961 Commits

Author SHA1 Message Date
Sungun Park
ca0f98c513 Merge branch 'rc/1.50.3' into release 2024-02-13 00:45:01 +00:00
Mathias Agopian
31b836282d fix a typo that broke the resourceallocator cache
the cache size is given in MiB not bytes, so we needed to convert it
to bytes.
2024-02-13 00:09:17 +00:00
Powei Feng
cdd9c4aebe [release] update base64 command (#7559)
Seems like a `-i` is now necessary for the command. Note that we recently startede using mac-mx machines.
2024-02-07 12:58:16 -08:00
Powei Feng
f3a61f100c [release] update base64 command (#7559)
Seems like a `-i` is now necessary for the command. Note that we recently startede using mac-mx machines.
2024-02-07 12:57:53 -08:00
Powei Feng
0774ce6b5e Bump version to 1.50.3 2024-02-06 17:40:08 +00:00
Powei Feng
60db518b75 Merge branch 'rc/1.50.2' into release 2024-02-06 17:40:08 +00:00
Powei Feng
7be9cdc7f8 Release Filament 1.50.2 2024-02-06 17:40:03 +00:00
Mathias Agopian
af792e3d18 Resource allocator cache params can be set in Engine::Config
FIXES=[323386395]
2024-02-05 10:26:26 -08:00
Ben Doherty
3c5316f1e9 Metal: schedule PresentDrawable for destruction on the main thread (#7535) 2024-02-01 10:47:37 -08:00
Ben Doherty
f93677548d Metal: schedule PresentDrawable for destruction on the main thread (#7535) 2024-02-01 10:45:58 -08:00
Mathias Agopian
1eff66e4ec wip: nullability attributes for backend 2024-02-01 10:36:42 -08:00
Mathias Agopian
34f8b9aa20 add nullability attributes to filament public APIs 2024-02-01 10:36:42 -08:00
Mathias Agopian
6d07443188 fixes builds fail when path to build directory contains spaces
fixes #7533
2024-02-01 10:35:06 -08:00
Powei Feng
1f33a6efd2 Merge branch 'rc/1.50.1' into release 2024-02-01 00:41:24 +00:00
Powei Feng
4127f619e1 Bump version to 1.50.2 2024-02-01 00:41:24 +00:00
Powei Feng
13a23703ad Release Filament 1.50.1 2024-02-01 00:41:19 +00:00
Powei Feng
57f6ca625c Update mac-continuous.yml (#7537)
ref: https://github.blog/changelog/2024-01-30-github-actions-introducing-the-new-m1-macos-runner-available-to-open-source/
2024-01-30 13:55:43 -08:00
Mathias Agopian
02f5903b67 fix external images
It turns out that many drivers require to rebind the external texture
after its content has changed, so we now always do that.

Bug #6188
2024-01-26 14:17:48 -08:00
Mathias Agopian
c39a870abc fix mobile debug builds, missing incluldes 2024-01-26 11:14:16 -08:00
Mathias Agopian
09b8008e17 add missing public headers in backend 2024-01-26 10:58:28 -08:00
Mathias Agopian
59b59cf6be add missing includes in libutils public headers 2024-01-26 10:58:00 -08:00
Mathias Agopian
0df6013263 update remote ui 2024-01-26 10:57:37 -08:00
Mathias Agopian
d5041fced7 support execinfo.h on Android 33 and above
Note: this currently requires to compile filament with API_LEVEL 33.
2024-01-25 14:50:29 -08:00
Mathias Agopian
9c54b8a777 fix typo that broke FSR upscaling 2024-01-25 10:34:47 -08:00
Mathias Agopian
8a534d0940 Fix missing/extra includes in all public headers 2024-01-24 23:13:14 -08:00
Mathias Agopian
130825422e Shadowmap texture vizualizer
The vizualizer is implemented inside filament itself and activated
using the debug registry. This intended for filament development use.
2024-01-24 11:28:11 -08:00
Powei Feng
984006ee25 geometry: allow additional attributes in TangentSpaceMesh (#7483)
- Add methods for adding attributes to the input mesh
 - Add method in TangentSpaceMesh for when user provides the
   tangents
 - Separate client-side Algorithm enum from implementation algorithm
   (AlgorithmImpl)
 - Fix CMake config for combining static libs
2024-01-23 16:16:35 -08:00
Sungun Park
b3cc4d11b8 Merge branch 'rc/1.50.0' into release 2024-01-23 21:10:19 +00:00
Sungun Park
8523f4e970 Bump version to 1.50.1 2024-01-23 21:10:19 +00:00
Sungun Park
b676002521 Release Filament 1.50.0 2024-01-23 21:10:06 +00:00
Sungun Park
6b7450dc0b Update material version to 50 2024-01-22 23:25:59 +00:00
Sungun Park
8bdf7bd1e5 Update material version to 50 2024-01-22 15:25:06 -08:00
Mathias Agopian
7b384fb5e8 remove all uses of our custom spinlock
This has caused issues and over time we have reduced the use of
spinlocks, it was only used in few places and we still have evidence
that it's causing ANRs.

We use utils::Mutex instead which is a low overhead mutex implementation
on Linux systems.

FIXES=[321101014]
2024-01-22 23:21:11 +00:00
Mathias Agopian
1262cb286c backend option to disable parallel shader compile 2024-01-22 12:43:27 -08:00
Powei Feng
20dc6d479b Merge branch 'rc/1.49.3' into release 2024-01-22 11:43:54 -08:00
Mathias Agopian
ee6f3fb1dc Add sharpening option to TAA
We use the RCAS algorithm from FSR1. This is useful for when TAA
upscaling is enabled and similar to to what FSR2 is doing.
2024-01-19 14:40:04 -08:00
Mathias Agopian
b9a9586abb taa: add a 32-sample mode and cleanup
32 sample may be more suited to 2x upsampling, it gives 8 samples per
high-res pixel (instead of 4). This is also what FSR 2.0 is using,
which is useful for comparing.
2024-01-19 14:40:04 -08:00
Mathias Agopian
6ee20a57aa remove all uses of our custom spinlock
This has caused issues and over time we have reduced the use of 
spinlocks, it was only used in few places and we still have evidence
that it's causing ANRs.

We use utils::Mutex instead which is a low overhead mutex implementation
on Linux systems.

FIXES=[321101014]
2024-01-19 12:05:07 -08:00
Mathias Agopian
368fa2bf39 fix IBLPrefilterContext::EquirectangularToCubemap
it was incorrectly mapping the equirect image to a cubemap due to a
typo in our overload of atan2 which was swapping its parameters.

atan2 is now removed, and we use atan(y,x) instead. Also modified the
code slightly so it matches almost exactly cmgen's.

FIXES=[320856413]
2024-01-18 21:10:21 -08:00
Mathias Agopian
8f642892b4 taa: fixes input filtering option=OFF when upscaling is off
when upscaling is off, the confidence value is always 1.
2024-01-18 21:07:13 -08:00
Powei Feng
0736f3c3b3 Add missing includes (#7501)
BUG=320668410
2024-01-17 11:08:58 -08:00
Powei Feng
5435a8ed3b Add missing includes (#7501)
BUG=320668410
2024-01-17 11:04:42 -08:00
Ben Doherty
5fd7a4e153 Don't render in background (#7486) 2024-01-12 10:17:00 -08:00
Sungun Park
20ff230b92 Fix a json parsing bug (#7490)
When parsing a lexeme, we use one less byte than it's intended to be for
comparing the current string.

This results in a success in cases like:
- true and truX
- false and falsX
- null and nulX
where X means an arbitrary character.

Fix this by the full intended length.
2024-01-11 12:47:11 -08:00
Ben Doherty
44ff79ad34 Metal: disable fast math (#7485) 2024-01-10 15:31:23 -08:00
Mathias Agopian
102d2db008 Bokeh aspect ratio (#7482)
* Bokeh aspect ratio

new DoF option to set the bokeh aspect ratio, this can be used to
simulate anamorphic lenses

* Update android/filament-android/src/main/java/com/google/android/filament/View.java

Co-authored-by: Powei Feng <powei@google.com>

* Update web/filament-js/filament.d.ts

Co-authored-by: Powei Feng <powei@google.com>

---------

Co-authored-by: Powei Feng <powei@google.com>
2024-01-10 15:25:28 -08:00
Powei Feng
6a7767f4e4 Bump version to 1.50.0 2024-01-10 13:57:25 -08:00
Powei Feng
5c9039e650 Release Filament 1.49.3 2024-01-10 13:55:33 -08:00
Eliza Velasquez
1203c24f06 Generate dummy stereo variants for FL0 mats
See #7415 for a more detailed description of why this change is necessary.

The remaining variants which are filtered from FL0 materials are all related to
lighting, so further hacks like this won't be necessary.

Future work involves properly supporting differing sets of variants based on
shader language.
2024-01-10 13:50:53 -08:00
Mathias Agopian
9704d27aeb TAA upscaling
This feature is still work in progress. TAA can now optionally
upscale by 4x (this disables dynamic resolution scaling).
2024-01-09 14:07:18 -08:00
Mathias Agopian
c6f2c3fc1c bloom: disable fireflies reduction when using TAA
TAA already does a fireflies reduction pass, so it's not needed when
applying bloom.
2024-01-09 14:04:01 -08:00
Mathias Agopian
2faf868341 fix missing includes (new CLion warnings) 2024-01-09 14:04:01 -08:00
Powei Feng
628d387cbd Fix typo oin MaterialCompiler (#7477) 2024-01-08 15:07:07 -08:00
Powei Feng
81f6260843 Fix typo oin MaterialCompiler (#7477) 2024-01-08 15:03:07 -08:00
Eliza Velasquez
75a1c6d7a8 Generate dummy stereo variants for FL0 mats
See #7415 for a more detailed description of why this change is necessary.

The remaining variants which are filtered from FL0 materials are all related to
lighting, so further hacks like this won't be necessary.

Future work involves properly supporting differing sets of variants based on
shader language.
2024-01-08 13:04:55 -08:00
Ben Doherty
1a9063d53a Metal: Always report material name in use-after-free detector (#7473) 2024-01-04 11:55:28 -08:00
Ben Doherty
8c76370e2d Fix ostream linking error when compiling Linux DSO (#7470) 2024-01-04 10:43:50 -08:00
Ben Doherty
6062b3c8c6 Fix ostream linking error when compiling Linux DSO (#7470) 2024-01-03 12:06:44 -08:00
Ben Doherty
d9186c44ba matdbg: Load codicon font (#7469) 2024-01-03 10:41:18 -08:00
Benjamin Doherty
bdc15a5c2d Bump version to 1.49.3 2024-01-02 14:53:07 -08:00
Benjamin Doherty
acfe9298d9 Merge branch 'rc/1.49.2' into release 2024-01-02 14:53:06 -08:00
Benjamin Doherty
6d06e00934 Release Filament 1.49.2 2024-01-02 14:52:58 -08:00
Ben Doherty
b4021a1fbc Add picking queries to gltf-viewer iOS sample (#7464) 2024-01-02 14:48:59 -08:00
Benjamin Doherty
2c9559f9b8 Add string.h header 2024-01-02 13:24:08 -08:00
Benjamin Doherty
57f6e5371b Add string.h header 2024-01-02 13:23:54 -08:00
Ben Doherty
e2da13f817 Fix a couple race conditions in Metal backend (#7457) 2023-12-26 15:33:37 -07:00
Mathias Agopian
addae314b4 on vulkan and metal resolve depth manuall with a shader
both vk and metal don't support depth resolves, and are currently
implemented in the backend. vk is buggy and they don't resolve the
same way that gl does.
2023-12-21 21:32:22 -08:00
Mathias Agopian
5e3c51822d add basic support for multisampled samplers
we don't document this feature yet because it's still WIP and 
intended to be used internally only at this point
2023-12-21 21:32:22 -08:00
Mathias Agopian
9d817a0cc9 TAA improvements (#7441)
* TAA improvements

- fix variance + AABB history clipping (we were computing the union
  of both AABB instead of the intersection).
- added a setting for the variance parameter
- added a setting for the jitter pattern
- improved some default values
- fixed a few comments
- smaller code tweaks to facilitate future improvements
2023-12-21 16:51:28 -08:00
Mathias Agopian
9fa3cbfcde add a way to set a log consumer
this is a private API to capture filament's logs.
2023-12-19 12:09:28 -08:00
Mathias Agopian
c81ece5c3c cleanup 2023-12-19 12:09:23 -08:00
princeku07
95ecc3d1e4 Make applyMaterialVariant public 2023-12-19 10:09:42 -08:00
Mathias Agopian
cabaa4323e add a way to set a log consumer
this is a private API to capture filament's logs.
2023-12-18 16:28:31 -08:00
Mathias Agopian
d6b2ec49b0 cleanup 2023-12-18 16:28:31 -08:00
Mathias Agopian
93004aa5bc add -fno-limit-debug-info to android debug builds
This is what the NDK does, it's needed to keep debug infos for the 
STL symbols (because the STL is provided by the platform and doesn't
have debug infos).
2023-12-18 16:28:06 -08:00
Mathias Agopian
27fa8ab46d don't use eglGetProcAddress on emscripten
In practice none of the extensions we would use eglGetProcAddress for
are supported. And we had a case where an extension was reported
supported but eglGetProcAddress didn't return the corresponding
entry point.

update web demos remote ui.

FIXES=[315033914]
2023-12-18 16:27:33 -08:00
Benjamin Doherty
aae48c1121 Bump version to 1.49.2 2023-12-18 13:35:38 -08:00
Benjamin Doherty
918ce935b8 Merge branch 'rc/1.49.1' into release 2023-12-18 13:35:37 -08:00
Benjamin Doherty
614990f413 Release Filament 1.49.1 2023-12-18 13:35:28 -08:00
Benjamin Doherty
6f37e07dba Bump MATERIAL_VERSION to 49 2023-12-18 09:57:31 -08:00
Benjamin Doherty
171b3279e0 Fix VulkanPlatformAndroidLinuxWindows for G3 2023-12-18 09:48:50 -08:00
Benjamin Doherty
da058028fd Fix VulkanPlatformAndroidLinuxWindows for G3 2023-12-18 09:48:20 -08:00
selcott
55c75453df Assert non-null texture in setTextureData 2023-12-15 09:22:01 -08:00
Mathias Agopian
f5eff10532 fix TAA on mobile devices.
The history neighborhood variance needed to be computed in highp.

fixes #6504
2023-12-14 13:44:15 -08:00
Ben Doherty
56a0364534 Metal: call frame scheduled callback on driver thread (#7414) 2023-12-14 11:28:29 -08:00
Mathias Agopian
c8aae3cdaa fix android build 2023-12-13 16:32:22 -08:00
Mathias Agopian
4f94c575c9 fix UiHelper
We shouldn't remove the listener callbacks when a surface is destroyed
by the system (e.g. screen off/on) because then we won't know when
it comes back.

But we still need to do this when the user calls UiHelper.detach() or
when we are attaching to a new surface.

Fixes #7424
2023-12-13 15:21:04 -08:00
Mathias Agopian
1a29f020c9 fix getUserWorldFromWorldMatrix() during the shadow passes
getUserWorldFromWorldMatrix() was always set to identity during the
shadow pass. It needs to be set the the same value as the main
camera.

FIXES=[315504607]
2023-12-13 10:58:20 -08:00
Mathias Agopian
d5847a3c6a fix per-light shadow "caster" flag update
The per-light shadow caster flag wasn't updated when a light was
toggled from casting to non-casting. This resulted in an out of date or
incorrect shadowmap to be used.

FIXES=[315859790]
2023-12-12 19:31:03 -08:00
Mathias Agopian
3dbae482d3 make a few PostProcessManager methods "package" public
we make PostProcessMaterial, getPostProcessMaterial() as well as
render(), commitAndRender() public (as in public to filament internals).

These methods have no reason to be private to PostProcessManager.
2023-12-12 11:18:52 -08:00
Mathias Agopian
27572e3be6 fix framegraph clear flags for imported targets
Imported targets always use the imported flags, not the flags that
where specified when the target was created. The clear flags
should be cleared after they've been used once in case that rendertarget
is reused by multiple passes.

For e.g. if the clear flag is set, the target will be cleared the first
time it is used, but if it's used again, we don't want to clear again, 
in that case we'll use the "local" flags used when the target was
created (as opposed to the imported flags).
2023-12-12 11:18:36 -08:00
Mathias Agopian
dedf276e26 Allow post-process materials to update the vertex position
we add a `position` field to PostProcessVertexInputs which the user
code can modify if needed.
2023-12-12 11:18:18 -08:00
Romain Guy
111ad96134 Update dependencies to latest (#7423) 2023-12-11 21:11:54 -08:00
Mathias Agopian
2a67152dda Dynamic "Material constants" support
Material constants (a.k.a: specialization constants) can only be set
during Material creation through Material::Builder.
This change somewhat relaxes that limitation by allowing constants to
be set at runtime on Material directly.

Currently this new API is still private and only supported on FMaterial.

This feature works by invalidating the HwProgram cache of the concerned
Material, causing a shader recompile per variant; so this API is costly
and should be used only for debugging or during app/game configuration.

The TAA material is modified to use constants instead of #defines for
various settings and those are exposed in TaaOptions as well is in 
ViewerGui. So with this change all aspects of the TAA material can
be changed at runtime.
2023-12-11 16:20:47 -08:00
Maximilien
2f62978e4c Changed the level 0 materials custom commands so that materials for all APIs are built.
This allows to turn off the OpenGL backend (using -DFILAMENT_SUPPORTS_OPENGL=OFF).
Previously, disabling the OpenGL backend was leading to a compilation error related to missing level 0 materials.
2023-12-11 10:40:34 -08:00
Benjamin Doherty
fdb90f2f49 Bump MATERIAL_VERSION to 49 2023-12-11 00:16:13 -08:00
Benjamin Doherty
b3a1cfe7c9 Bump version to 1.49.1 2023-12-11 00:13:01 -08:00
Benjamin Doherty
d273838e07 Merge branch 'rc/1.49.0' into release 2023-12-11 00:13:00 -08:00
Benjamin Doherty
5418b8c4cc Release Filament 1.49.0 2023-12-11 00:12:52 -08:00
Benjamin Doherty
724cab3623 Update NEW_RELEASE_NOTES.md to reflect cherry-pick 2023-12-11 00:10:29 -08:00
Benjamin Doherty
a1de8c924d Update RELEASE_NOTES 2023-12-11 00:09:20 -08:00
Benjamin Doherty
d9faea264a Fix bump-version script on Darwin 2023-12-10 19:53:12 -08:00
Benjamin Doherty
72765a5b0a Fix bump-version script on Darwin 2023-12-10 19:52:39 -08:00
Benjamin Doherty
e1beabaa98 Bump MATERIAL_VERSION to 49 2023-12-10 19:43:28 -08:00
Eliza Velasquez
ebaee14b8b Generate dummy skinning variants for FL0 mats
Even if skinning is not fully implemented on FL0, we have clients which depend
on materials with skinning variants that otherwise could easily be converted to
FL0 materials. There are two proper ways to deal with this:

1. Support skinning/morphing in Feature Level 0.

2. Allow ESSL 1.0 code and ESSL 3.0 code to support different sets of variants.

However, the simplest solution is to just include skinning/morphing variants,
but disable codegen for ESSL 1.0 code, making them identical to the base
variants. This shouldn't increase the file size much due to the dictionary
deflation. Of course, skinning will not work correctly on FL0, but this has
always been the case. Future work here would be to properly implement one of the
two solutions described above.
2023-12-10 19:19:17 -08:00
Powei Feng
d4f08dafbb vk: workaround swiftshader spirv no-op issue (#7417)
Swiftshader runs spirv validation before compilation. However,
the validation does not like having Nop (no-op) in the input.
So we skip instructions instead of writing no-op for the
output of `workaroundSpecConstant`.

Also, fix issue to keep the value in the original shader if a
specialization wasn't provided.
2023-12-08 12:05:20 -08:00
Powei Feng
753fb102c4 spirv-headers: fix failed headers check (#7416)
Due to a cmake misconfiguration, we are installing spirv as part
of the release, public headers. It's now corrected.
2023-12-08 12:05:12 -08:00
Powei Feng
b219113a55 vk: spec const workaround through runtime spirv modification (#7399)
To avoid driver edge cases with spec const, we manually change the
spec const into a regular constant in the spirv.

FIXES=310603393
2023-12-08 12:05:03 -08:00
Powei Feng
6116c2f9eb vk: workaround swiftshader spirv no-op issue (#7417)
Swiftshader runs spirv validation before compilation. However,
the validation does not like having Nop (no-op) in the input.
So we skip instructions instead of writing no-op for the
output of `workaroundSpecConstant`.

Also, fix issue to keep the value in the original shader if a
specialization wasn't provided.
2023-12-08 10:32:18 -08:00
Ben Doherty
95fdba0c5d Add camera toe-in setting for stereoscopic testing (#7411) 2023-12-08 10:03:57 -08:00
mdagois
9c56517370 Added an option to shorten compile time in Visual Studio (#7408) 2023-12-08 09:18:31 -08:00
Powei Feng
08f06f4298 spirv-headers: fix failed headers check (#7416)
Due to a cmake misconfiguration, we are installing spirv as part
of the release, public headers. It's now corrected.
2023-12-07 17:21:25 -08:00
Eliza Velasquez
b1794a82ca Generate dummy skinning variants for FL0 mats
Even if skinning is not fully implemented on FL0, we have clients which depend
on materials with skinning variants that otherwise could easily be converted to
FL0 materials. There are two proper ways to deal with this:

1. Support skinning/morphing in Feature Level 0.

2. Allow ESSL 1.0 code and ESSL 3.0 code to support different sets of variants.

However, the simplest solution is to just include skinning/morphing variants,
but disable codegen for ESSL 1.0 code, making them identical to the base
variants. This shouldn't increase the file size much due to the dictionary
deflation. Of course, skinning will not work correctly on FL0, but this has
always been the case. Future work here would be to properly implement one of the
two solutions described above.
2023-12-07 15:37:23 -08:00
Powei Feng
56bd2871ec vk: spec const workaround through runtime spirv modification (#7399)
To avoid driver edge cases with spec const, we manually change the
spec const into a regular constant in the spirv.

FIXES=310603393
2023-12-07 14:49:12 -08:00
Mathias Agopian
5a9c47bcb5 Improve blitting APIs
- deprecate blit(), renamed to blitDEPRECATED. It's only used in one
  place in copyFrame() now. We can't void it because we don't have
  access to the texture from the RenderTarget.

- add a new blit() api that works with textures instead of render
  targets.

- add a new resolve() api that works with textures instead of render
  targets. doesn't support scaling.

- always use a shader when scaling in the framegraph
  (there was only one place where we used blit)

- use the new blit() for:
	- for mipmap generation on vk (fixme)
	- copying the depth buffer to avoid ssao feedback loop

- use the new resolve() for:
    - manually resolving MSAA color/depth

- Simplify the resolve APIs on the filament side

- implement generateMipmaps for the vulkan backend

simplify MetalBlitter

- remove metal blit workarounds
- don't issue a blit from a renderpass, this only affect Renderer::copyFrame
- We only handle a single texture now (instead of color+depth), so we
  can simplify the code a lot.
  Didn't touch the "slowpath" much, but it now assumes it blits color or
  depth, not both.
2023-12-06 11:22:53 -08:00
Ben Doherty
51c2fa2728 Fix typo in getStereoscopicOptions Java (#7410) 2023-12-06 11:03:51 -08:00
Mathias Agopian
334c71de52 material depth variant caching should only happen for surface materials (#7407) 2023-12-05 14:28:32 -08:00
Mathias Agopian
001098098b on ANGLE we now use a thread pool for parallel shader compilation (#7405)
* on ANGLE we now use a thread pool for parallel shader compilation

in general we now prefer using a thread pool instead of the KHR
extension, because we have less control over how the queue is
managed by the driver.

ANGLE supports many threads very well, so we use cpu_threads/2 for the
pool size, at background priority.

* Update filament/backend/src/opengl/ShaderCompilerService.h

Co-authored-by: Ben Doherty <bendoherty@google.com>

* Update filament/backend/src/opengl/ShaderCompilerService.h

Co-authored-by: Ben Doherty <bendoherty@google.com>

---------

Co-authored-by: Ben Doherty <bendoherty@google.com>
2023-12-05 14:28:11 -08:00
Sungun Park
9140d44b29 Bump version to 1.49.0 2023-12-05 11:39:05 -08:00
Sungun Park
8b0d65768a Merge branch 'rc/1.48.0' into release 2023-12-05 11:34:11 -08:00
Sungun Park
35a00b1d84 Release Filament 1.48.0 2023-12-05 11:33:03 -08:00
Powei Feng
b517a45213 Fix sed in bump-version.sh for Linux (#7406) 2023-12-05 11:28:42 -08:00
Powei Feng
f8057b9a5b Add spirv-headers as a separate third_party repo (#7401)
* Add spirv-headers as a separate third_party repo

Previously, we pulled in spirv-headers as part of spirv-tools. We
still keep this behavior but move the content of the repo to
third_party.

We introduce a script for updating spirv-tools, and update the patch
file. The same script will also pull in a dependent spirv-headers.
2023-12-05 10:02:15 -08:00
Mathias Agopian
d9ee526a72 fix jni Scene.getEntities
we were using the version of ReleaseIntArrayElements that doesn't
always keep changes to the array.
2023-12-05 09:28:50 -08:00
Powei Feng
349bf7be38 Update MATERIAL_VERSION to 48 2023-12-01 11:01:36 -08:00
Powei Feng
164d7507bb Update MATERIAL_VERSION to 48 2023-12-01 11:00:53 -08:00
Powei Feng
a01d282f14 Add intermediate.h include to builtinResource.h (#7388) 2023-12-01 11:00:08 -08:00
Eliza Velasquez
b007e9137e Fix ESSL 1.0 external samplers
matc was re-introducing the incorrect extension for external samplers in ESSL
1.0 code after spirv optimizations.
2023-11-30 14:07:23 -08:00
Mathias Agopian
06bae26e1b better handle invalid programs in release builds (#7389)
* better handle invalid programs in release builds

Until now invalid program would basically be undefined behavior,
which in practice was a crash via a null pointer dereference.
With this change, invalid programs cause drawing ops to become no-ops.

Additionally fixed an unsynchronized access of a the variable containing
the program id. I don't think it would have caused issues though.

FIXES=[311775564]

Co-authored-by: Powei Feng <powei@google.com>
2023-11-29 09:39:35 -08:00
Rene Sepulveda
1be3e13427 Add namespace qualifiers to satisfy gcc (-fpermissive) 2023-11-29 09:38:19 -08:00
Mathias Agopian
e7c67d1adb Fix a possible NPE in UiHelper
We were not unregistering the TextureView or SurfaceHolder callbacks on
detach, so they could fire and access an null'ed RenderSurface

FIXES=[308443790]
2023-11-28 10:24:16 -08:00
Powei Feng
9b67d80961 Add intermediate.h include to builtinResource.h (#7388) 2023-11-28 10:23:31 -08:00
Powei Feng
b066678a37 matdbg: fix backend selection 2023-11-27 16:54:49 -08:00
Powei Feng
b10d5c58aa backend: Add linux as platform for tests (#7375)
- Also fixed a leak in `test_FeedbackLoops`
2023-11-27 15:48:31 -08:00
Powei Feng
3c3296a114 Bump version to 1.48.0 2023-11-27 14:22:16 -08:00
Powei Feng
61501ba122 Merge branch 'rc/1.47.0' into release 2023-11-27 14:20:27 -08:00
Powei Feng
000bff5ce1 Release Filament 1.47.0 2023-11-27 14:19:52 -08:00
Powei Feng
99ba40e965 Update MATERIAL_VERSION to 47 2023-11-27 19:40:34 +00:00
Powei Feng
47d45a64fa Update MATERIAL_VERSION to 47 2023-11-27 11:40:00 -08:00
Eliza Velasquez
4116af7971 Enable optimizations for ESSL 1.0 code
The CL introducing the ESSL 1.0 chunk in materials inadvertently disabled
optimizations for said code. This commit reintroduces those optimizations and
fixes associated bugs which manifested. In particular, spirv-cross was
generating uints for bools; this has been fixed with a hack. Additionally,
spirv-cross is now compiled with exceptions enabled so that matc can gracefully
fail and show the code which failed to compile rather than abruptly aborting.
2023-11-27 19:32:59 +00:00
Eliza Velasquez
2fab93faff Fix typo in depth_main.fs 2023-11-27 19:32:47 +00:00
Eliza Velasquez
b92c5cab07 Incorporate feedback 2023-11-27 19:32:34 +00:00
Eliza Velasquez
8ed9678cbe Update NEW_RELEASE_NOTES.md 2023-11-27 19:32:21 +00:00
Eliza Velasquez
731e52a3c1 Add option to matc to disable ESSL 1.0 codegen 2023-11-27 19:32:03 +00:00
Eliza Velasquez
687b6da800 Enable preprocessor optimization of ESSL 1.0
Since #7358 is blocked by an upstream spirv-cross issue, we can at least do a
bit of preprocessor optimization for ESSL 1.0 code in the meantime and introduce
the FILAMENT_EFFECTIVE_VERSION preprocessor definitions.
2023-11-27 19:31:52 +00:00
Eliza Velasquez
2bd48f58ff Enable optimizations for ESSL 1.0 code
The CL introducing the ESSL 1.0 chunk in materials inadvertently disabled
optimizations for said code. This commit reintroduces those optimizations and
fixes associated bugs which manifested. In particular, spirv-cross was
generating uints for bools; this has been fixed with a hack. Additionally,
spirv-cross is now compiled with exceptions enabled so that matc can gracefully
fail and show the code which failed to compile rather than abruptly aborting.
2023-11-16 14:59:01 -08:00
Mathias Agopian
f2f4f556b5 fix a lot of issues with the backend tests
With these fixes failing tests are:

OpenGL:
BackendTest.FeedbackLoops

Metal:
BackendTest.ColorResolve
BackendTest.BufferObjectUpdateWithOffset
BasicStencilBufferTest.StencilBufferMSAA

Vulkan:
Many failures still
2023-11-16 14:56:07 -08:00
Mathias Agopian
5aa15f061b fix scissor/viewport state tracking 2023-11-16 13:13:27 -08:00
Eliza Velasquez
63db4f0bf0 Fix typo in depth_main.fs 2023-11-16 12:57:54 -08:00
Eliza Velasquez
7797f5af38 Incorporate feedback 2023-11-16 12:47:46 -08:00
Eliza Velasquez
f9def09d17 Update NEW_RELEASE_NOTES.md 2023-11-16 12:47:46 -08:00
Eliza Velasquez
494934f34e Add option to matc to disable ESSL 1.0 codegen 2023-11-16 12:47:46 -08:00
Eliza Velasquez
caae42fdf6 Enable preprocessor optimization of ESSL 1.0
Since #7358 is blocked by an upstream spirv-cross issue, we can at least do a
bit of preprocessor optimization for ESSL 1.0 code in the meantime and introduce
the FILAMENT_EFFECTIVE_VERSION preprocessor definitions.
2023-11-16 12:47:46 -08:00
Benjamin Doherty
8f6c4841cb Fix LoadImage backend tests 2023-11-15 14:42:25 -08:00
Ben Doherty
64f4c097ac Remove debug code in MetalShaderCompiler (#7368) 2023-11-15 13:55:34 -08:00
Ben Doherty
5d0c06e3f2 Remove debug code in MetalShaderCompiler (#7368) 2023-11-15 13:50:59 -08:00
Ben Doherty
7d31a7f7ea Move stereoscopic debug options to Settings framework (#7347) 2023-11-15 10:51:58 -08:00
Powei Feng
0191e1fe46 Bump version to 1.47.0 2023-11-14 21:38:45 -08:00
Powei Feng
21093067db Merge branch 'rc/1.46.0' into release 2023-11-14 21:34:54 -08:00
Powei Feng
4f4a439f8b Release Filament 1.46.0 2023-11-14 21:34:31 -08:00
Eliza Velasquez
1b10e7d4f3 Add FILAMENT_ENABLE_FEATURE_LEVEL_0 option
This allows clients to selectively exclude feature level 0 support.
2023-11-14 16:52:26 -08:00
Eliza Velasquez
2306b56ea7 Add FILAMENT_ENABLE_FEATURE_LEVEL_0 option
This allows clients to selectively exclude feature level 0 support.
2023-11-14 16:50:44 -08:00
Eliza Velasquez
5b2d3ac225 Fix some ES2 issues
Fix specification of mipmaps when generating textures.

Fix oversight where emulated UBOs would not replace uniforms when swapped out
for another.
2023-11-14 16:29:23 -08:00
Antonio Maiorano
422dfcc1e6 Add missing include for latest glslang (#7365)
This change in glslang removes the include of "intermediate.h" from
GlslangToSpv.h:
62de186c33

As a result, the definition of "class TIntermediate" is removed, and
will fail compilation of MaterialCompiler.cpp when glslang is updated to
a version including the aforementioned change. We fix this by adding an
explicit include to this header in MaterialCompiler.cpp.

Co-authored-by: Powei Feng <powei@google.com>
2023-11-14 16:00:34 -08:00
Antonio Maiorano
eebb40d30d Add missing include for latest glslang (#7365)
This change in glslang removes the include of "intermediate.h" from
GlslangToSpv.h:
62de186c33

As a result, the definition of "class TIntermediate" is removed, and
will fail compilation of MaterialCompiler.cpp when glslang is updated to
a version including the aforementioned change. We fix this by adding an
explicit include to this header in MaterialCompiler.cpp.

Co-authored-by: Powei Feng <powei@google.com>
2023-11-14 15:23:39 -08:00
Powei Feng
0d304393f4 Small compilation fixes (#7363)
- gltfio: Enable -Wall -Werror for gltfio_core
 - gltfio: Fix various errors that were missed warnings
 - matdbg: switch from std::atomic_uint64_t to
   std::atomic<uint64_t> for older clang
2023-11-14 14:21:52 -08:00
Eliza Velasquez
766e4f2ec9 Fix some ES2 issues
Fix specification of mipmaps when generating textures.

Fix oversight where emulated UBOs would not replace uniforms when swapped out
for another.
2023-11-14 13:39:01 -08:00
Powei Feng
b0a584c915 Small compilation fixes (#7363)
- gltfio: Enable -Wall -Werror for gltfio_core
 - gltfio: Fix various errors that were missed warnings
 - matdbg: switch from std::atomic_uint64_t to
   std::atomic<uint64_t> for older clang
2023-11-14 13:26:48 -08:00
Mathias Agopian
57aa99e964 Fix (again) the shadow transform
- we were not using the correct field in ShadowMapManager
- we were not computing the transform correctly, it should applied
  after the local transform, not before.

FIXES=[299310624]
2023-11-13 22:29:52 -08:00
Mathias Agopian
cfc133fcf1 repair visible shadow status
FIXES=[309519433]
2023-11-13 22:29:26 -08:00
Powei Feng
8eade6be1f Revert "engine: move setFrontFaceWindingInverted from View to MaterialInstance (#7331)" (#7360)
This reverts commit 038f07cb34.
2023-11-13 22:27:56 -08:00
Powei Feng
d5bab43ceb Revert "engine: move setFrontFaceWindingInverted from View to MaterialInstance (#7331)" (#7360)
This reverts commit 038f07cb34.
2023-11-13 22:26:03 -08:00
Mathias Agopian
485ae1b324 add Material.compile and Engine.flush java bindings (#7348)
* Add Material.compile() Java binding.

* Add Engine.flush() java binding

* Add Scene.forEach java binding

update the Android gltf-viewer sample to precompile all variants of all
materials in the scene, similarly to the desktop sample.
2023-11-13 10:43:52 -08:00
Powei Feng
2250664e58 Update Material_VERSION to 46 2023-11-13 10:40:02 -08:00
Powei Feng
75e8961109 Update Material_VERSION to 45 2023-11-13 10:37:16 -08:00
Mathias Agopian
0e9b2eda0a Fix (again) the shadow transform
- we were not using the correct field in ShadowMapManager
- we were not computing the transform correctly, it should applied
  after the local transform, not before.

FIXES=[299310624]
2023-11-10 15:33:06 -08:00
Mathias Agopian
a900bc69fb don't precompile variants in gltfio
we recently added calls to Material::compile in gltfio to precompile
materials are they are discovered. that wasn't a good call, because 
this should be the responsibility of the app, not of gltfio, at least
not without an option.

This is now done in gltf_viewer. We need something similar for 
Android.

Bugs #7318, #7336
2023-11-10 15:30:44 -08:00
Ben Doherty
edbfefda0a Fix Java FeatureLevel Enum (#7344) 2023-11-10 14:41:17 -08:00
Mathias Agopian
1a0b5ddc14 fix [Pixel]BufferDescriptor functor callbacks
and attempt to make functor<->callback code less confusing.
2023-11-10 13:15:29 -08:00
Mathias Agopian
e8bed52b3f repair visible shadow status
FIXES=[309519433]
2023-11-10 08:58:48 -08:00
Ben Doherty
06e8b1d689 Support up to 4 side-by-side stereoscopic eyes (#7328) 2023-11-09 15:34:50 -08:00
Benjamin Doherty
8a27cc8b7f Bump version to 1.46.0 2023-11-09 10:11:39 -08:00
Benjamin Doherty
d786d59ea1 Merge branch 'rc/1.45.1' into release 2023-11-09 10:09:56 -08:00
Benjamin Doherty
9c351b28ab Release Filament 1.45.1 2023-11-09 10:09:36 -08:00
Ben Doherty
cbc3bb3326 Fix mix-precision quaternion conversions (#7339) 2023-11-08 12:38:57 -08:00
Ben Doherty
79116905aa Fix mix-precision quaternion conversions (#7339) 2023-11-08 12:38:23 -08:00
Benjamin Doherty
763950ec18 Revert "fix a couple shadow stability bugs"
This reverts commit 1b0db0fca2.
2023-11-08 12:33:24 -08:00
Powei Feng
0eb851ff8c Update debug.cpp (#7337) 2023-11-08 10:55:28 -08:00
Powei Feng
038f07cb34 engine: move setFrontFaceWindingInverted from View to MaterialInstance (#7331)
Moving setFrontFaceWindingInverted to MaterialInstance will enable
finer control over face inversion and aligns better with Vulkan's
pipeline definition (see VkGraphicsPipelineCreateInfo).
2023-11-06 14:06:00 -08:00
Sungun Park
62adb234d1 Cleanup RenderPassNode::declareRenderTarget (#7332)
There's no functional change. Remove an unused local variable
`outgoingEdges` to save CPU resource. Tidy up an usage of local variable
to slightly improve readability.
2023-11-06 13:22:30 -08:00
Powei Feng
53d8d2a41f vk: fix headless semaphore wait logic (#7335)
We removed the no-op queue submit for headless in PR #7264. This
means that a semaphore will not be waited on and caused a
validation error.  Here, we simply don't acquire that semaphore
for present.

Also reorganzied the pSempaphores array code for better
readability.

Fixes #7334
2023-11-03 17:14:31 -07:00
Mathias Agopian
974a69a273 update remote ui and web samples 2023-11-02 07:15:46 -07:00
Mathias Agopian
7d694ee85e repair gltf_viewer drag&drop
Drag and dropping a gltf folder was broken:
- the handle didn't find the gltf file on drag&drop
- the ResourceLoader cached the asset path
- don't exit(1) when drag&dropping an invalid file
2023-11-02 06:00:45 -07:00
Powei Feng
46e6664c3d matdbg: fix matinfo use case (#7325)
- Make sure matinfo works by selecting a default backend in the
   absence of activeShaders.
 - Add options to select backend in matinfo mode.
 - Workaround cursor misplacement for monaco
 - Refactor menu sections into a common element.
2023-11-01 21:42:42 -07:00
Eliza Velasquez
4fd7c418e5 Remove invalid FeatureLevel0Sampler3D test 2023-11-01 21:17:22 +00:00
Eliza Velasquez
a3fdca7997 Fix basic post processing on ES2
First, this commit introduces some very simple bugfixes regarding ES2
compatibility related to postprocessing.

Second, this commit adds support for creating textures specified as R8, SRGB8,
and SRGB8_A8 in ES2. R8 is trivial: just use GL_LUMINANCE instead. The sRGB
formats, however, are maybe a bit more controversial. As implemented, they
instead just use the equivalent non-sRGB formats. This is of course technically
incorrect. There are a few approaches to how to add sRGB compatibility for ES2
that I can think of.

1. Do a bunch of complex shader nonsense in matc. Maybe even traversing the AST
and ensuring any texture lookup of a texture flagged as sRGB uses some
compatibility function. This would require static analysis to track if samplers
are reassigned to another variable, for example. This of course also breaks down
if you don't know at compile time if the shader will receive an RGB or an sRGB
sampler, or if the shader should be able to support both RGB or sRGB samplers.
Really only worth mentioning here for the sake of completion.

2. You could also generate simple compatibility functions to look up each
sampler, which would only apply to FL0 materials.

First, we would have to extend the material format to be able to explicitly
"color" a sampler as sRGB or not, like:

```
parameters : [
    {
        type : sampler2d,
        name : albedo,
        precision : medium,
        colorSpace : srgb,
    },
    {
        type : sampler2d,
        name : normal,
        precision : medium,
        colorSpace : linear,
    }
],
```

Then, the following GLSL code would be generated.

```glsl
\#if __VERSION__ == 100
vec4 texture_albedo(vec2 position) {
  return sRGBtoLinear(texture2D(materialParams_albedo, position));
}
vec4 texture_normal(vec2 position) {
  return texture2D(materialParams_albedo, position);
}
\#else
vec4 texture_albedo(vec2 position) {
  return texture(materialParams_albedo, position);
}
vec4 texture_normal(vec2 normal) {
  return texture(materialParams_normal, position);
}
\#endif
```

Finally, at runtime, if a sampler is "colored" one way or the other, we would
verify that only the appropriate kinds of samplers are bound.

I'm actually very partial to this solution. Since sRGB compatibility is only a
concern on ES2, we can generate this code only for FL0 shaders, which already
require GLSL shader authors to care about ESSL 1.0 compatibility by calling the
appropriate `textureXX` functions. Additionally, it provides a layer of
high-level validation that texture lookups are correct, even if a real ES2
context is not available on the device being tested.

3. Leave it entirely up to the client. (What this commit does.) This leaves
client code ripe for making mistakes, but luckily, we can go back and do
solution 2 whenever. If specifying a color space for a sampler remains optional,
then if this feature is retrofitted in the future, client code will continue to
compile.
2023-11-01 18:23:27 +00:00
Eliza Velasquez
7f1704481e Force post-processing materials to be unlit
This both fixes compilation of FL0 ES2 materials and unsaddles post-processing
materials with a lot of requirements added on by being considered lit.
2023-11-01 18:23:27 +00:00
Eliza Velasquez
e6b962b038 Further improve feature level 0 support
Enable a limited subset of materials in PostProcessManager for FL0.

Create new function Material::getFeatureLevel() in C++ and Java.

Create missing Material::getReflectionMode() method in Java.
2023-11-01 18:23:27 +00:00
Eliza Velasquez
668fb07ac2 matc: support basic post-process materials in FL0
This change hard-codes writing the post process output at index 0 (i.e. color)
to gl_FragColor when generating ESSL 1.0 shaders. Any other outputs (besides
depth) are discarded with a warning, but as far as I can tell, no such cases yet
exist in Filament.
2023-11-01 18:23:27 +00:00
Eliza Velasquez
f64eef02a3 Miscellaneous feature level 0 fixes
Fix edge case where an empty struct could be generated in an ESSL 1.0 shader.

Include _maskThreshold and _doubleSided in ESSL 1.0 shaders.

Add GL_OES_standard_derivatives extension to ESSL 1.0 shaders. According to
gpuinfo.org, this has 96% device coverage and supports both Mali-400 and Adreno
(TM) 304.

Remove 3D sampler support from ESSL 1.0 shaders. This extension is only
supported by 62% of devices.

Change filagui material to a FL0 material.
2023-11-01 18:23:27 +00:00
Eliza Velasquez
239e98ccec Add basic Emacs support 2023-11-01 17:28:26 +00:00
Mathias Agopian
919cfae6b2 improve ShadowMapManager slightly
don't use FixedCapacityVector to store pointers to active shadowmaps,
that's just not needed. They're all stored un a static array already
and directional and spot shadows are partitioned. 

This saves a couple heap allocations as well a an pointer dereference.
2023-11-01 09:59:32 -07:00
Mathias Agopian
0b9389430b avoid to dereference nullable objects (#7322)
Add missing @NonNull annotations

FIXES=[308443790]
2023-10-31 16:12:03 -07:00
Eliza Velasquez
1d157677d1 Add proper language server support
compile_commands.json was being generated, but hidden away inside of the cmake
build directories. This change makes build.sh link it to the main project dir
and adds some associated .gitignore entries. Now compile_commands.json is
properly read when starting clangd from Emacs, for example, and probably many
other editors.
2023-10-31 22:54:38 +00:00
Powei Feng
ef3f0cb326 vk: assert updateImage is called with non-empty size (#7315)
BUG=303073160
2023-10-31 10:36:07 -07:00
Mathias Agopian
fcf53f2c3e fix SSR artifact when enabling SSR
When we enable SSR the first time, the SSR buffer is not initialized,
this can result in the color pass fragment shader aborting, which in 
turn prevents the SSR history buffer from being initialized (since 
it's made from the result of the color pass), repeating the cycle.

In some other case, the system somehow recovers but we still see a
flicker when enabling SSR.

The solution here is to disable SSR in the shader until the history
buffer is ready (i.e. a frame later).
2023-10-31 10:32:14 -07:00
Mathias Agopian
13571868de don't use a spinlock for the HandleArena
We've seen hangs/ANR that are not well understood on that spinlock, so
for now we're going back to mutexes, which, on android, are very 
efficient under low contention (no syscall).

FIXES=[308029108]
2023-10-30 15:16:01 -07:00
Mathias Agopian
0f2c89b140 improvements to EntityManagers and Filament APIs (#7302)
* prevent public classes from being created on the stack

- we used to to this by deleting operator delete, but this prevented
  the internal "F" classes from being virtual; which can be useful
  when using EntityManger::Listener.
  now we just make the destructor protected in each class.

- EntityManger::Listener now has a virtual destructor so that
  objects could be correctly destroyed from Listener*

* improve EntityManger and Component managers

- all component managers now have the same "base" API
    - getComponentCount()
	- empty()
    - getEntity()
    - getEntities()

- Scene now has getEntityCount()

- EntityManager now has getEntityCount()

- all component manager implement gc() the same way, by calling destroy()

- SingleInstanceComponentManager::gc() that calls removeComponent() has
  been removed because it's dangerous. removeComponent() is often
  not enough, some additional cleanup might be needed.
2023-10-30 15:15:00 -07:00
Mathias Agopian
f8b70e8ec5 fix a Transform component leak in CameraManager
CameraManager creates a Transform component for each Camera component
is not already present. However, it didn't destroy the transform
component when it's itself destroyed. the leaked transform component
would eventually be garbage collected, but caused significant
slow down and memory pressure. This is because camera components are
created every frame for the shadow maps.

FIXES=[303914944]
2023-10-30 15:14:52 -07:00
Mathias Agopian
7abdea5a2e improve BlobCache API and compatibility
- the insert and retrieve handlers can now be set/unset independently.
  this could be useful for debugging.

- program caching is disabled if the GL implementation doesn't support it.

- removed unused code

FIXES=[307549547]
2023-10-30 15:13:28 -07:00
Mathias Agopian
d6fda03b06 fix logging typo 2023-10-30 15:08:56 -07:00
Mathias Agopian
a76addd2bf disable multiple context support on WGL
The reason is that some implementations of WGL require all contexts to
be created on the same thread, which we're not necessarily doing here.

fixes #7078
2023-10-27 15:53:48 -07:00
Mathias Agopian
b1f7731dbe fix max lod level computation in IBLPrefilter
FIXES=[308012116]
2023-10-27 15:28:24 -07:00
Mathias Agopian
e3e12dbf73 Make sure to unbind imported textures when destroying one
fixes #7280
2023-10-27 15:28:07 -07:00
Benjamin Doherty
31a75029f0 Update RELEASE_GUIDE with npm and CocoaPods instructions 2023-10-27 16:22:47 -04:00
Mathias Agopian
e5c24cc718 Fix dangling pointer where destroying a samplergroup 2023-10-27 13:10:21 -07:00
Mathias Agopian
2b86c8df6f cleanup and better bone weight checks
- only check/log in debug builds
- use epsilon = 2e-7 * double(tempPairCount)
- compute boneWeightsSum in double
- don't modify the weights if they're within the threshold

FIXES=[306565054]
2023-10-27 11:33:03 -07:00
Mathias Agopian
4d8e6eefa1 don't use a spinlock for the HandleArena
We've seen hangs/ANR that are not well understood on that spinlock, so
for now we're going back to mutexes, which, on android, are very 
efficient under low contention (no syscall).

FIXES=[308029108]
2023-10-27 11:32:38 -07:00
Ben Doherty
8aeec2ba35 Fix iOS transparent rendering sample (#7300) 2023-10-27 13:44:28 -04:00
Eliza Velasquez
4eb4fd5aba Explicitly prevent upgrading from feature level 0 2023-10-26 22:20:57 +00:00
Eliza Velasquez
56355231bd Allow explicitly initializing at feature level 0
This change does three main things. First, it adds an option to the Engine
Builder to pick the feature level at which to instantiate Filament. The only
real practical purpose of allowing this is to be able to instantiate at feature
level 0. Secondly, it allows feature level 0 to properly work on non-ES2
devices. Thirdly, it changes both Android and desktop hellotriangle samples to
explicitly opt-in to feature level 0.

Unfortunately, feature levels are used in two different, somewhat contradictory
ways presently in Filament, which can make reasoning about this change a bit
confusing. From a client perspective, feature levels refer to buckets of
capabilities which are guaranteed to be supported. Internally, there is a
separate "feature level" stored internally at the Driver subclass level which
generally corresponds to the maximum supported feature level, but is also
referenced when activating workarounds for limited devices. For example, Uniform
Buffer Objects are not supported in ES2, however, Filament supports emulating
them such that the client does not need to care at all; a supported feature is a
supported feature. But internally, Filament uses this "Driver" feature level to
determine whether or not a given workaround is needed. There were several cases
where the "active feature level" was being examined in order to activate these
workarounds rather than the "driver feature level", which was incorrect.

Why should non-ES2-only devices want to activate feature level 0? Allowing this
behavior 1. makes feature level 0 more consistent with the behavior of other
feature levels and 2. allows clients a layer of validation that their software
will work on all devices supported by Filament if they explicitly opt into it.

Consistency: Filament guarantees that any given device which supports a given
feature level will also support running on every feature level below, except for
feature level 0. This change removes that exception.

Validation: It's not perfect, and there will likely be bugs and unexpected
differences in behavior between ES2 and non-ES2 devices that crop up in the
future between two devices running on the same feature level. However, it's at
least a basic high level layer of validation that enables more rapid testing
workflows directly via desktop versions of Filament rather than having to fiddle
with something like ANGLE to get perfect GLES 2.0 compliance. Additionally, it
expands options for automated testing (with the same caveats).

This change has been tested on both the desktop and Android versions of
hellotriangle.
2023-10-26 22:20:57 +00:00
Powei Feng
9ccb8fce31 matdbg: UI refresh (#7301) 2023-10-26 14:14:01 -07:00
Mathias Agopian
e674420e9c improvements to EntityManagers and Filament APIs (#7302)
* prevent public classes from being created on the stack

- we used to to this by deleting operator delete, but this prevented
  the internal "F" classes from being virtual; which can be useful
  when using EntityManger::Listener.
  now we just make the destructor protected in each class.

- EntityManger::Listener now has a virtual destructor so that
  objects could be correctly destroyed from Listener*

* improve EntityManger and Component managers

- all component managers now have the same "base" API
    - getComponentCount()
	- empty()
    - getEntity()
    - getEntities()

- Scene now has getEntityCount()

- EntityManager now has getEntityCount()

- all component manager implement gc() the same way, by calling destroy()

- SingleInstanceComponentManager::gc() that calls removeComponent() has
  been removed because it's dangerous. removeComponent() is often
  not enough, some additional cleanup might be needed.
2023-10-26 13:10:43 -07:00
Mathias Agopian
8a9cbcfb99 fix a Transform component leak in CameraManager
CameraManager creates a Transform component for each Camera component
is not already present. However, it didn't destroy the transform
component when it's itself destroyed. the leaked transform component
would eventually be garbage collected, but caused significant
slow down and memory pressure. This is because camera components are
created every frame for the shadow maps.

FIXES=[303914944]
2023-10-26 13:05:35 -07:00
Mathias Agopian
f0d5cd3fa1 improve BlobCache API and compatibility
- the insert and retrieve handlers can now be set/unset independently.
  this could be useful for debugging.

- program caching is disabled if the GL implementation doesn't support it.

- removed unused code

FIXES=[307549547]
2023-10-25 22:15:25 -07:00
Powei Feng
cb3e808e8d Bump version to 1.45.1 2023-10-25 15:18:59 -07:00
Powei Feng
b2e0b97bad Merge branch 'rc/1.45.0' into release 2023-10-25 15:16:19 -07:00
Powei Feng
73b0751ccf Release Filament 1.45.0 2023-10-25 15:15:35 -07:00
Powei Feng
cc95a4a7a3 vk: remove unused platform GGP (#7298) 2023-10-25 11:00:35 -07:00
Ben Doherty
d76cf643c5 Improve Metal vertex buffer bindings (#7293) 2023-10-25 12:14:02 -04:00
Powei Feng
6e249c4c1b matdbg: material info and fix resizing (#7295) 2023-10-24 21:56:42 -07:00
Mathias Agopian
9c0cbed214 OpenGLBlobCache: be more robust when shader fails to compile
- don't call BlobCache if link status false
- don't assume glGetProgramiv never fails
- don't assume malloc never fails

FIXES=[307549547]
2023-10-24 20:45:52 -07:00
Powei Feng
c531a9c077 filamat: Fix MaterialInfo::userMaterialHasCustomDepth init (#7292)
Leaving it uninitialized leads to msan failure.
2023-10-24 20:45:34 -07:00
Mathias Agopian
af0c6a7fe9 OpenGLBlobCache: be more robust when shader fails to compile
- don't call BlobCache if link status false
- don't assume glGetProgramiv never fails
- don't assume malloc never fails

FIXES=[307549547]
2023-10-24 16:42:01 -07:00
Powei Feng
7b7dfad552 filamat: Fix MaterialInfo::userMaterialHasCustomDepth init (#7292)
Leaving it uninitialized leads to msan failure.
2023-10-24 15:43:24 -07:00
Powei Feng
deb3eb0b11 matdbg: fix deadlock and add experimental UI (#7275)
- Ensure that waiting on lock times out so that we don't lock
   up a thread when the client is gone.
 - Add an experimental folder to matdbg/web/ for the new
   UI work.
2023-10-24 13:48:29 -07:00
Sungun Park
d3016adaff FFilamentAsset has root nodes' scene-mask set
The transient property `mRootNotes` in FAssetLoader is built when a new
root asset is created and referenced whenever a new instance is created.
So it incurs an undefined behavior when a previously created asset tries
creating a new instance after a newly created asset has already created
via the same asset loader.

Move this transient property to each asset so that they can reference it
when a new instance is created.

This partially fixes #7269
2023-10-23 15:30:49 -07:00
Sungun Park
6c29542fad Cleanup function signatures
There's no functional change in this commit.

Make some parameter names more legible by renaming them and put output
parameters to the right of their function.
2023-10-23 15:30:49 -07:00
Sungun Park
0d2a96d630 Remove transient property mAsset from FAssetLoader
The temporary variable has been used to store the current instance of
FFilamentAsset being loaded for easy access from internal methods.  This
causes a crash as to a complex scenario as follows.

val asset1 = assetLoader.createAsset(assetBuffer1)
val instance1 = assetLoader.createInstance(asset1)
val asset2 = assetLoader.createAsset(assetBuffer2)
val instance2 = assetLoader.createInstance(asset1)

As the first step of fixing this issue, remove the transient property
`mAsset` from FAssetLoader. This commit alone doesn't resolve the issue,
and more commits are following.

Consolidate the low level version of createInstance, which takes a
pointer to cgltf_data type, into the high level version as the latter
one uses a parameter for FFilamentAsset instead of referencing mAsset.

Update all other relevant methods to take a FFilamentAsset pointer
instead of cgltf_data.

This partially fixes #7269
2023-10-23 15:30:49 -07:00
Ben Doherty
d3fe46765f Implement Metal parallel shader compilation (#7205) 2023-10-23 17:08:09 -04:00
Mathias Agopian
892f94e3c4 attempt to repair PlatformEGLHeadLess
It had been broken for a while. Here we attempt to repair it by moving
a lot of its functionality into PlatformEGL.
2023-10-23 11:13:02 -07:00
Mathias Agopian
f75f7039f4 Add support for stenciled swapchains in EGL
Support for GLX, WGL and WebGL is still missing.

partially fixes #7232
2023-10-23 11:13:02 -07:00
Mathias Agopian
8303d6b28e EGL: fix typos in config creation
thankfully it didn't seem to cause harm.
2023-10-23 11:13:02 -07:00
Mathias Agopian
0f9a2dd6af Froxel visualization debug option
The setting can be changed at runtime using a debug property.
2023-10-23 10:02:52 -07:00
Mathias Agopian
626621fb1c minor filament benchmarks cleanup 2023-10-23 10:02:26 -07:00
Powei Feng
2a1f762e23 Update MATERIAL_VERSION to 45 2023-10-22 22:19:41 -07:00
Powei Feng
2b78fd8359 Update MATERIAL_VERSION to 45 2023-10-22 22:18:43 -07:00
Ben Doherty
4dd98e63e4 Create use-after-free detector for Metal textures (#7250) 2023-10-20 17:16:55 -04:00
Ben Doherty
9d181a172a Create use-after-free detector for Metal textures (#7250) 2023-10-20 17:15:42 -04:00
Benjamin Doherty
d4b9d1e023 Update NEW_RELEASE_NOTES.md to reflect cherry-pick 2023-10-20 17:09:50 -04:00
Ben Doherty
1b7187f427 Support stencil buffer when post-processing is disabled (#7227) 2023-10-20 17:08:30 -04:00
Powei Feng
b62991d967 vk: support stencil format in swapchain (#7277)
Fixes #7233
FIXES=302197523
2023-10-19 13:43:23 -07:00
Mathias Agopian
562ea65d5c Increase FrameGraph Arena to 256KiB
It was possible to run out of space with the Bistro scene and 
everything enabled.
2023-10-19 12:21:31 -07:00
Mathias Agopian
6498cf5b64 dynamic shadowmap visualization (#7274)
* debugging PCF mode

This mode always uses a hard PCF and takes a 
slightly slower code path.

* dynamic shadowmap visualization

The directional shadowmap visualizer is implemented behind a 
specialization constant. Add the DebugRegistry infrastructure to be
able to update the spec-constant at runtime and have a subset of 
all materials invalidated.

This allows to toggle the visualization at runtime using a debug
property.

This is also a proof of concept that we can update spec-constants
at runtime; we could probably leverage this work for engine-wide
shader configurations.

* Update main.fs

* Update filament/src/details/Material.cpp

Co-authored-by: Powei Feng <powei@google.com>

---------

Co-authored-by: Powei Feng <powei@google.com>
2023-10-19 12:18:00 -07:00
Adrian Perez
3c77d2c3f5 StructureOfArrays can push_back move-only types 2023-10-19 12:13:28 -07:00
Powei Feng
960c6170fe vk: optimize headless swapchain (#7264)
- Remove queue submit call when using headless swapchain. It was meant
    to emulate a real swapchain, but queue submits are expensive.
 - Add option to remove flush and wait when window resizes. If a
    headless platform uses this signal to refresh the swapchain, we
    don't necessarily need it to also flush and wait before the refresh.
 - Refactor VulkanPlatform customizations
2023-10-19 11:43:24 -07:00
mackong
37c2fe31d5 samples: support apply all animations in gltf_viewer 2023-10-19 08:45:25 -07:00
Mathias Agopian
21b51caf3d add Renderer::getClearOptions (#7272)
FIXES=[243846268]
2023-10-18 15:16:01 -07:00
Benjamin Doherty
0dddd94eab Fix missing SkinningBuffer include 2023-10-18 13:45:08 -07:00
Benjamin Doherty
76dbc08176 Fix missing SkinningBuffer include 2023-10-18 13:44:31 -07:00
Mathias Agopian
1b0db0fca2 fix a couple shadow stability bugs
- shadows are now stable (in stable mode) when an IBL rotation is
  used.

- fix the shadow transform option which didn't work when an IBL rotation
  was used

- also use the x-axis as a reference for the "up" direction when
  computing the light space matrix so that we don't fall into the
  degenerate case when the light points straight down, which is a
  common case

FIXES=[299310624]
2023-10-17 12:26:43 -07:00
mackong
163f02035f fix ubershader index for transmission&volume material (#7244)
Co-authored-by: Mathias Agopian <mathias@google.com>
2023-10-16 12:15:47 -07:00
Mathias Agopian
14263efbea fix mixed-precision quaternion math
We follow the same rules as C++, e.g. float * double -> double
2023-10-16 10:52:36 -07:00
Sungun Park
7c6103a458 Update BUILDING.md for the latest instruction (#7267)
- filament can be built with Visual Studio 2022 as well.
- Fix the link to the Windows SDK.
2023-10-14 00:13:54 -07:00
Powei Feng
92846305f5 matdbg: change from websocket to GET (#7263)
- Use a hanging-GET approach to reduce dependency on websockets.
 - Also add mutex to protect access to MaterialRecords, which is
   written to/read from from multiple threads.
2023-10-13 14:43:34 -07:00
Benjamin Doherty
2ef0244266 Fix build failures due to filamat lite removal 2023-10-13 11:34:47 -07:00
Powei Feng
d5ebca0c49 vk: clean up depth formats (#7262)
To prepare for allowing stencil formats in attachments.
2023-10-13 11:21:29 -07:00
Ben Doherty
38ceee8d75 Support stencil buffer when post-processing is disabled (#7227) 2023-10-13 10:04:39 -07:00
Ben Doherty
fc6744ba75 Add AgX tonemapper (#7236) 2023-10-13 09:44:29 -07:00
Ben Doherty
0774bf9501 Remove problematic GlslangToSpv option: emitNonSemanticShaderDebugInfo (#7260) 2023-10-13 09:37:56 -07:00
Ben Doherty
8f5b2fd230 Update glslang to 277d09e679f0f4d9469c463c00cb11c6a040e65f (#7261) 2023-10-13 09:37:49 -07:00
Powei Feng
078a17469a matdbg: refactor EDIT command (#7255)
The websocket code for parsing the EDIT command is pretty verbose.
Proposing that we move to a HTTP POST request instead.

Also moved the API handler code out of DebugServer.h for clarity.
2023-10-12 10:40:57 -07:00
Eliza Velasquez
0887e388db matinfo: further refactor out redundant code 2023-10-11 16:17:48 -07:00
Eliza Velasquez
e4a57cedf9 matinfo: add support for viewing ESSL1 code 2023-10-11 16:17:48 -07:00
Ben Doherty
d92bdce852 Remove problematic GlslangToSpv option: emitNonSemanticShaderDebugInfo (#7260) 2023-10-11 15:13:41 -07:00
Ben Doherty
76bf906856 Update glslang to 277d09e679f0f4d9469c463c00cb11c6a040e65f (#7261) 2023-10-11 15:01:35 -07:00
Mustafa Uzun
b5e23162df fix: CameraInfo.clipTransform typo 2023-10-11 11:02:32 -07:00
Powei Feng
eeb53606c8 doc: fix viewer page again (#7254)
- reverse the link and original relationship between
   docs/viewer/filament-viewer.js and
   web/filament-js/filament-viewer.js
 - symlink in github pages does not seem to link to outside of the
   /doc directory (it does not get pulled in during deploy).
2023-10-10 13:32:22 -07:00
Mathias Agopian
dd23f271e3 cleanup Camera code and docs
- setProjection and setLensProjection are now less special, they can
  now be entirely implemented by the user thanks to two new helper
  functions. Everything can now be done with setCustomProjection.

- fix some out-dated comments

- remove dead code

- reorder methods in Camera.h
2023-10-10 13:14:37 -07:00
Powei Feng
e78a06797c doc: pin viewer lit to a specific version (#7253)
- Pin lit to version 2.8.0 (to fix a breakage caused by new
   release).
 - Update viewer filament version to latest
 - Use symbolic link instead of having two copies of the same
   file. (Could we consider removing `filament-viewer.js` in
   `web/filament-js/` ?)
 - Update `web/filament-js/README.md`
2023-10-10 13:06:24 -07:00
Ben Doherty
6279613b79 Metal: support float16 operations in --optimize-size mode (#7249) 2023-10-09 14:15:07 -07:00
Powei Feng
78d433cafa Update NEW_RELEASE_NOTES.md 2023-10-05 22:43:39 -07:00
Mathias Agopian
6590f62052 handle more generic projections for shadowFar
When we update the Far plane in the projection matrix, we assumed the
shape of the matrix. This fell appart when the projection matrix was
(for instance) a blend between an ortho and perspective projection.

We now do this more generally, that is, with less assumptions on the
projection matrix shape.
2023-10-05 16:31:39 -07:00
Jacob Su
f1f7aeb14f fix java Engine.Builder method access modifiers. 2023-10-05 14:21:58 -07:00
Mathias Agopian
5190b03f89 the visibility type can be 8 bits instead of 16.
It was changed to 16 a while back to handle more shadows, but since
then we changed the culling algorithm and 8 bits is enough again.
2023-10-05 14:20:48 -07:00
Powei Feng
2cd492ed38 Fix build failures due to filamat lite removal try 4 (#7228) 2023-10-04 10:25:15 -07:00
Mathias Agopian
07975868fe update cgltf to latest v1.13
FIXES=[239321615]
2023-10-04 10:19:39 -07:00
Powei Feng
5ab526cbb0 Fix build failures due to filamat lite removal try 3 (#7226) 2023-10-03 15:23:03 -07:00
Powei Feng
df935b75e5 engine: add job system thread count configuration (#7223)
- Plumb Engine::Config in Java
 - Add Engine::Builder for Java
 - Add jobSystemThreadCount to Engine::Config

BUG=303129581
2023-10-03 15:22:46 -07:00
Powei Feng
b4d6f975c1 Fix build failures due to filamat lite removal #2 (#7224) 2023-10-03 12:39:50 -07:00
Powei Feng
2cf86454bb vk: refactor VulkanProgram (#7221) 2023-10-03 10:49:45 -07:00
Benjamin Doherty
5572097fb3 Fix build failures due to filamat lite removal 2023-10-03 09:07:37 -07:00
Benjamin Doherty
274191036f Bump version to 1.45.0 2023-10-02 15:29:31 -07:00
Benjamin Doherty
13afbc2876 Merge branch 'rc/1.44.0' into release 2023-10-02 15:27:51 -07:00
Benjamin Doherty
480cfb2820 Release Filament 1.44.0 2023-10-02 15:27:29 -07:00
Mathias Agopian
39d555d115 get rid of filmat-lite, which was poorly maintained 2023-10-02 14:18:25 -07:00
Powei Feng
df50f4c25d vk: layout transition clean up (#7217)
- Return the correct SubresourceRange for depth attachments
 - Fix transition for when one layout within mulitple mip-levels
   is different
 - Use implicit layout transition for renderpasses
 - Fix access mask for sampler in vertex shaders
 - Use unordered_map for VkSubresourceRange in VulkanTexture
2023-10-02 10:54:44 -07:00
Eliza Velasquez
a7bb0a60fb Revert "Remove now-redundant feature level 0 materials"
This reverts most of commit 9a6b8bf24e. The hello
triangle sample remains unreverted.

The original commit inadvertently broke screen space reflections, and perhaps
other features when the default material was used. The source of the issue is
that MaterialBuilder.cpp (correctly) filters out variants that aren't supported
in feature level 0 materials, including screen space reflections.

Unfortunately, while the "feature level 0 compatibility" feature itself was
intended to make creating duplicate materials like this redundant in client
code, unfortunately, it seems the best solution for resolving this issue is to
simply keep these redundant materials in the core.

To elaborate: clients should expect that feature level 0 materials that they
create work on /all/ feature levels /exactly/ or /close to exactly/ identically.
This includes restricting more advanced features that theoretically could be
available on a higher feature level, like SSR. It's already true that if a user
would like to optionally opt-in to a more advanced material which takes
advantage of more advanced features, they would have to maintain two separate
versions of that material: one for feature level 3 and one for feature level 1.
It should be no different in this case.

However, the materials built into the engine core are an exception to this
expectation. Given that feature level 0 was tacked on after the fact with fewer
features, there must /by necessity/ have been a new material introduced for both
the default material and the default skybox specifically for feature level 0
with fewer features than extant client apps expected to be included by default.
I imagine if filament were to be rebuilt from the ground up, this exception
wouldn't exist. However, the end result is this somewhat messy redundancy.
2023-09-29 16:49:40 -07:00
Eliza Velasquez
976f304ee7 Revert "Remove now-redundant feature level 0 materials"
This reverts most of commit 9a6b8bf24e. The hello
triangle sample remains unreverted.

The original commit inadvertently broke screen space reflections, and perhaps
other features when the default material was used. The source of the issue is
that MaterialBuilder.cpp (correctly) filters out variants that aren't supported
in feature level 0 materials, including screen space reflections.

Unfortunately, while the "feature level 0 compatibility" feature itself was
intended to make creating duplicate materials like this redundant in client
code, unfortunately, it seems the best solution for resolving this issue is to
simply keep these redundant materials in the core.

To elaborate: clients should expect that feature level 0 materials that they
create work on /all/ feature levels /exactly/ or /close to exactly/ identically.
This includes restricting more advanced features that theoretically could be
available on a higher feature level, like SSR. It's already true that if a user
would like to optionally opt-in to a more advanced material which takes
advantage of more advanced features, they would have to maintain two separate
versions of that material: one for feature level 3 and one for feature level 1.
It should be no different in this case.

However, the materials built into the engine core are an exception to this
expectation. Given that feature level 0 was tacked on after the fact with fewer
features, there must /by necessity/ have been a new material introduced for both
the default material and the default skybox specifically for feature level 0
with fewer features than extant client apps expected to be included by default.
I imagine if filament were to be rebuilt from the ground up, this exception
wouldn't exist. However, the end result is this somewhat messy redundancy.
2023-09-29 23:48:47 +00:00
Mathias Agopian
21ea99a1d9 fade shadows out at shadowFar distance
Instead of a hard cutoff, we fade shadows out at
the shadowFar distance if active, fading occurs
over about 10% of the shadowFar distance.

- this works only for the directional shadow
  (other lights don't use shadowFar).
2023-09-28 16:21:27 -07:00
Mathias Agopian
f0b0505207 fix materials that write to the depth or use discard
these materials would not generate proper structure or shadow buffer,
because they used a special variant that in most case removed the
user code.

now when the user code writes the depth or calls discard, the user
shader is kept.
2023-09-28 16:19:17 -07:00
Mathias Agopian
11997d007a cleanup ASTHelpers, GLSLTools
- fix IDE warnings
- rename ASTUtils to ASTHelpers to match filename
- break dependency of ASTHelper on GLSLTools
- break dependency of GLSLTools on MaterialInfo
2023-09-28 16:19:17 -07:00
Powei Feng
0650b13358 vk: remove uneeded log in readpixels 2023-09-28 15:27:17 -07:00
Powei Feng
6cd851e77e Update MATERIAL_VERSION to 44 2023-09-28 15:27:06 -07:00
Powei Feng
c0327c95d4 vk: remove uneeded log in readpixels 2023-09-28 15:24:11 -07:00
Romain Guy
e1dfea0f12 Fix masked materials (#7215) 2023-09-27 21:36:28 -07:00
Powei Feng
b2cef9bfee Update MATERIAL_VERSION to 44 2023-09-27 16:32:17 -07:00
Powei Feng
f6b979c8e4 vk: refactor VulkanResources.h (#7206)
- Make the size of FixedSizedVulkanResources declarable
- Rename `acquire`
2023-09-27 15:01:33 -07:00
Powei Feng
87a8cb3872 Bump version to 1.44.0 2023-09-27 11:30:16 -07:00
Powei Feng
a4869eaf19 Merge branch 'rc/1.43.1' into release 2023-09-27 11:11:53 -07:00
Powei Feng
2e9dd03222 Release Filament 1.43.1 2023-09-27 11:10:42 -07:00
Eliza Velasquez
9a6b8bf24e Remove now-redundant feature level 0 materials 2023-09-26 22:46:51 +00:00
Eliza Velasquez
3aee9402d5 Add ESSL1 block to feature level 0 mats
Previously, when a material was expilicitly built for feature level 0, it was
necessary to write ESSL 1.0 code which was incompatible with the OpenGL feature
level 1 implementation in Filament. Rather than adjust the Filament
implementation so that feature level 1 uses the same workarounds as feature
level 0 to emulate GLES 3.0+ features, it's more runtime-efficient to have matc
embed ESSL 1.0 and 3.0 shaders as separate chunks and load the corresponding
one based on the active feature level.

Feature level 0 material shaders must still be written in what is effectively
ESSL 1.0. To assist with this, a small set of compatibility definitions are
introduced when building the ESSL 3.0 variant. These are virtually all
textureXXX() functions, and end up either optimized out or inlined by glslang.

One final effect of this change is that external and 3D samplers are now
properly supported in feature level 0 materials.
2023-09-26 22:46:51 +00:00
Eliza Velasquez
fcf5866fc3 Add .editorconfig file
This informs a broad variety of text editors of some of the formatting
conventions of the filament project. It may be possible to replicate some of the
more specific formatting conventions like brace placement in a future commit.
2023-09-26 22:46:51 +00:00
Powei Feng
864a854c13 vk: remove unnecessary flush and gc (#7207) 2023-09-26 14:05:36 -07:00
Jacob Su
5a18cdf0f6 Fix: remove duplicated links inside ios samples. (#7190) 2023-09-25 14:19:00 -07:00
Jacob Su
6ebd22f474 Fix test_gltfio cmake duplicate libs and failed UT. (#7199)
* fix test-gltfio cmake duplicated libs links.

* fix test_gltfio failed UT.
2023-09-25 12:52:46 -07:00
Ben Doherty
9f5c1bb0dc Move CallbackManager out of OpenGL driver (#7194) 2023-09-25 11:07:43 -07:00
Powei Feng
1740220f6a Update MATERIAL_VERSION to 43 2023-09-25 10:10:25 -07:00
Powei Feng
9cbf741ac2 Update MATERIAL_VERSION to 43 2023-09-25 10:08:56 -07:00
Mathias Agopian
04c8845284 skinning: fix a possible crash/assert
all textures declared in a shader must be bound, in the case of the
new skinning API, if less than 4 bone weights are used the texture is 
not needed, but it must still be bound.
2023-09-24 14:20:33 -07:00
Mathias Agopian
d82b7a529c skinning: minor cleanup 2023-09-24 14:20:33 -07:00
Mathias Agopian
72c60b35d7 fix TAA when camera is far from the origin
The intermediate transforms needed to be kept in double precision.
2023-09-22 10:10:18 -07:00
fvbj
5d30435620 Extend skinning for >4 bones per vertex (#6772)
* Add skinning and morphing samples to check functionality
* Implement skinning for more than four bones pair vertex

The API allows defining an unlimited number of bone indices and weights of primitives. Data is defined in building process of the renderable manager. Backward compatibility with the original solution.
Skinning of vertices is calculated on GPU, data is transferred to the vertex shader in the texture.
2023-09-21 13:04:12 -07:00
Mathias Agopian
d02231dce9 fix build most recent emscripten build 2023-09-21 13:02:35 -07:00
Powei Feng
ea70fe540f vulkan: cache vertex buffer info (#7181)
To reduce CPU on draw(), we move the VertexBuffer related metadata
out of draw into a cache. We store the cache info outside of the
actual VulkanVertexBuffer class since VulkanVertexBuffer subclass
HwVertexBuffer, which is a Handle meant to be minimal.

This  means that we need to cache on the heap, but it should be ok
since the caching is only for scene set-up and not per-frame.
2023-09-20 12:19:43 -07:00
Romain Guy
cb9bfaeb83 Fix build warnings (#7189)
The latest macOS toolchain triggers warnings for duplicate libraries at
link time. This is caused by our dependency chains.

Also remove an inlining warning in Kotlin and unnecessary warnings in
build.sh when doing a clean or generating web docs.
2023-09-20 10:16:55 -07:00
Mathias Agopian
3dcdb202d1 new API to return a Material's supported variants
FIXES=[297456590]
2023-09-19 15:48:39 -07:00
Mathias Agopian
da73f89a23 improve/simplify "Focus on shadow casters" algorithm
Functionally this shouldn't be too different, but we have some
improvements:
- better detection of "no shadows" cases
- more computations done in light-space, which should result in
  better light frustum.
- spit the code into several static functions
2023-09-19 14:15:52 -07:00
Mathias Agopian
9474798c75 remove anamorphic bloom feature
This features didn't work well, had a lot of artifacts and generally
wasn't very useful. This kind of effect should be accomplished
differently.

This is an API break because BloomOptions::anamorphism has been removed.
2023-09-18 16:09:21 -07:00
Ben Doherty
13df6652ae Add npm deploy GitHub workflow (#7184) 2023-09-18 15:11:58 -07:00
Ben Doherty
4f925f07e8 Add CocoaPods Deploy GitHub workflow (#7176) 2023-09-18 14:39:20 -07:00
Benjamin Doherty
5deccffdce Bump version to 1.43.1 2023-09-18 14:23:12 -07:00
Benjamin Doherty
a76eacba67 Merge branch 'rc/1.43.0' into release 2023-09-18 14:22:06 -07:00
Benjamin Doherty
8f64756721 Release Filament 1.43.0 2023-09-18 14:21:43 -07:00
Benjamin Doherty
ee31ca6fc0 Fix RELEASE_NOTES version 2023-09-18 14:20:30 -07:00
Mathias Agopian
9f62dc2f2f Improve bloom
- add a quality option
- remove the ping-pong code, we'll disable for GPU that don't work
  instead.
- improve quality by doing a better first downscale
  (using a 5x5 gaussian).
- improve performance by using a 9 tap filter instead of 13 for
  in most cases
- fix usages of setMinMaxLevels as it resets the base level to "min"
2023-09-18 11:51:51 -07:00
Benjamin Doherty
7dd6686087 Correct version to 1.43.0 2023-09-18 11:07:05 -07:00
Mathias Agopian
9c0fb67b87 repair "focus on shadow casters" features
This broken recently because of a typo in refactoring.
2023-09-15 14:32:56 -07:00
Ben Doherty
0b97de425f Fix, Metal setMinMaxLevel called between render passes (#7175) 2023-09-15 11:48:52 -07:00
Powei Feng
373c5710b1 Fix the output of Windows release workflow (#7171)
This probably broke due to a change from old windows command prompt to powershell. The way to refer to env variables in powershell is `$Env:var`
2023-09-14 21:23:02 -07:00
Powei Feng
fdffd93949 vulkan: fix fence deadlock (#7173)
Should only reset fences in VulkanCommands::gc
2023-09-14 17:11:49 -07:00
Powei Feng
91b55bdea1 vulkan: fix fence deadlock (#7173)
Should only reset fences in VulkanCommands::gc
2023-09-14 17:09:33 -07:00
Powei Feng
976b8c6ae4 vulkan: refactor debug defines (#7170)
Add more fine grain debug defines and add some systrace markers.
2023-09-14 15:30:42 -07:00
Powei Feng
f0a0a9b2e1 Bump version to 1.42.3 2023-09-13 23:05:11 -07:00
Powei Feng
e743e9243f Merge branch 'rc/1.42.2' into release 2023-09-13 22:59:52 -07:00
Powei Feng
c80fbfdf17 Release Filament 1.42.2 2023-09-13 22:59:33 -07:00
Powei Feng
d4d03e4a35 vulkan: reset fences instead create/destroy (#7169)
We allocate all the fences beforehand to reduce calls to
vkCreateFence.

Also remove blocking code in `getFenceStatus` since there is
not a usecase that would require that.
2023-09-13 21:17:31 -07:00
Ben Doherty
2114995489 Implement setMinMaxLevels for Metal (#7158) 2023-09-13 12:33:10 -07:00
Powei Feng
ec30ddd2fb vulkan: implicitly free command buffers (#7167)
We were calling vkFreeCommandBuffers directly, but resetting
the buffers implicitly (when vkBeginCommandBuffer is called)
seems to be a lot more performant.

Also, cleaned up destructor for VkBuffer to no longer require
a separate terminate() method.
2023-09-13 11:21:04 -07:00
Mathias Agopian
c35a60808c repair ShadowOptions::shadowFar
the shadow far plane (shadowFar) was only partially taken into account
2023-09-13 10:29:05 -07:00
Eliza Velasquez
58017a0e6a Tweak documentation
This is admittedly a very nitpicky change.

For most of the changes, I went through the various Markdown files and added
language names to the source blocks for better syntax highlighting on GitHub. It
also makes it easier to copy and paste commands without copying the leading `$`.
I avoided changing anything in `third_party`.

Additionally, I added some instructions for compiling the Android samples on the
command line and fixed some typos.
2023-09-12 20:28:46 +00:00
mackong
d9c2893976 Fix possible change of scale sign when decomposing matrix (#7138)
Co-authored-by: Mathias Agopian <mathias@google.com>
2023-09-12 12:28:02 -07:00
Mathias Agopian
1ff0a2dd6d Improvements to shadowing
- use the geometric normal to apply the shadow bias. This affects
  cascades > 0 and spot/point lights.

- use the scene's origin as a reference point for stabilizing the
  shadowmap, this is more robust.

- clamp directional shadowmap correctly to the 1-texel border, which
  needs to be reachable, as it is a valid value.

- don't snap the shadowmap to texel boundaries if stable mode is not
  active (before we only didn't do it based on lispsm). Stable mode can
  make the shadow unstable when both the camera and the scene move 
  together, so it's better to have a more predictable API where
  "stable" mode means that the snapping occurs and doesn't otherwise.

- add "far origin" distance slider to the debug ui

FIXES=[299310624]
2023-09-12 10:14:21 -07:00
Mathias Agopian
da96b45827 apply shadow settings to all lights in the scene
this "fixes" point light shadows in gltf_viewer, which were never
enabled
2023-09-12 10:08:07 -07:00
Powei Feng
8e13d53b1e vulkan: fix swapchain leak (#7161) 2023-09-11 23:23:04 -07:00
Powei Feng
9ac4fa63ed vulkan: fix swapchain leak (#7161) 2023-09-11 23:21:52 -07:00
Ben Doherty
ff274c8387 Fix incorrect SamplerParams operators (#7150) 2023-09-11 22:47:25 -07:00
Jacob Su
8d621561f3 fix ios samples missing functional header (#7153) (#7154)
1. ios sample: ios/samples/hello-gltf;
2. ios sample: ios/samples/hello-pbr;
2023-09-11 16:25:06 -07:00
Ben Doherty
2c63a5ad7a Fix use of -Wno-deprecated-register flag and MSVC (#7156) 2023-09-11 13:40:02 -07:00
Ben Doherty
d328b64dce Only call .put if the tick op was canceled (#7152) 2023-09-11 12:58:29 -07:00
Ben Doherty
3ab7780c38 Fix incorrect SamplerParams operators (#7150) 2023-09-08 13:06:52 -07:00
Mathias Agopian
6fd5d45295 fix shadow stability when eye is far from world origin
in stable mode the scale was ever so slightly varying with the 
camera position, because it was calculated from the camera frustum in
world-space, this variation was amplified when the camera is far from 
the origin, which eventually caused the modulo needed for snapping the
shadowmap projection to widely vary, leading to the instability.

We now calculate the camera frustum sphere in view space, which is
guaranteed to be constant. If "shadow caster mode" is chosen, we 
quantize the scale a little bit so it stays constant.

The snapping code itself has been cleaned.
2023-09-08 11:19:19 -07:00
Benjamin Doherty
e08e1c209b Bump version to 1.42.2 2023-09-07 17:59:26 -07:00
Benjamin Doherty
de310ed9ad Merge branch 'rc/1.42.1' into release 2023-09-07 17:55:24 -07:00
Benjamin Doherty
4f11a02889 Release Filament 1.42.1 2023-09-07 17:54:47 -07:00
Ben Doherty
a57225cf2e Fix race condition in ShaderCompilerService (#7147) 2023-09-07 17:46:45 -07:00
Romain Guy
dc818abadc Quiet popd/pushd 2023-09-07 14:48:11 -07:00
Jacob Su
c808d8f260 More robust validation in build.sh 2023-09-07 09:42:48 -07:00
Mathias Agopian
e0a37de1ba cleanup 2023-09-07 00:09:02 -07:00
Mathias Agopian
4e2a791e81 Directional shadows can now be transformed
This setting only exists for artistic reasons.

b/297095805
2023-09-07 00:09:02 -07:00
Romain Guy
60627f836f Update cmgen README 2023-09-06 09:07:40 -07:00
Romain Guy
22c9d350c1 Fix docs 2023-09-06 08:45:45 -07:00
Ben Doherty
d11a782c3e Set CMAKE_OSX_DEPLOYMENT_TARGET in CMakeLists.txt as opposed to build.sh (#7143) 2023-09-05 17:46:19 -07:00
Powei Feng
b4a45f126a vulkan: fix ordering when using debugMarkerExt (#7141) 2023-09-05 16:23:09 -07:00
Mathias Agopian
35a8e73675 remove Program operator=
It is currently not needed, AFAIK, and it was not implemented properly.
2023-09-05 12:19:37 -07:00
Jacob Su
8d7743ce67 Bug: fix underscores literal operators c++ compatible error. (#7136) 2023-09-05 10:03:51 -07:00
mackong
682585be4a Fix TextureFlags for sheenRoughnessMap (#7139) 2023-09-05 10:00:12 -07:00
Romain Guy
9046d04de4 Fix a crash in gltfio when not using ubershaders (#7135) 2023-09-02 12:45:58 -07:00
Powei Feng
d01b29fa01 vulkan: properly set bool spec onst (#7125)
We wrote a bool directly into 4 bytes (as the first byte). This has two issues:
 - the other 3 bytes are not initialized
 - should be writing VK_TRUE/FALSE instead
2023-09-01 14:17:18 -07:00
Mathias Agopian
8cba3d4366 Revert "workaround another PowerVR compiler bug "
This reverts commit 58f96be2c4.

This caused material files to increase in size significantly. It turns
out that glslang has to generate a copy for each parameter that is
passed to a function as a non-const parameter.


This revert will break IMG devices again, but that should be the case
only on debug builds. Release builds lose the const qualifier by 
virtue of going through spirv. We'll try to address this some other 
way later.
2023-09-01 14:16:59 -07:00
mackong
0f21c6c90b Use flatmat for mat parameters (#7131)
Co-authored-by: Romain Guy <romainguy@curious-creature.com>
2023-09-01 12:39:07 -07:00
Powei Feng
cc6a6454d7 vulkan: properly set bool spec onst (#7125)
We wrote a bool directly into 4 bytes (as the first byte). This has two issues:
 - the other 3 bytes are not initialized
 - should be writing VK_TRUE/FALSE instead
2023-09-01 10:59:30 -07:00
Romain Guy
609dc97c4a Remove warnings 2023-08-31 11:58:06 -07:00
Ben Doherty
be1e51ad91 Update FrameCompletedCallback using directive (#7128) 2023-08-30 16:29:28 -07:00
Ben Doherty
29ce1cad84 Transition setFrameCompletedCallback to take a CallbackHandler (#7103) 2023-08-30 16:28:53 -07:00
Ben Doherty
6a967ad007 Make destroyFence asynchronous (#7127) 2023-08-30 16:28:36 -07:00
Powei Feng
fe2bb3d9a4 Merge branch 'rc/1.42.0' into release 2023-08-28 13:26:35 -07:00
Mathias Agopian
cd7973bdcf Revert "workaround another PowerVR compiler bug "
This reverts commit 58f96be2c4.

This caused material files to increase in size significantly. It turns
out that glslang has to generate a copy for each parameter that is
passed to a function as a non-const parameter.


This revert will break IMG devices again, but that should be the case
only on debug builds. Release builds lose the const qualifier by 
virtue of going through spirv. We'll try to address this some other 
way later.
2023-08-28 10:43:26 -07:00
Powei Feng
e6384e0e92 Bump version to 1.42.1 2023-08-22 13:41:42 -07:00
Powei Feng
e63dc17f54 Fix missing createFence (#7076)
Continuing from #7072
2023-08-21 10:59:42 -07:00
Powei Feng
af338cf2ec Update MaterialEnums.h (#7098) 2023-08-21 10:52:25 -07:00
Benjamin Doherty
f1d8a04337 Bump version to 1.42.0 2023-08-15 17:13:50 -07:00
Benjamin Doherty
b13497e2a0 Merge branch 'rc/1.41.0' into release 2023-08-15 17:12:01 -07:00
Powei Feng
175c9f9966 Bump version to 1.41.0 2023-08-09 10:43:02 -07:00
Powei Feng
1a50420b46 Merge branch 'rc/1.40.5' into release 2023-08-09 10:32:50 -07:00
Powei Feng
bbad75a012 vulkan: fix fence initialization (#7038)
Previously, we have a VulkanSync with a default constructor that
allows us to have sync objects that returns error when
actual fences are not yet present.  We need to replicate that
with VulkanFence since sync objects have been removed from the
API.

Fixes #7034
2023-08-07 10:17:15 -07:00
Powei Feng
25c08f19e3 vulkan: fix TSAN in readpixels (#7023) (#7028) 2023-08-02 11:00:48 -07:00
Benjamin Doherty
59063fb7b4 Bump version to 1.40.5 2023-08-01 15:39:35 -07:00
Benjamin Doherty
71f60de0ad Merge branch 'rc/1.40.4' into release 2023-08-01 15:38:37 -07:00
Powei Feng
743661109d vulkan: fix TSAN in readpixels 2023-08-01 15:35:06 -07:00
Benjamin Doherty
5014cbb023 Bump version to 1.40.4 2023-07-26 12:53:47 -07:00
Benjamin Doherty
4142e7a1cf Merge branch 'rc/1.40.3' into release 2023-07-26 12:52:12 -07:00
Benjamin Doherty
a721e648b7 Bump version to 1.40.3 2023-07-17 15:50:20 -06:00
Benjamin Doherty
358e89ef08 Merge branch 'rc/1.40.2' into release 2023-07-17 15:49:29 -06:00
Benjamin Doherty
e023e90e7a Bump version to 1.40.2 2023-07-12 13:44:52 -07:00
Benjamin Doherty
a5b5b0c3a7 Merge branch 'rc/1.40.1' into release 2023-07-12 13:43:33 -07:00
Powei Feng
b7b4d3c295 Update Material version to match 1.40.1 2023-07-10 14:27:51 -07:00
Benjamin Doherty
a445c4e156 Merge branch 'rc/1.40.0' into release 2023-06-26 13:29:06 +08:00
Benjamin Doherty
1432c59499 Add missing atomic header 2023-06-26 13:23:38 +08:00
Benjamin Doherty
4c7c10fad0 Bump version to 1.40.1 2023-06-26 13:17:45 +08:00
Benjamin Doherty
7c8a0d1967 Bump MATERIAL_VERSION to 40 2023-06-26 10:33:02 +08:00
Powei Feng
11fbacea20 Update MaterialEnums.h to v39 2023-06-20 11:57:42 -07:00
Powei Feng
8cd70454c3 Update MaterialEnums.h to v39 2023-06-20 11:56:40 -07:00
Benjamin Doherty
1d988182fc Bump version to 1.40.0 2023-06-20 23:55:00 +08:00
Benjamin Doherty
dba49f00df Merge branch 'rc/1.39.0' into release 2023-06-20 23:45:04 +08:00
Ben Doherty
89c0b44da9 Re-enable Metal half conversion, only register SimplificationPasses for Metal (#6883) 2023-06-20 14:27:44 +08:00
Ben Doherty
4f450fd5c4 Temporarily disable Metal relaxed to half pass due to spirv-cross bug (#6880) 2023-06-20 14:27:37 +08:00
Mathias Agopian
f0943cfca2 align uniform buffer to 16 bytes
this is needed on armv7 because we use alignas to get strcture-alignment,
but that also implies (to the compiler) that the structure itself
is aligned properly.
2023-06-20 14:23:47 +08:00
Mathias Agopian
9aa52b79d4 helpers to fix jank when resizing a TextureView (#6889)
There was two related issues:
- we need to "latch" the new TextureView size when its resized. That
  can only be done by recreating the EGLSurface (i.e. recreating the
  SwapChain). UiHelper now calls onNativeWindowChanged in the case of
  the TextureView resize, so clients can recreate their SwapChain.
- we also needed to make sure that all current filament frames have
  finished to render (i.e. the last eglSwapBuffers has been called) so
  that they don't pick-up a new size (this happens after
  eglSwapBuffers) that doesn't match the viewport.

Fixes b/282220665
2023-06-12 23:13:48 -07:00
Powei Feng
a82125dbbd vulkan: minor fixes (#6874)
- flush() and wait() before destroying a swapchain
 - Make sure the debug marker extension is enabled under correct
   circumstances.
 - Change shared_ptrs to unique_ptrs and raw pointers.
 - Rename most teardown methods to terminate()
2023-06-07 11:45:09 -07:00
Benjamin Doherty
24fcb299b5 Bump version to 1.39.0 2023-06-07 11:32:33 -05:00
Benjamin Doherty
753aa9ca61 Merge branch 'rc/1.38.0' into release 2023-06-07 11:31:21 -05:00
Benjamin Doherty
a33eada7ec Bump version to 1.38.0 2023-06-07 11:29:02 -05:00
Benjamin Doherty
676a2be874 Correct MATERIAL_VERSION to 37 2023-06-05 17:20:18 -05:00
Powei Feng
c95d466cdd Fix missing include (#6858) 2023-06-01 16:28:05 -07:00
Powei Feng
3d8ac384ad Update MATERIAL_VERSION to 38 (#6852) (#6854) 2023-05-31 15:00:34 -07:00
Powei Feng
57424cc7e9 Merge branch 'rc/1.37.0' into release 2023-05-30 16:22:31 -07:00
Powei Feng
78fe4ba547 vulkan: allow for headless linux builds (#6836) 2023-05-24 17:19:42 -07:00
Benjamin Doherty
a3d25cd22b Bump version to 1.37.0 2023-05-15 16:57:27 -07:00
Benjamin Doherty
56bf841bac Merge branch 'rc/1.36.0' into release 2023-05-15 16:55:48 -07:00
Ben Doherty
dfdf0db794 Fix incorrect target passed to glBindFramebuffer (#6807) 2023-05-15 15:28:44 -07:00
Benjamin Doherty
bf6bd4eca8 Fix Android CI release build 2023-05-09 13:22:47 -04:00
Benjamin Doherty
36c69dda9c Fix Android CI release build 2023-05-09 13:22:20 -04:00
Benjamin Doherty
f14fdc11f5 Bump MATERIAL_VERSION to 36 2023-05-08 18:07:13 -04:00
Benjamin Doherty
0157487b1f Bump version to 1.36.0 2023-05-08 18:06:59 -04:00
Benjamin Doherty
1716c856c3 Merge branch 'rc/1.35.0' into release 2023-05-08 18:05:07 -04:00
Benjamin Doherty
31521c70c2 Fix Android CI release build 2023-05-01 18:45:34 -04:00
Benjamin Doherty
8cac90d81e Bump MATERIAL_VERSION to 35 2023-05-01 13:37:24 -04:00
Benjamin Doherty
388b4f5efb Bump version to 1.35.0 2023-05-01 13:36:40 -04:00
Benjamin Doherty
f9b5a7f301 Merge branch 'rc/1.34.0' into release 2023-05-01 13:35:16 -04:00
Benjamin Doherty
00c78cc225 Bump MATERIAL_VERSION to 34 2023-05-01 12:20:50 -04:00
Powei Feng
2bdb8b560c vulkan: fix RenderPass size (#6775) (#6776) 2023-04-27 14:09:46 -07:00
Benjamin Doherty
a974fddd4c Bump version to 1.34.0 2023-04-25 13:10:24 -04:00
Benjamin Doherty
83a3e243da Merge branch 'rc/1.33.0' into release 2023-04-25 13:09:13 -04:00
Benjamin Doherty
f286e308bf Bump version to 1.33.0 2023-04-19 14:11:43 -04:00
Benjamin Doherty
29af3be2e3 Merge branch 'rc/1.32.4' into release 2023-04-19 14:10:10 -04:00
Mathias Agopian
6623dcbebf fix specification constant injection in glsl
- boolean where handled as int
- always cast float to float()
2023-04-17 10:29:02 -04:00
Powei Feng
89dc43f361 vulkan: fix spec constant bool size 2023-04-17 10:28:44 -04:00
Ben Doherty
b77aac43ea Fix float spec constant formatting (#6731) 2023-04-17 10:28:24 -04:00
Benjamin Doherty
0d63fa02ee Fix build when exceptions disabled 2023-04-17 10:28:00 -04:00
Benjamin Doherty
e187bc442d Bump version to 1.32.4 2023-04-11 11:19:53 -07:00
Benjamin Doherty
d62268fbfb Merge branch 'rc/1.32.3' into release 2023-04-11 11:17:47 -07:00
Benjamin Doherty
3b73e3de60 Bump version to 1.32.3 2023-04-05 12:56:23 -07:00
Benjamin Doherty
7685736d6c Merge branch 'rc/1.32.2' into release 2023-04-05 12:55:12 -07:00
Benjamin Doherty
b8a3a7f221 Bump version to 1.32.2 2023-04-05 12:53:18 -07:00
Benjamin Doherty
732628acf5 Merge branch 'rc/1.32.1' into release 2023-03-21 15:48:49 -07:00
Powei Feng
7bd00d2b30 Revert "vulkan: async readPixels (#6605)" (#6644)
This reverts commit 6f25e8ae5a.

Reason for revert: breaks clients that were expecting synchronous readPixels for vulkan
2023-03-16 13:05:48 -07:00
Benjamin Doherty
0759797e61 Include atomic in VulkanCommands.h 2023-03-15 11:15:24 -07:00
Benjamin Doherty
fddc5160c7 Bump version to 1.32.1 2023-03-14 10:48:50 -07:00
Benjamin Doherty
2a579c460f Merge branch 'rc/1.32.0' into release 2023-03-14 10:46:27 -07:00
Ben Doherty
efdc801cff Vulkan: fix stack use-after-free (#6639) 2023-03-14 10:24:18 -07:00
Benjamin Doherty
dd654c575a Bump MATERIAL_VERSION to 32 2023-03-13 14:04:13 -07:00
Benjamin Doherty
0c2599b6ca Bump version to 1.32.0 2023-03-09 10:40:39 -08:00
Benjamin Doherty
979d6742e0 Merge branch 'rc/1.31.7' into release 2023-03-09 10:39:41 -08:00
Mathias Agopian
4a7a033d04 Fix timer query breakage on webgl 2023-03-08 12:33:11 -08:00
Ben Doherty
5c2bbcb4a1 Remove compute shader assertion in GLProgram 2023-03-08 10:56:24 -08:00
Benjamin Doherty
dec903a0ee Bump version to 1.31.7 2023-03-01 11:26:50 -08:00
Benjamin Doherty
e9475b322b Merge branch 'rc/1.31.6' into release 2023-03-01 11:24:13 -08:00
Benjamin Doherty
d9aead8aac Linux CI: use clang 14 2023-02-21 16:06:48 -08:00
Powei Feng
ce148ebeb1 Update linux clang to version 10 (#6556) 2023-02-21 16:00:07 -08:00
Powei Feng
e1315fbaa7 Update CI Ubuntu version 2023-02-21 15:57:00 -08:00
Benjamin Doherty
1690549392 Bump version to 1.31.6 2023-02-21 13:18:22 -08:00
Benjamin Doherty
c73710e343 Merge branch 'rc/1.31.5' into release 2023-02-21 13:16:47 -08:00
Benjamin Doherty
092a0489da Bump version to 1.31.5 2023-02-13 11:53:24 -08:00
Benjamin Doherty
1c6279366f Merge branch 'rc/1.31.4' into release 2023-02-13 11:51:37 -08:00
Benjamin Doherty
2ac85049a9 Bump version to 1.31.4 2023-02-01 11:55:09 -08:00
Benjamin Doherty
ea428a27d1 Merge branch 'rc/1.31.3' into release 2023-02-01 11:54:04 -08:00
Benjamin Doherty
3bef718238 Bump version to 1.31.3 2023-01-24 14:28:21 -08:00
Benjamin Doherty
e9daaa0503 Merge branch 'rc/1.31.2' into release 2023-01-24 14:27:02 -08:00
Benjamin Doherty
6d3ea21993 Bump version to 1.31.2 2023-01-18 14:25:45 -08:00
Benjamin Doherty
f39127605d Merge branch 'rc/1.31.1' into release 2023-01-18 14:23:59 -08:00
Ben Doherty
db476f00f7 Vulkan: fix stack-use-after-scope ASAN error in VulkanContext.cpp (#6463) 2023-01-18 11:59:10 -08:00
Powei Feng
e54d8ec0a4 [vulkan] enumerate: Returning size = 0 is ok (#6450) 2023-01-17 10:14:45 -08:00
Benjamin Doherty
b95ac854b4 Bump version to 1.31.1 2023-01-10 12:52:49 -05:00
Benjamin Doherty
96443f6dac Merge branch 'rc/1.31.0' into release 2023-01-10 12:50:52 -05:00
Benjamin Doherty
e5e3dff9fa Fix Windows CI build 2022-12-20 11:01:49 -07:00
Benjamin Doherty
31d9da38fd Fix CI build 2022-12-20 09:40:07 -07:00
Benjamin Doherty
c5a59bb6b1 Bump version to 1.31.0 2022-12-19 11:14:41 -08:00
Benjamin Doherty
e6854148d7 Merge branch 'rc/1.30.0' into release 2022-12-19 11:12:54 -08:00
Ben Doherty
001daf0350 Fix memory leak in Scene (#6387) 2022-12-16 14:38:38 -08:00
Mathias Agopian
d26c48110d Fix Android continuous and presubmit builds
Since cmake 3.25 LINUX is automatically set based on CMAKE_SYSTEM_NAME, 
which the android cmake files are setting to "Linux". This created an
inconsistant state in our build system.
2022-12-15 11:54:33 -08:00
Mathias Agopian
15a070f0c9 Fix unused variables in release builds 2022-12-12 11:17:42 -08:00
Mathias Agopian
93230b06cf Fix Android continuous and presubmit builds
Since cmake 3.25 LINUX is automatically set based on CMAKE_SYSTEM_NAME, 
which the android cmake files are setting to "Linux". This created an
inconsistant state in our build system.
2022-12-12 10:44:49 -08:00
Mathias Agopian
4c55fc3c48 fix TransformManager high precision mode
The residual would be corrupted when using transactions.
2022-12-12 10:06:18 -08:00
Mathias Agopian
a4797a631e Fix a WebGL program link failure
WebGL complained about:

Precisions of uniform block 'ShadowUniforms' member
'ShadowUniforms.shadows.texelSizeAtOneMeter' differ between VERTEX and
FRAGMENT shaders.


this field didn't have a precision qualifier, this might be specific to
WebGL or a Chrome bug, unsure. Either we fix it by specifying all
qualifiers.
2022-11-29 23:32:21 -08:00
Mathias Agopian
5868e857aa Fix a WebGL program link failure
WebGL complained about:

Precisions of uniform block 'ShadowUniforms' member 
'ShadowUniforms.shadows.texelSizeAtOneMeter' differ between VERTEX and 
FRAGMENT shaders.


this field didn't have a precision qualifier, this might be specific to
WebGL or a Chrome bug, unsure. Either we fix it by specifying all
qualifiers.
2022-11-29 23:31:03 -08:00
Benjamin Doherty
32ba063bd7 Fix CI build 2022-11-21 16:58:37 -08:00
Benjamin Doherty
02f0839f9b Bump version to 1.30.0 2022-11-21 15:45:27 -08:00
Benjamin Doherty
e50f3c4c91 Merge branch 'rc/1.29.0' into release 2022-11-21 15:44:36 -08:00
Ben Doherty
7ca15bfc16 Use GitHub script for release asset upload (#6308) 2022-11-21 15:39:55 -08:00
Benjamin Doherty
b7b7afb62a Bump version to 1.29.0 2022-11-09 16:43:08 -08:00
Benjamin Doherty
b8ff6a9ad9 Merge branch 'rc/1.28.3' into release 2022-11-09 16:38:02 -08:00
Ben Doherty
a1dcb4f259 Skip rendering renderables with missing geometry (#6281) 2022-11-09 16:35:31 -08:00
Benjamin Doherty
c3501393fd Fix RELEASE_NOTES 2022-11-09 16:34:42 -08:00
Benjamin Doherty
a03de75a4e Update Python requirements.txt to fix CI 2022-11-04 11:35:21 -07:00
Benjamin Doherty
0c3d59aba3 Update Python requirements.txt to fix CI 2022-11-03 16:15:29 -07:00
Benjamin Doherty
ce1987f291 Fix iOS build 2022-11-01 15:10:57 -04:00
Benjamin Doherty
7eae926ed7 Bump version to 1.28.3 2022-11-01 13:03:19 -04:00
Benjamin Doherty
606af52eda Merge branch 'rc/1.28.2' into release 2022-11-01 13:02:05 -04:00
Benjamin Doherty
8ea30aea80 Update RELEASE_NOTES for 1.28.2 2022-11-01 13:00:04 -04:00
Ben Doherty
72dea695ba Fix G3 reported ubsan warning (#6257) 2022-11-01 12:52:10 -04:00
daemyung jang
1c0cf56ed0 Fix glTF breaking issue (#6239) 2022-10-28 13:08:50 -04:00
Benjamin Doherty
bcb4eb35cd Bump version to 1.28.2 2022-10-25 13:09:21 -04:00
Benjamin Doherty
bce2676335 Merge branch 'rc/1.28.1' into release 2022-10-25 13:08:12 -04:00
Benjamin Doherty
510097d722 Force spirv-cross to keep Metal argument buffer equal sized 2022-10-25 12:10:54 -04:00
Ben Doherty
5784118a12 Fix, don't use encoder.device (#6212) 2022-10-20 13:08:04 -04:00
Ben Doherty
b5de0f2d23 Metal: fix simulator regression due to arg buffers (#6181) 2022-10-18 16:29:18 -04:00
Ben Doherty
04df3f4dad Metal: fix use-after-free ASAN error (#6203) 2022-10-18 16:29:12 -04:00
Benjamin Doherty
093ca4d623 Revert Android API_LEVEL change
glDispatchCompute requires Android API level 21, however bumping our
required minimum from 19 to 21 caused some clients' builds to fail.
Commenting-out that line for now to proceed with the 1.28.0 upgrade.
2022-10-18 16:28:58 -04:00
Mathias Agopian
a3ed1558d3 fix a directional light shadowing issue
the shadow frustum could be be smaller than expected.
2022-10-14 15:42:03 -07:00
Benjamin Doherty
fe379070ae Bump version to 1.28.1 2022-10-13 12:38:40 -04:00
Benjamin Doherty
ead0a2f597 Merge branch 'rc/1.28.0' into release 2022-10-13 12:37:07 -04:00
Benjamin Doherty
5dbc593f90 Fix assertion in ShadowMap due to vertexCount 2022-10-12 18:16:53 -04:00
Ben Doherty
d47dc12bf0 G3 fixes for 1.28.0 (#6174) 2022-10-12 14:48:37 -04:00
Philip Rideout
4392f63e57 Vulkan: fix black screen regression
The VulkanProgram constructor was bailing out early and emitting a
warning because it saw that one of the stages wasn't fulfilled.
However it's okay for a pipeline to be missing a compute program.

Fixes regression that started with fabba73b1.
2022-10-12 14:48:25 -04:00
Ben Doherty
f64b7aca79 Fix Metal argument buffer encoding (#6150) 2022-10-10 10:07:41 -07:00
Ben Doherty
3762ec5750 Fix Android release build (#6133) 2022-09-30 11:48:13 -07:00
Ben Doherty
4591fdd9c0 Fix Android release build (#6133) 2022-09-30 11:47:41 -07:00
Benjamin Doherty
d3068d2a0e Bump version to 1.28.0 2022-09-29 13:13:18 -07:00
Benjamin Doherty
ca0a4bc23a Merge branch 'rc/1.27.2' into release 2022-09-29 13:12:20 -07:00
Ben Doherty
433c163c61 Fixes for 1.27.2 release (#6125) 2022-09-29 13:08:39 -07:00
Mathias Agopian
c867fda883 don't attempt to use more texture units than present
to emulate the bindless API in the gl backend we always used the highest
texture unit available. However at feature level 3, we support up to 62
textures, so the that max was bumped to 62 -- however, where we're not
on a feature level 2 device, that texture unit doesn't exist.

Instead we now always use binding 31, which is guaranteed to exist by 
EGL's minspec.
2022-09-29 12:36:28 -07:00
Ben Doherty
566540ae6d Fix generic/Mutex.h not installed on Linux (#6117) 2022-09-28 16:49:03 -07:00
Benjamin Doherty
759f490dae Bump version to 1.27.2 2022-09-20 11:06:02 -07:00
Benjamin Doherty
867d4d44f5 Merge branch 'rc/1.27.1' into release 2022-09-20 11:04:58 -07:00
Ben Doherty
65747d5877 Update glslang to c0cf8ad87 (master) (#6076) 2022-09-19 12:38:05 -07:00
Benjamin Doherty
bd357f6076 Bump version to 1.27.1 2022-09-13 10:15:12 -07:00
Benjamin Doherty
9e960b7d45 Merge branch 'rc/1.27.0' into release 2022-09-13 10:14:24 -07:00
Ben Doherty
0c54d4a6a1 Add return case to fix G3 compiler error (#6062) 2022-09-12 12:34:56 -07:00
Philip Rideout
526e846a81 gltfio: fix crash when material is unspecified 2022-09-06 22:13:19 -04:00
Philip Rideout
8bcfa373d4 gltfio: do not clear out the texture slots too early.
This broke asyncGetLoadProgress() and caused WebGL to crash reliably
because ResourceLoader got destroyed too soon.

Bug was introduced with de7dfc2ea6.

I intend to cherry pick this to rc/1.27.0, which is where it was
introduced, so there's no need to update the release notes.
2022-09-06 22:09:25 -04:00
Ben Doherty
db9a0f2c1f Fix uberarchive not included in PodSpec (#6024) 2022-09-01 12:18:20 -04:00
Benjamin Doherty
646b1e2193 Bump version to 1.27.0 2022-08-31 12:04:55 -04:00
Benjamin Doherty
41bd30f81d Merge branch 'rc/1.26.0' into release 2022-08-31 12:04:08 -04:00
Mathias Agopian
bd626aea27 material instanced property must be false by default.
this fix auto-instancing.
2022-08-26 18:28:56 -04:00
Benjamin Doherty
8a03f75485 Bump version to 1.26.0 2022-08-23 12:07:33 -04:00
Benjamin Doherty
dc9594fbdf Merge branch 'rc/1.25.6' into release 2022-08-23 12:05:44 -04:00
Benjamin Doherty
fad0b533c0 Update RELEASE_NOTES for 1.25.6 2022-08-23 12:03:28 -04:00
Benjamin Doherty
4d773e9453 Bump version to 1.25.6 2022-08-16 11:03:07 -07:00
Benjamin Doherty
e0c610b013 Merge branch 'rc/1.25.5' into release 2022-08-16 11:01:04 -07:00
Benjamin Doherty
0da4f36c33 Bump version to 1.25.5 2022-08-16 10:54:00 -07:00
Ben Doherty
11d17e1db3 Fix SamplerGroup update issue (#5928) 2022-08-16 10:47:18 -07:00
Philip Rideout
b8c318d923 WASM: Allow clients to enable pthreads.
Filament does not yet fully support threads with WASM, but this is a
baby step in that direction.

To enable experimental pthreads support, enable the WEBGL_PTHREADS CMake
option. This will enable pthreads support in `gltfio` and `utils`, which
is known to work, but not when served with GitHub Pages.

The web server must emit COOP, COEP and CORP headers, so our build
instructions now recommend the use of `emrun` for local testing.

This also changes our demos so that they do not use unpkg, which
does not work when using `emrun`, due to cross-origin restrictions.
2022-08-12 15:43:28 -07:00
Philip Rideout
e563cc6f5e gltfio: add 'detach' methods to allow ownership transfer
These new methods allow gltfio to be integrated into internal Google
libraries.
2022-08-09 14:37:44 -07:00
Benjamin Doherty
cb8914ab96 Merge branch 'rc/1.25.4' into release 2022-08-02 11:47:52 -07:00
Alan Eneev
2fecda7bdc Headless EGL: Fallback to a 24-bit depth buffer 2022-07-29 10:10:46 -07:00
Alan Eneev
543d8efb25 Fix PlatformEGLHeadless build and add a build.sh option to build EGL
I have disabled building SDL with headless EGL, because
SDL_config_minimal.h doesn't work with EGL, and I don't know how to
implement a an SDL config that would work with EGL.

I have verified that this works in a separate project, and that it
compiles in this project.

```
$ ./build.sh -e release
$ find ./out/ -name "*EGL*"
./out/cmake-release/filament/backend/CMakeFiles/backend.dir/src/opengl/platforms/PlatformEGL.cpp.o
./out/cmake-release/filament/backend/CMakeFiles/backend.dir/src/opengl/platforms/PlatformEGLHeadless.cpp.o
./out/cmake-release/libs/bluegl/CMakeFiles/bluegl.dir/src/BlueGLLinuxEGL.cpp.o
```
2022-07-29 10:10:40 -07:00
Benjamin Doherty
396b1079a7 Bump version to 1.25.4 2022-07-26 09:57:34 -07:00
Benjamin Doherty
eedcd9f8cb Merge branch 'rc/1.25.3' into release 2022-07-26 09:56:22 -07:00
Ben Doherty
ab252b210c Fix config object related build failures (#5814) 2022-07-26 09:51:55 -07:00
Benjamin Doherty
1071b8ea90 Bump version to 1.25.3 2022-07-18 15:15:03 -07:00
Benjamin Doherty
a9c5bbf185 Merge branch 'rc/1.25.2' into release 2022-07-18 15:14:02 -07:00
Philip Rideout
8dd4bff7a7 gltfio: minor fixups to prep for g3 integration. 2022-07-18 15:11:36 -07:00
Philip Rideout
77c54446af gltfio: use openLocalTransformTransaction API. 2022-07-18 15:11:30 -07:00
Benjamin Doherty
4a0bc0af57 Bump version to 1.25.2 2022-07-11 15:51:55 -07:00
Benjamin Doherty
a171e75e70 Merge branch 'rc/1.25.1' into release 2022-07-11 15:50:10 -07:00
Benjamin Doherty
f5ffa092fe Bump version to 1.25.1 2022-07-06 19:09:34 -07:00
Benjamin Doherty
8de5fdd551 Merge branch 'rc/1.25.0' into release 2022-07-06 19:07:42 -07:00
Benjamin Doherty
ecca3abe98 Bump version to 1.25.0 2022-07-01 12:09:29 -07:00
Benjamin Doherty
8570e35224 Merge branch 'rc/1.24.0' into release 2022-07-01 12:08:48 -07:00
Benjamin Doherty
84df9f9a03 Update version to 1.24.0 2022-07-01 12:03:38 -07:00
Benjamin Doherty
5f93fb9613 Update RELEASE_NOTES for 1.24.0 2022-06-30 16:37:03 -07:00
Benjamin Doherty
75f77fdbdd Metal: Allow Filament to disregard MTLTexture pixelFormat when importing 2022-06-30 16:04:28 -07:00
Ben Doherty
7825d582c2 CocoaPods: include uberz library (#5752) 2022-06-30 14:07:55 -07:00
Ben Doherty
5deb0ba933 CocoaPods: include uberz library (#5752) 2022-06-30 14:04:02 -07:00
Philip Rideout
18e917aaf2 ImGuiHelper: add support for Y flip. (#5748)
Reflects a change from Betty and should be cherry picked to v1.23.3

Users could customize the ImGuiHelper camera, but they had no control
over the scissor coordinates. This allows them to use vertically flipped
coordinates, which, unfortunately, is required for MediaPipe
integration.
2022-06-30 12:05:53 -07:00
Benjamin Doherty
a165f3890a Vulkan: Support VMA_DYNAMIC_VULKAN_FUNCTIONS off 2022-06-21 15:09:26 -07:00
Philip Rideout
f88b6d9c97 Do not trigger UB with string_view. 2022-06-20 21:11:52 -07:00
Mathias Agopian
db8ecd9952 configure render primitive pool so it works with msvc
a 64 bytes pool seems to work with both clang and msvc, unfortunately,
c++ doesn't let us know the allocator object size at compile time
for map containers, so we have to guess.
2022-06-16 13:13:33 -07:00
Benjamin Doherty
b5ec06c2d2 Fix build breakage 2022-06-15 17:05:53 -07:00
Benjamin Doherty
dfbac8385e Merge branch 'rc/1.23.2' into release 2022-06-15 11:49:06 -07:00
Ben Doherty
eaab737b2c NoopDriver: return unique handles (#5697) 2022-06-15 11:28:08 -07:00
Benjamin Doherty
45991cda0a Update RELEASE_NOTES for 1.23.2 2022-06-13 15:27:03 -07:00
Philip Rideout
09a016bb6f Remove an 'API Change' warning from one item. 2022-06-09 14:33:49 -07:00
Benjamin Doherty
31607d355d Bump version to 1.23.2 2022-06-06 15:18:57 -07:00
Benjamin Doherty
a67d50b9e2 Merge branch 'rc/1.23.1' into release 2022-06-06 15:16:53 -07:00
Benjamin Doherty
8d42f53c80 Update RELEASE_NOTES for 1.23.1 2022-06-06 15:14:59 -07:00
Benjamin Doherty
5e21a55bce Revert "iOS: implement headleass swapchain (#5486)"
This reverts commit 0e5ba60cb6.

This causes issues in G3, still need to investigate the root cause.
2022-06-06 13:17:04 -07:00
Benjamin Doherty
85930ea2e8 Bump version to 1.23.1 2022-06-01 14:58:29 -07:00
Benjamin Doherty
4b3cde8b39 Merge branch 'rc/1.23.0' into release 2022-06-01 14:55:20 -07:00
Benjamin Doherty
eedfa85355 Update RELEASE_NOTES for 1.23.0 2022-06-01 14:54:50 -07:00
Benjamin Doherty
bb54c6c807 Release Filament 1.23.0 2022-06-01 14:53:48 -07:00
Benjamin Doherty
c23f905858 Update RELEASE_NOTES for 1.23.0 2022-06-01 14:53:23 -07:00
Benjamin Doherty
8c7be0a1d0 Revert "iOS: implement headleass swapchain (#5486)"
This reverts commit 0e5ba60cb6.

This causes issues in G3, still need to investigate the root cause.
2022-05-31 11:29:38 -07:00
Benjamin Doherty
878497b3d5 Bump version to 1.23.0 2022-05-24 15:19:51 -07:00
Benjamin Doherty
a155561769 Merge branch 'rc/1.22.2' into release 2022-05-24 15:17:55 -07:00
Benjamin Doherty
8b86a0ed2e Revert "iOS: implement headleass swapchain (#5486)"
This reverts commit 0e5ba60cb6.

This causes issues in G3, still need to investigate the root cause.
2022-05-24 20:13:10 +00:00
Benjamin Doherty
26f9a9b122 Update RELEASE_NOTES for 1.22.2 2022-05-23 12:40:47 -07:00
Ben Doherty
6e5f6978fb Add ktxreader and viewer libs to CocoaPods (#5573) 2022-05-18 16:24:51 -07:00
Ben Doherty
0d31d7b2de Add ktxreader and viewer libs to CocoaPods (#5573) 2022-05-18 16:13:26 -07:00
Benjamin Doherty
dd862b7e0a Bump version to 1.22.2 2022-05-17 18:03:38 -07:00
Benjamin Doherty
52065f2cbd Merge branch 'rc/1.22.1' into release 2022-05-17 18:01:37 -07:00
Benjamin Doherty
77c02d5831 Update RELEASE_NOTES for 1.22.1 2022-05-17 17:51:37 -07:00
Ben Doherty
f7e4c8d16d Add package name back to AndroidManifest.xml to fix G3 (#5569) 2022-05-17 17:51:21 -07:00
Philip Rideout
b7410474ff Workaround: partially revert "don't issue a flush..."
This is a temporary workaround for a memory corruption issue observed on
some devices from a specific vendor. We will try to make this workaround
more targeted in a subequent change.

Partial revert for b2cdf9f2b4.
2022-05-16 10:39:55 -07:00
Benjamin Doherty
384cc4ebf6 Bump version to 1.22.1 2022-05-09 13:09:39 -07:00
Benjamin Doherty
7c0643f122 Merge branch 'rc/1.22.0' into release 2022-05-09 13:07:07 -07:00
Benjamin Doherty
58abae3067 Release Filament 1.22.0 2022-05-09 13:07:02 -07:00
Benjamin Doherty
4742693869 Update RELEASE_NOTES for 1.22.0 2022-05-09 13:04:55 -07:00
Benjamin Doherty
6c39e474ea Bump version to 1.22.0 2022-05-02 13:19:43 -04:00
Benjamin Doherty
5c8977c906 Merge branch 'rc/1.21.3' into release 2022-05-02 13:18:06 -04:00
Benjamin Doherty
6b3cc2e2f3 Update RELEASE_NOTES for 1.21.3 2022-05-02 13:12:51 -04:00
Ben Doherty
76b2edd6ea Move toCompressedFilamentEnum to Ktx1Reader header (#5493) 2022-04-29 14:50:24 -04:00
Benjamin Doherty
8ebb37d011 Bump version to 1.21.3 2022-04-25 16:58:19 -04:00
Benjamin Doherty
3366db83ef Merge branch 'rc/1.21.2' into release 2022-04-25 16:57:25 -04:00
Benjamin Doherty
5415254aac Update RELEASE_NOTES for 1.21.2 2022-04-25 16:50:53 -04:00
Benjamin Doherty
caacc61602 Bump version to 1.21.2 2022-04-13 10:21:09 -06:00
Benjamin Doherty
e902df19b2 Merge branch 'rc/1.21.1' into release 2022-04-13 10:15:50 -06:00
Benjamin Doherty
40b372dda7 Update RELEASE_NOTES for 1.21.1 2022-04-12 14:36:26 -06:00
Ben Doherty
fd330a98aa Releases: specify Python dependencies in requirements.txt file (#5402) 2022-04-08 14:35:51 -07:00
Ben Doherty
90c23a7d5d Correctly prepare color grading as subpass programs (#5384) 2022-04-04 13:24:01 -07:00
Benjamin Doherty
5710304114 Bump version to 1.21.1 2022-04-04 13:03:39 -07:00
Benjamin Doherty
0f684820bc Merge branch 'rc/1.21.0' into release 2022-04-04 13:01:35 -07:00
Benjamin Doherty
e1d2d6ade6 Update RELEASE_NOTES for 1.21.0 2022-04-04 12:58:54 -07:00
Mathias Agopian
e0b6f2ca71 Change version number to 1.21.0 2022-04-04 12:41:38 -07:00
Benjamin Doherty
2a35ee279b Bump version to 1.20.6 2022-03-28 18:47:38 -07:00
Benjamin Doherty
5abf780360 Merge branch 'rc/1.20.5' into release 2022-03-28 18:45:34 -07:00
Benjamin Doherty
181f158ea9 Update RELEASE_NOTES for 1.20.5 2022-03-28 18:42:41 -07:00
Benjamin Doherty
f4087fc81d Bump version to 1.20.5 2022-03-21 14:45:34 -07:00
Benjamin Doherty
82793f9b82 Merge branch 'rc/1.20.4' into release 2022-03-21 14:43:15 -07:00
Benjamin Doherty
537576e84a Update RELEASE_NOTES for 1.20.4 2022-03-21 14:42:35 -07:00
Benjamin Doherty
375d1f55e3 Release Filament 1.20.4 2022-03-21 14:38:14 -07:00
Benjamin Doherty
597218963f Update RELEASE_NOTES for 1.20.4 2022-03-21 14:37:17 -07:00
Philip Rideout
0d29b3ddc8 WebGL endFrame: restore additional default state. 2022-03-21 09:43:45 -07:00
Philip Rideout
e0e3b42623 WebGL: reset VAO and texture bindings when frame ends. 2022-03-21 09:43:33 -07:00
Benjamin Doherty
0995ca6614 Bump version to 1.20.4 2022-03-14 14:19:36 -07:00
Benjamin Doherty
3d741fc8d4 Merge branch 'rc/1.20.3' into release 2022-03-14 14:17:54 -07:00
Benjamin Doherty
c20772b458 Update RELEASE_NOTES for 1.20.3 2022-03-10 12:56:56 -08:00
Benjamin Doherty
4a6b659098 Bump version to 1.20.3 2022-03-07 14:12:28 -08:00
Benjamin Doherty
2b93f08ca5 Merge branch 'rc/1.20.2' into release 2022-03-07 14:10:07 -08:00
Benjamin Doherty
fca62b8fff Update RELEASE_NOTES for 1.20.2 2022-03-07 14:08:41 -08:00
Benjamin Doherty
dee6d9de2c Bump version to 1.20.2 2022-03-02 15:51:38 -08:00
Benjamin Doherty
574e3e7521 Merge branch 'rc/1.20.1' into release 2022-03-02 15:47:21 -08:00
Ben Doherty
29fdf82ac5 Make VulkanContext compatible with vk_mem_alloc dev 2022-03-02 12:10:20 -08:00
Benjamin Doherty
c8cf2a54e8 Update RELEASE_NOTES for 1.20.1 2022-03-02 11:47:51 -08:00
Ben Doherty
bfd32e67d4 Apply some clang-tidy fixes (#5281) 2022-03-02 11:43:51 -08:00
Benjamin Doherty
95915367fa Bump version to 1.20.1 2022-02-24 13:59:17 -08:00
Benjamin Doherty
b769cfda62 Merge branch 'rc/1.20.0' into release 2022-02-24 13:57:10 -08:00
Benjamin Doherty
40ac88dfed Bump MATERIAL_VERSION to 20 2022-02-24 13:53:43 -08:00
Benjamin Doherty
6d96082f07 Bump version to 1.20.0 2022-02-24 13:53:43 -08:00
Benjamin Doherty
21f913db1c Update RELEASE_NOTES for 1.20.0 2022-02-24 13:53:42 -08:00
Ben Doherty
699a578966 Add SSR variant filter to matc (#5256) 2022-02-24 10:36:06 -08:00
Philip Rideout
babbfa1394 Vulkan: remove layout checks from SwapChain.
These asserts will come back in the next Filament release, which has
much cleaner layout tracking.
2022-02-23 13:48:13 -08:00
Philip Rideout
73f0d58e10 Vulkan: fix backend tests (especially ReadPixels Y flip). 2022-02-23 11:18:29 -08:00
Philip Rideout
9cb4b74bbd Fix web apps that have multiple Filament viewers.
This fixes the emscripten binding errors that we've been seeing
with the <filament-viewer> test page, which prevented us from
including web in the last few Filament releases.

The binding errors were caused by double-initializing the emscripten
module.

I fixed this by allowing clients (e.g. FilamentViewer) to call
Filament.init() more than once. We now accumulate a list of "on ready"
callbacks that get triggered after the emscripten module becomes ready.

As far as I can tell, multiple canvases were actually always broken, and
the viewer test page worked in the past only because we got lucky.
2022-02-22 16:12:13 -08:00
Ben Doherty
f2c8456971 Attempt to fix Windows CI builds (#5205) 2022-02-15 11:12:19 -08:00
Benjamin Doherty
e571600c30 Bump version to 1.19.1 2022-02-14 13:48:59 -08:00
Benjamin Doherty
e84c94d3eb Merge branch 'rc/1.19.0' into release 2022-02-14 13:46:45 -08:00
Benjamin Doherty
e791d4818f Update RELEASE_NOTES for 1.19.0 2022-02-14 13:45:28 -08:00
Philip Rideout
cdadb43e50 WebGL: another fix for BufferDescriptor bindings.
The previous code would convert each element of the source data
into 8 bit-per-element, but we wnat to preserve the original format
that the user provides.

The new solution is to use `slice()` which is a robust way to clone
all the data in a typed array.

This fixes the new regression with Triangle that Ben caught.
2022-02-09 15:44:26 -08:00
Philip Rideout
edaff60fbf WebGL: remove buffer sharing optimization.
If emscripten grows the heap inside one of our BufferDescriptor binding
functions, then the old heap becomes "detached" and an error can
occur.

This fixes the issue seen with the Parquet demo that Ben caught.
2022-02-09 12:51:06 -08:00
Benjamin Doherty
3f64e46557 Bump version to 1.19.0 2022-02-08 09:27:14 -08:00
Benjamin Doherty
32dab23bc6 Update RELEASE_NOTES for 1.18.0 2022-02-08 09:25:41 -08:00
Benjamin Doherty
362de7dd31 Merge branch 'rc/1.18.0' into release 2022-02-08 09:22:21 -08:00
Benjamin Doherty
6b01fbb903 Correct version to 1.18.0 2022-02-08 09:08:49 -08:00
Philip Rideout
4934d9f7bc Vulkan: fix leak when apps do not draw anything. 2022-02-07 14:27:06 -08:00
Ben Doherty
ed73955b00 Initialize useLegacyMorphing to fix MSAN warning (#5164) 2022-02-07 11:28:03 -08:00
Ben Doherty
3f1f2726c4 Add a MAX_LEGACY_MORPH_TARGETS definition (#5163) 2022-02-07 11:27:45 -08:00
Philip Rideout
1a7bd7ea8d Rewrite VulkanPipelineCache (without changing its API).
All three types of caches (descriptor sets, pipelines, and pipeline
layouts) are now managed in exactly the same way. They all use an LRU
eviction scheme that is based on a count of command buffer flush
events.

Vulkan objects can only be destroyed if there are no in-flight command
buffers that reference them, so an easy way to know when it is safe to
evict a given entry is to wait for "N" flushes after its last use, where
"N" is the number of command buffers in the command buffer ring.

Another big simplification is that there are no more dirty flags,
instead we store two sets of state vectors for each type of cache: the
"currently bound" state, and the "current requirements" state.
2022-02-03 17:37:40 -08:00
Ben Doherty
946ea43436 Fix sampler overflow check in SamplerBindingMap (#5143) 2022-02-02 11:04:48 -08:00
Philip Rideout
7f42385f5f VulkanPipelineCache: fix spurious SEGV.
Fixes #5142 by replacing unsafe pointers with map keys.

One of the differences between robin_map and unordered_map is the
following:

    pointers to keys or values in the map are invalidated in the same
    way as iterators to these keys-values

Therefore it is unsafe to track the pointer to a value that is stored
in a robin_map.
2022-02-01 16:34:34 -08:00
Philip Rideout
8845ac2b75 VulkanPipelineCache: code cleanup and minor fixes.
This is mostly just code cleanup. One actual bug was the fact that the
dummy sampler was re-created every time a new pipeline layout was
created.

It also felt strange to use `auto&` to refer to a C-style array. I
changed this into a `std::array` which is more consistent with other
fixed size arrays in this class.
2022-02-01 16:34:27 -08:00
Ben Doherty
da85001d4d Support legacy morphing (#5129)
Support legacy morphing (morphing with targets supplied via VertexAttributes) for older clients. This gives clients more time to transition over to the new MorphTargetBuffer API.
2022-01-27 16:08:46 -08:00
Mathias Agopian
8d15079937 update material version to 18 2022-01-26 14:32:46 -08:00
daemyung jang
adc542b5cd Bind samplers to specified shader stages (#5036)
Co-authored-by: Ben Doherty <bendoherty@google.com>
Co-authored-by: Mathias Agopian <mathias@google.com>
2022-01-26 14:31:30 -08:00
Philip Rideout
72feb044a3 Vulkan now supports offsets when uploading texture data.
This allows `MorphStressTest` to work on Vulkan.

However, `Horse` is still broken because it provides positions but not
tangents.  Separate fix for that is coming.

Partial fix for #5109.
2022-01-26 14:31:00 -08:00
Philip Rideout
c0ba260ddf Vulkan: clean up image layout management.
This fixes validation errors and makes a first pass at simplification.
VulkanTexture now tracks image layout using RangeMap, which paves the
way for further simplification.
2022-01-26 14:30:45 -08:00
Philip Rideout
2da215e8e7 RangeMap: improve naming convention, etc. 2022-01-26 14:30:23 -08:00
Philip Rideout
4d0368b5f1 RangeMap: improve the auto-merge functionality. 2022-01-26 14:30:09 -08:00
Philip Rideout
d11c78857d utils: introduce RangeMap container and unit test.
This will allow the Vulkan backend to efficiently track the subresource
image layouts for each texture.

This is a sparse container for a series of ordered non-overlapping
integer intervals, where each interval maps to a concrete value.
2022-01-26 14:29:45 -08:00
Mathias Agopian
e829d90c4a A morphing buffer must be bound when skinning is active
This is because we're using the same program variant for skinning
and morphing, in the skinning-only case, the buffer won't be accessed
in the shader, but it must be present.

fixes #5085
2022-01-26 14:20:25 -08:00
Mathias Agopian
a8006acd33 [GL backend] fix sampler binding bug
When a program had an unused SamplerInterfaceBlock, other samplers
could be bound to the wrong TMU

Fixes #5088
2022-01-26 14:20:12 -08:00
Benjamin Doherty
86ec502040 Bump version to 1.17.1 2022-01-24 12:52:22 -08:00
Benjamin Doherty
b19a73cc50 Merge branch 'rc/1.17.0' into release 2022-01-24 12:50:22 -08:00
Benjamin Doherty
a99c695932 Update RELEASE_NOTES for 1.17.0 2022-01-24 12:45:05 -08:00
Benjamin Doherty
8cd720b53a Bump version to 1.17.0 2022-01-18 13:54:40 -08:00
Benjamin Doherty
04df79e58f Merge branch 'rc/1.16.1' into release 2022-01-18 13:53:02 -08:00
Benjamin Doherty
a4b3717762 Update RELEASE_NOTES for 1.16.1 2022-01-18 13:48:36 -08:00
Benjamin Doherty
1035e442ee Bump version to 1.16.1 2022-01-10 10:47:46 -08:00
Benjamin Doherty
60d3638f15 Merge branch 'rc/1.16.0' into release 2022-01-10 10:44:36 -08:00
Benjamin Doherty
3e7d3c9035 Update RELEASE_NOTES for 1.16.0 2022-01-10 10:41:14 -08:00
jeanlemotan
eb360be2ad Fixed cubemap update 2022-01-10 10:34:49 -08:00
Benjamin Doherty
485ac8704d Bump version to 1.16.0 2022-01-04 11:55:19 -08:00
Benjamin Doherty
dc74540423 Merge branch 'rc/1.15.2' into release 2022-01-04 11:53:02 -08:00
Benjamin Doherty
96219c22db Update RELEASE_NOTES for 1.15.2 2022-01-04 11:49:50 -08:00
Romain Guy
f3b7048775 Add missing JNI impl (#4959) 2022-01-04 11:43:19 -08:00
Romain Guy
aaed6fb376 Fix rounding math 2022-01-04 11:43:13 -08:00
Romain Guy
35a5d3310f Fix preprocessor test 2021-12-13 13:53:20 -07:00
Benjamin Doherty
9da79a1d2d Bump version to 1.15.2 2021-12-13 11:08:28 -07:00
Benjamin Doherty
595b355d1b Merge branch 'rc/1.15.1' into release 2021-12-13 11:06:45 -07:00
Benjamin Doherty
3a67d769f4 Update RELEASE_NOTES for 1.15.1 2021-12-06 16:54:28 -08:00
Benjamin Doherty
6fb536a937 Bump version to 1.15.1 2021-12-06 11:17:37 -08:00
Benjamin Doherty
bb460d78d8 Merge branch 'rc/1.15.0' into release 2021-12-06 11:14:00 -08:00
Benjamin Doherty
838835a715 Update RELEASE_NOTES for 1.15.0 2021-12-02 11:52:04 -08:00
Romain Guy
63acd53e23 Use __ANDROID__ instead of ANDROID 2021-11-30 11:15:49 -08:00
Romain Guy
fcd2d0457b Workaround for a build system issue 2021-11-30 11:15:42 -08:00
Benjamin Doherty
58fc26461b Bump version to 1.15.0 2021-11-29 14:08:28 -08:00
Benjamin Doherty
fd82f6b04e Merge branch 'rc/1.14.2' into release 2021-11-29 14:07:03 -08:00
Benjamin Doherty
cef3200533 Add additional RELEASE_NOTES for 1.14.2 2021-11-29 10:21:56 -08:00
Ben Doherty
634500c398 Fix, avoid divide-by-zero inside makeBone (#4889) 2021-11-24 16:29:33 -08:00
Ben Doherty
b3e294ac54 Fix Metal depth comparison initialization (#4886) 2021-11-23 12:09:43 -08:00
Benjamin Doherty
2bf7535ad0 Update RELEASE_NOTES for 1.14.2 2021-11-22 10:16:00 -08:00
Benjamin Doherty
3315f75de9 Bump version to 1.14.2 2021-11-22 10:12:44 -08:00
Benjamin Doherty
bbe7dbfa92 Merge branch 'rc/1.14.1' into release 2021-11-22 10:11:06 -08:00
Benjamin Doherty
5697922a65 Update RELEASE_NOTES for 1.14.1 2021-11-17 12:04:26 -08:00
Ben Doherty
4da83df2b9 Fix material compilation error with device vertex domain (#4865)
A recent refactor was causing the following error when the vertex domain
was set to `device`:
```
ERROR: main.vs:23: 'material' : undeclared identifier
ERROR: main.vs:23: 'materialVertex' : no matching overloaded function found
```
2021-11-17 12:03:55 -08:00
Ben Doherty
8f156d6588 Android: re-enable VSM cascade fix (#4863) 2021-11-17 10:19:09 -08:00
Benjamin Doherty
cec0871c11 Bump version to 1.14.1 2021-11-15 10:09:45 -08:00
Benjamin Doherty
41a809368b Merge branch 'rc/1.14.0' into release 2021-11-15 10:07:56 -08:00
Romain Guy
ea53eb9290 Skip task incompatible with configuration caching (#4831) 2021-11-09 15:56:00 -08:00
Benjamin Doherty
05875057c9 Update RELEASE_NOTES for 1.14.0 2021-11-09 15:51:13 -08:00
Benjamin Doherty
44125926d1 Disable configuration-cache 2021-11-08 17:05:34 -08:00
Benjamin Doherty
60734349de Bump version to 1.14.0 2021-11-08 11:52:50 -08:00
Benjamin Doherty
fbfd5ec0ec Merge branch 'rc/1.13.0' into release 2021-11-08 11:50:19 -08:00
Ben Doherty
a74a95cc65 Call VirtualMachineEnv::JNI_OnLoad for non-Android Java builds (better fix) (#4779) 2021-11-04 13:28:17 -07:00
Benjamin Doherty
bc0ea16ff0 Update RELEASE_NOTES for 1.13.0 2021-11-02 13:15:53 -07:00
Mathias Agopian
b2dc8aa84c Fix typo that broke the directional shadowmap 2021-11-02 13:11:44 -07:00
Benjamin Doherty
9987e8b6ab Bump version to 1.13.0 2021-11-01 14:59:11 -07:00
Benjamin Doherty
0d9bdcc008 Merge branch 'rc/1.12.11' into release 2021-11-01 14:55:58 -07:00
Benjamin Doherty
b5c634045e Update RELEASE_NOTES for 1.12.11 2021-10-28 16:13:50 -07:00
Ben Doherty
1f05531d53 Call VirtualMachineEnv::JNI_OnLoad for non-Android Java builds (better fix) (#4779) 2021-10-28 16:02:23 -07:00
Benjamin Doherty
88f382f0e3 Revert "refactor colorgrading materials"
This reverts commit fb86a77cf8.
2021-10-28 15:56:50 -07:00
Benjamin Doherty
3e59925900 Remove problematic configuration-cache setting for release build 2021-10-28 15:56:26 -07:00
Benjamin Doherty
ea404f8d4f Remove problematic configuration-cache setting for release build 2021-10-26 10:09:51 -07:00
Benjamin Doherty
602a550d93 Bump version to 1.12.11 2021-10-25 12:30:37 -07:00
Benjamin Doherty
12abbe2d23 Merge branch 'rc/1.12.10' into release 2021-10-25 11:06:18 -07:00
Benjamin Doherty
fb0ee97588 Update RELEASE_NOTES for 1.12.10 2021-10-25 11:00:06 -07:00
Ben Doherty
56ef48c9c3 Fix, call VirtualMachineEnv::JNI_OnLoad for non-Android Java builds (#4749) 2021-10-25 10:39:24 -07:00
Benjamin Doherty
47c3dd3dd1 Revert "refactor colorgrading materials"
This reverts commit fb86a77cf8.
2021-10-25 10:37:52 -07:00
Benjamin Doherty
c181648bfa Bump version to 1.12.10 2021-10-20 12:15:51 -07:00
Benjamin Doherty
0cf78b3abe Merge branch 'rc/1.12.9' into release 2021-10-20 12:12:33 -07:00
Ben Doherty
22889a7ad9 Fix VirtualMachineEnv.cpp with older JNI versions (#4752) 2021-10-20 11:16:55 -07:00
Ben Doherty
a14451d0ac Fix, call VirtualMachineEnv::JNI_OnLoad for non-Android Java builds (#4749) 2021-10-20 10:41:34 -07:00
Benjamin Doherty
5dfdab10b7 Revert "refactor colorgrading materials"
This reverts commit fb86a77cf8.
2021-10-19 16:13:21 -07:00
Benjamin Doherty
d6f2e3b8e9 Update RELEASE_NOTES for 1.12.9 2021-10-14 18:22:20 -07:00
Benjamin Doherty
df30517743 Bump version to 1.12.9 2021-10-11 11:06:43 -07:00
Benjamin Doherty
8f80643c1a Merge branch 'rc/1.12.8' into release 2021-10-11 11:02:56 -07:00
Benjamin Doherty
5aea9be2fb Update RELEASE_NOTES for 1.12.8 2021-10-07 11:03:45 -07:00
Benjamin Doherty
ad02e483d0 Bump version to 1.12.8 2021-10-04 10:26:57 -07:00
Benjamin Doherty
f463d53036 Merge branch 'rc/1.12.7' into release 2021-10-04 10:21:42 -07:00
Benjamin Doherty
b8d4408524 Update RELEASE_NOTES for 1.12.7 2021-10-04 10:18:56 -07:00
Benjamin Doherty
fef70be848 Bump version to 1.12.7 2021-09-27 11:15:58 -07:00
Benjamin Doherty
bdb12d9b24 Merge branch 'rc/1.12.6' into release 2021-09-27 11:12:42 -07:00
Ben Doherty
43ad283a83 Fix, missing call to setTransparencyMode (#4674) 2021-09-24 13:32:49 -07:00
Benjamin Doherty
2e4936afc4 Revert "fix a race in jobsystem"
This reverts commit 2feb0ad325.
2021-09-24 11:24:44 -07:00
Benjamin Doherty
891ffabd11 Update RELEASE_NOTES for 1.12.6 2021-09-23 17:25:25 -07:00
Ben Doherty
e2c19498b4 Metal: don't call createTextureViewWithsSwizzle directly (#4662) 2021-09-23 11:28:04 -07:00
Ben Doherty
c32630b265 Fix MetalBlitter crash when shader contains warnings (#4663) 2021-09-23 11:27:56 -07:00
Alexey Pelykh
bf21e78d02 Podspec: include headers in nested directories (#4658) 2021-09-22 16:48:12 -07:00
Benjamin Doherty
525d4e08a3 Bump version to 1.12.6 2021-09-20 10:17:18 -07:00
Benjamin Doherty
2e9bf6d694 Merge branch 'rc/1.12.5' into release 2021-09-20 10:14:37 -07:00
Benjamin Doherty
e845f01d85 Update RELEASE_NOTES for 1.12.5 2021-09-20 10:10:56 -07:00
Benjamin Doherty
bef48be7b4 Bump version to 1.12.5 2021-09-13 10:47:09 -07:00
Benjamin Doherty
b54fdc9e6e Merge branch 'rc/1.12.4' into release 2021-09-13 10:43:06 -07:00
Benjamin Doherty
cedbf2e30b Update RELEASE_NOTES for 1.12.4 2021-09-13 10:40:55 -07:00
Benjamin Doherty
592f8d1b0d Bump version to 1.12.4 2021-09-08 11:05:44 -07:00
Benjamin Doherty
29612a684e Merge branch 'rc/1.12.3' into release 2021-09-08 11:03:53 -07:00
Benjamin Doherty
e6d5807399 Bump version to 1.12.3 2021-08-30 11:46:51 -07:00
Benjamin Doherty
fa2553251f Merge branch 'rc/1.12.2' into release 2021-08-30 11:44:00 -07:00
Benjamin Doherty
7387718852 Update RELEASE_NOTES for 1.12.2 2021-08-30 11:41:52 -07:00
Ben Doherty
a503a6209a Fix regression with clipSpaceTransform (#4552) 2021-08-26 14:14:58 -07:00
Ben Doherty
ce3e5f74e8 Fix Metal STREAM buffers (#4543) 2021-08-25 09:48:34 -07:00
Ben Doherty
f37112358e Fix missing bookkeeping in bindUniformBufferRange (#4538) 2021-08-24 09:51:40 -07:00
Benjamin Doherty
f368b14621 Bump version to 1.12.2 2021-08-23 12:41:33 -07:00
Benjamin Doherty
6960b1148a Merge branch 'rc/1.12.1' into release 2021-08-23 12:39:38 -07:00
Benjamin Doherty
3cc23aac25 Update RELEASE_NOTES for 1.12.1 2021-08-23 10:45:59 -07:00
Benjamin Doherty
11dc8740f2 Fix stack-use-after-scope error 2021-08-20 15:40:06 -07:00
Ben Doherty
4b797cff88 Windows: fix incorrect CRT flags with Visual Studio generators (#4516) 2021-08-20 10:56:41 -07:00
Ben Doherty
fe1c1736cd Fix, potential null dereferences in MetalBlitter (#4481) 2021-08-10 16:30:11 -07:00
Benjamin Doherty
4058ef5f09 Bump version to 1.12.1 2021-08-09 12:06:47 -07:00
Benjamin Doherty
d25ca01624 Merge branch 'rc/1.12.0' into release 2021-08-09 12:03:33 -07:00
Benjamin Doherty
d96f87dbbf Update RELEASE_NOTES for 1.12.0 2021-08-09 12:01:46 -07:00
Ben Doherty
8a2e31023f Attempt to fix TSAN failure in ColorGrading.cpp (#4447) 2021-08-05 11:54:07 -07:00
Benjamin Doherty
1ea8e171d9 Bump version to 1.12.0 2021-08-03 11:18:56 -07:00
Benjamin Doherty
e2be3dd0ac Merge branch 'rc/1.11.2' into release 2021-08-03 11:15:19 -07:00
Romain Guy
1c51164e7b Fix inverse tone mapping issues (#4437)
Bring color grading back into the Rec.709 color space to match
previous behaviors. This change also implements an exact inverse
tone map function for the "Filmic" operator.
2021-08-02 20:54:44 -07:00
Benjamin Doherty
f190f03530 Update RELEASE_NOTES for 1.11.2 2021-08-02 11:17:42 -07:00
Ben Doherty
055fc7cbc1 Attempt to fix Windows CI by turning off concurrent builds (#4395) 2021-07-30 14:09:07 -07:00
Ben Doherty
9073fc3dc3 Attempt to fix Windows CI by turning off concurrent builds (#4395) 2021-07-28 10:49:37 -07:00
Ben Doherty
2409dc9bc4 Expose Engine::flush (#4385) 2021-07-26 15:09:14 -07:00
Benjamin Doherty
6586c8d70b Bump version to 1.11.2 2021-07-26 12:50:29 -07:00
Benjamin Doherty
ac0c94da69 Merge branch 'rc/1.11.1' into release 2021-07-26 12:48:36 -07:00
Ben Doherty
d19d6a72b0 Expose Engine::flush (#4385) 2021-07-26 12:10:19 -07:00
Benjamin Doherty
c81b5d98ef Update RELEASE_NOTES for 1.11.1 2021-07-26 10:05:41 -07:00
Mathias Agopian
756866675f fix an issue where a sampler could fail to be updated
SamplerGroup was comparing texture handles to decide if a texture needed
to be updated, however, texture handles are (quickly) recycled and
therefore can't be used for that purpose. e.g. if a texture is destroyed,
its handle could be reused quickly by another texture, if that texture 
is now set on the SamplerGroup, it will ignore it, thinking it's not
different.
2021-07-21 14:46:44 -07:00
Benjamin Doherty
ebcd4925f7 Bump version to 1.11.1 2021-07-19 10:24:01 -07:00
Benjamin Doherty
13bed4fdf9 Merge branch 'rc/1.11.0' into release 2021-07-19 10:20:29 -07:00
Benjamin Doherty
1dae5c6b6c Update RELEASE_NOTES for 1.11.0 2021-07-19 10:18:07 -07:00
Benjamin Doherty
8e6663e4b0 Bump version to 1.11.0 2021-07-12 10:49:43 -07:00
Benjamin Doherty
ba804444b8 Merge branch 'rc/1.10.7' into release 2021-07-12 10:47:12 -07:00
Benjamin Doherty
58cfb85004 Update RELEASE_NOTES for 1.10.7 2021-07-12 10:45:24 -07:00
Mathias Agopian
ab46481b45 Fix colorgrading as subpass
We were inserting the colorgrading subpass command between the
refracted and blended objects, instead of after all of them.

Another bad side effect of this was to trigger the refraction pass for
no reason.
2021-07-12 10:35:01 -07:00
Mathias Agopian
4296782399 don't crash when refraction pass is empty 2021-07-12 09:53:37 -07:00
Benjamin Doherty
ef375a7103 Bump version to 1.10.7 2021-07-07 11:02:30 -07:00
Benjamin Doherty
fd258b7765 Merge branch 'rc/1.10.6' into release 2021-07-07 10:59:12 -07:00
Benjamin Doherty
147de8d372 Update RELEASE_NOTES for 1.10.6 2021-07-07 10:57:25 -07:00
Philip Rideout
eb2a1928b6 ShadowMapManager: fix MSAN use-of-uninitialized-value.
The operator!= in std::array compares SPLIT_COUNT elements, which
is potentially greater than cascadeCount, which was the number of
initialized elements in splitPercentages.
2021-07-07 10:51:18 -07:00
Benjamin Doherty
35b033102f Bump version to 1.10.6 2021-06-28 12:07:58 -07:00
Benjamin Doherty
7bc65421a9 Merge branch 'rc/1.10.5' into release 2021-06-28 12:04:31 -07:00
Benjamin Doherty
736514cf37 Update RELEASE_NOTES for 1.10.5 2021-06-28 12:04:27 -07:00
Mathias Agopian
db0158dae8 fix webgl build
we need to select opengl on webgl.
2021-06-28 08:27:46 -07:00
Benjamin Doherty
e706695ed1 Bump version to 1.10.5 2021-06-21 11:18:26 -07:00
Benjamin Doherty
e8877ffe2d Merge branch 'rc/1.10.4' into release 2021-06-21 11:16:10 -07:00
Benjamin Doherty
1fd5d9dae6 Update RELEASE_NOTES for 1.10.4 2021-06-21 11:14:35 -07:00
Ben Doherty
cd48089318 filament-utils-android: fix string literal conversions (#4137) 2021-06-15 16:19:32 -06:00
Benjamin Doherty
6379ab22c9 Bump version to 1.10.4 2021-06-14 11:06:42 -06:00
Benjamin Doherty
0bf02b75d5 Merge branch 'rc/1.10.3' into release 2021-06-14 11:04:25 -06:00
Benjamin Doherty
c4259b5598 Update RELEASE_NOTES for 1.10.3 2021-06-14 11:02:45 -06:00
Ben Doherty
6b3c1179bc Include sample-gltf-viewer with Android releases (#4099) 2021-06-08 10:41:48 -07:00
Philip Rideout
c1a0e61e8e Fix FixedCapacityVector destructor.
std::allocator::deallocate() expects the same value that was given
during allocate().

Interestingly, this bug did not manifest any issues (even with ASAN) on
some platforms.
2021-06-07 15:56:33 -07:00
Benjamin Doherty
fc06298ed4 Bump version to 1.10.3 2021-06-07 11:22:04 -07:00
Benjamin Doherty
4ca87b188c Merge branch 'rc/1.10.2' into release 2021-06-07 11:19:51 -07:00
Benjamin Doherty
f1f60c3e0d Turn off warnings as errors for spirv-tools 2021-06-07 11:18:56 -07:00
Benjamin Doherty
76d21b56af Update RELEASE_NOTES for 1.10.2 2021-06-07 11:18:09 -07:00
Benjamin Doherty
0ab0e50a4f Turn off warnings as errors for spirv-tools 2021-06-02 11:16:42 -07:00
Benjamin Doherty
34f4c06a5c Bump version to 1.10.2 2021-06-01 11:17:32 -07:00
Benjamin Doherty
6de36f1e53 Merge branch 'rc/1.10.1' into release 2021-06-01 11:15:49 -07:00
Benjamin Doherty
2a9a3b1ac2 Update RELEASE_NOTES for 1.10.1 2021-06-01 11:13:04 -07:00
Benjamin Doherty
84b73a3770 Bump version to 1.10.1 2021-05-24 10:52:02 -07:00
Benjamin Doherty
662a10e273 Merge branch 'rc/1.10.0' into release 2021-05-24 10:49:10 -07:00
Benjamin Doherty
ecce02502e Update RELEASE_NOTES for 1.10.0 2021-05-24 10:47:39 -07:00
Philip Rideout
d17875aea1 ImGuiHelper: fix support for custom images.
The texture binding in the material instance needs to be restored to the
glyph atlas when a custom image is not in use.
2021-05-21 12:49:37 -07:00
Philip Rideout
b8897a68f9 matc: detect missing end brace.
matc was failing to report certain kinds of syntax errors and would
read out-of-bounds memory.

This change casuses the flare material to fail.
2021-05-21 09:20:33 -07:00
Ben Doherty
84efd4871e API CHANGE: remove some Camera, Engine, and View deprecated APIs (#3965) 2021-05-19 12:02:13 -07:00
Benjamin Doherty
85ea5a6b70 Bump version to 1.10.0 2021-05-17 11:00:42 -07:00
Benjamin Doherty
77891acb92 Merge branch 'rc/1.9.25' into release 2021-05-17 10:57:30 -07:00
Benjamin Doherty
74fe102035 Update RELEASE_NOTES for 1.9.25 2021-05-17 10:54:38 -07:00
Philip Rideout
25cc554925 WASM: fix "Missing field" error for lensFlare. 2021-05-17 10:06:56 -07:00
Ben Doherty
d787a521b5 Fix DIST_DIR setting for Windows builds (#3945) 2021-05-12 11:16:30 -07:00
Ben Doherty
46e52c71e1 Fix DIST_DIR setting for Windows builds (#3945) 2021-05-12 11:15:49 -07:00
Philip Rideout
1dad27a172 Repair WebGL and fix potential INVALID_OPERATION.
We should take care not to call glVertexAttribPointer when there is
no bound ARRAY_BUFFER (i.e. when its binding is zero).

This fixes the black screen seen with some WebGL samples after
the recent memory leak fix related to the new BufferObject API.
2021-05-10 13:38:56 -07:00
Benjamin Doherty
60d230b380 Bump version to 1.9.25 2021-05-07 21:42:48 -07:00
Benjamin Doherty
d7cb38e706 Merge branch 'rc/1.9.24' into release 2021-05-07 21:39:17 -07:00
Benjamin Doherty
ce00cca6ee Update RELEASE_NOTES for 1.9.24 2021-05-07 21:36:49 -07:00
Philip Rideout
d627d57bad Second memory leak fix. (#3906)
Fixes #3888.
2021-05-06 14:25:46 -07:00
Philip Rideout
8ffc776f1c Fix horrible memory leak in the GL driver. (#3894)
This leak was introduced in the following PR on April 7.
https://github.com/google/filament/pull/3775

The guilty party has been contacted and properly admonished for his
transgression.

This was tested by adding the following code after applyAnimation in
gltf_viewer.cpp

        static int nframes = 0;
        if (!gpath.empty() && nframes++ > 100) {
            static int count = 0;
            printf("reloading %d\n", count++);
            nframes = 0;
            app.resourceLoader->asyncCancelLoad();
            app.resourceLoader->evictResourceData();
            app.viewer->removeAsset();
            app.assetLoader->destroyAsset(app.asset);
            loadAsset(gpath, app);
            loadResources(gpath, app);
        }
2021-05-04 18:22:08 -07:00
Benjamin Doherty
be032b52c1 Bump version to 1.9.24 2021-05-03 10:42:19 -07:00
Benjamin Doherty
4388e81e5f Merge branch 'rc/1.9.23' into release 2021-05-03 10:40:49 -07:00
Benjamin Doherty
71a185d139 Update RELEASE_NOTES for 1.9.23 2021-05-03 10:38:59 -07:00
Benjamin Doherty
d2cf5985ac Bump version to 1.9.23 2021-04-26 10:55:10 -07:00
Benjamin Doherty
debcbb8e5c Merge branch 'rc/1.9.22' into release 2021-04-26 10:52:43 -07:00
Benjamin Doherty
b9dd62c7d3 Update RELEASE_NOTES for 1.9.22 2021-04-26 10:47:11 -07:00
Benjamin Doherty
dc2b430f34 Bump version to 1.9.22 2021-04-26 10:39:16 -07:00
Benjamin Doherty
e5ef4e8868 Update RELEASE_NOTES for 1.9.21 2021-04-19 11:36:37 -07:00
Benjamin Doherty
c0d6cd3ac3 Merge branch 'rc/1.9.21' into release 2021-04-19 11:34:23 -07:00
Benjamin Doherty
b63ab2dc19 Update RELEASE_NOTES for 1.9.21 2021-04-19 11:32:31 -07:00
Benjamin Doherty
5d8dad561c Bump version to 1.9.21 2021-04-12 11:38:27 -07:00
Benjamin Doherty
8933be1ae2 Merge branch 'rc/1.9.20' into release 2021-04-12 11:36:17 -07:00
Benjamin Doherty
6b66b48b1d Update RELEASE_NOTES for 1.9.20 2021-04-12 11:35:45 -07:00
Benjamin Doherty
9c23eb6e33 Release Filament 1.9.20 2021-04-12 11:34:39 -07:00
Benjamin Doherty
baea54a3fc Update RELEASE_NOTES for 1.9.20 2021-04-12 09:20:31 -07:00
Benjamin Doherty
d9f800454c Bump version to 1.9.20 2021-04-05 11:14:17 -07:00
Benjamin Doherty
f9ee0de07a Merge branch 'rc/1.9.19' into release 2021-04-05 11:11:13 -07:00
Benjamin Doherty
2786d0a9f7 Update RELEASE_NOTES for 1.9.19 2021-04-05 11:10:50 -07:00
Benjamin Doherty
491e8032e6 Release Filament 1.9.19 2021-04-05 11:10:20 -07:00
Benjamin Doherty
ef3f13f5d3 Update RELEASE_NOTES for 1.9.19 2021-04-05 11:09:38 -07:00
Benjamin Doherty
bcb5b2d790 Implement Metal BufferObjects 2021-04-05 11:01:36 -07:00
Ben Doherty
02de3f2e2a Fix, regression with Metal buffers (#3715) 2021-03-29 14:21:34 -07:00
Benjamin Doherty
0e7bb53c07 Bump version to 1.9.19 2021-03-29 11:09:32 -07:00
Benjamin Doherty
759109d478 Merge branch 'rc/1.9.18' into release 2021-03-29 11:06:55 -07:00
Benjamin Doherty
54d5af6edf Update RELEASE_NOTES for 1.9.18 2021-03-29 11:04:57 -07:00
Philip Rideout
38fbe47ced RenderTarget: fix NPE when depth is not present.
This fixes a recent regression that would occur when a RenderTarget
does not have a depth attachment.
2021-03-29 10:32:22 -07:00
Benjamin Doherty
f066c925ba Bump version to 1.9.18 2021-03-22 10:16:34 -07:00
Benjamin Doherty
994fdf4e1d Merge branch 'rc/1.9.17' into release 2021-03-22 10:12:20 -07:00
Benjamin Doherty
50b50d65e3 Update RELEASE_NOTES for 1.9.17 2021-03-22 10:11:02 -07:00
Mathias Agopian
0aaa985649 Fix a hang in JobSystem
The hang was caused by a subtle race. When a job is completed, its 
thread must signal all the threads that might be waiting on this job.
The signaling code was attempting to signal only the minimum number
of threads -- this was important especially in the case where no threads
were waiting, then the call to notify() could be avoided.

Unfortunately, for performance reasons we're not calling notify() with
the condition lock held, this meant that between the time the number of 
waiting threads was latched and the time of the notify() call, more
threads could enter their condition variable wait(), and it would
then be possible for these threads to wake up, instead of the thread
we were trying to wake up (the one waiting on the job).

It would then get stuck forever.

This bug was introduced in 2df639133b


Also add some debugging code for this kind of failure (disabled)
2021-03-18 09:57:25 -07:00
Benjamin Doherty
29564f8eae Bump version to 1.9.17 2021-03-15 10:15:25 -07:00
Benjamin Doherty
c15db68a5b Merge branch 'rc/1.9.16' into release 2021-03-15 10:12:56 -07:00
Benjamin Doherty
3452fb3e56 Update RELEASE_NOTES for 1.9.16 2021-03-15 10:10:40 -07:00
Mathias Agopian
35eb8e7be1 Revert GL backend handle tracking
This wasn't very useful in the first place because we're recycling
handles very quickly. Additionally there was a race condition
which cause false positives.

This reverts commit bc6acd5c5a.
This reverts commit 3a15756c78.
2021-03-12 13:02:55 -08:00
Mathias Agopian
834b774128 Fixes some issues with imported rendertargets
We were not declaring the attachments it was using.

Fixes #3628
2021-03-12 08:26:16 -08:00
Benjamin Doherty
5aa0eb9f9d Bump version to 1.9.16 2021-03-08 10:09:14 -08:00
Benjamin Doherty
d9a6e2e649 Merge branch 'rc/1.9.15' into release 2021-03-08 10:07:00 -08:00
Benjamin Doherty
cb823b16a1 Update RELEASE_NOTES for 1.9.15 2021-03-08 10:05:19 -08:00
Ben Doherty
0bd41e877e Downgrade Linux GitHub Actions environment to fix error (#3588) 2021-03-01 11:21:11 -08:00
Benjamin Doherty
ecc3e73967 Bump version to 1.9.15 2021-03-01 11:13:44 -08:00
Benjamin Doherty
464b4c24f9 Merge branch 'rc/1.9.14' into release 2021-03-01 11:11:53 -08:00
Benjamin Doherty
32367516e8 Update RELEASE_NOTES for 1.9.14 2021-03-01 11:10:06 -08:00
Philip Rideout
1709a55606 filamat: Fix data race with SPIRV error registrations. 2021-02-26 10:34:36 -08:00
Benjamin Doherty
58b4455979 Bump version to 1.9.14 2021-02-22 10:32:56 -08:00
Benjamin Doherty
ea1dede19c Merge branch 'rc/1.9.13' into release 2021-02-22 10:29:57 -08:00
Benjamin Doherty
20ea3381fa Update RELEASE_NOTES for 1.9.13 2021-02-22 10:28:21 -08:00
Philip Rideout
7aa6fccd7c Modernize Python print syntax to appease external codebase. 2021-02-19 11:38:18 -08:00
Philip Rideout
adbd54f4f8 Change abs() to std::abs to appease external codebase. 2021-02-19 11:38:18 -08:00
Philip Rideout
9d54261f18 Move vk_mem_alloc to its own cpp.
This will improve parity between the GitHub repo and its sister codebase
within Google.
2021-02-19 11:38:18 -08:00
Philip Rideout
a97757c9ae Avoid uninitialized reads in computeVisibilityMasks.
I verified that this code is not enabled in our GitHub builds, and I
verified that the MSAN error goes away.
2021-02-19 11:38:18 -08:00
Philip Rideout
7c79d9f89d Remove some Windows line endings.
These line endings cause annoying diffs when comparing Filament's GitHub
source with its twin sister within Google.
2021-02-19 11:38:18 -08:00
Benjamin Doherty
bf0914f813 Bump version to 1.9.13 2021-02-16 10:48:24 -08:00
Benjamin Doherty
b96bc30fbd Merge branch 'rc/1.9.12' into release 2021-02-16 10:44:13 -08:00
Benjamin Doherty
62b50eb8ba Update RELEASE_NOTES for 1.9.12 2021-02-16 10:42:41 -08:00
Ben Doherty
b4932e384a matc: Use consistent params for semantic code analysis (#3524)
When running semantic analysis on a material, we were arbitrarily choosing the first code gen permutation to analyze. So, running matc with arguments --api metal versus --api all would run analysis on slightly different shader code. This causes bugs when flags passed to glslang differ during semantic analysis. This change updates all semantic analysis to always use the same shader code.
2021-02-08 14:06:47 -08:00
Benjamin Doherty
5e68dc5f8d Bump version to 1.9.12 2021-02-08 09:58:15 -08:00
Benjamin Doherty
f222f1b925 Merge branch 'rc/1.9.11' into release 2021-02-08 09:51:08 -08:00
Ben Doherty
22e4a54782 Update RELEASE_NOTES for 1.9.11 2021-02-08 09:50:23 -08:00
Benjamin Doherty
1289922c5f Release Filament 1.9.11 2021-02-08 09:48:50 -08:00
Ben Doherty
2839c352b8 Update RELEASE_NOTES for 1.9.11 2021-02-08 09:47:18 -08:00
Ben Doherty
6a01cbc312 Draft: fix TSAN issue by using lock in ColorGrading constructor (#3510)
This is a second attempt to fix Google3 TSAN failures seen inside of the ColorGrading constructor.

Related first attempt: #3462
2021-02-05 14:06:40 -08:00
Benjamin Doherty
6193156556 Bump version to 1.9.11 2021-02-01 11:49:12 -08:00
Benjamin Doherty
fe23aa917d Merge branch 'rc/1.9.10' into release 2021-02-01 11:45:47 -08:00
Benjamin Doherty
eb8a29a332 Update RELEASE_NOTES for 1.9.10 2021-02-01 11:45:17 -08:00
Ben Doherty
0626902530 Fix sporatic data race warning seen in Google3 (#3462) 2021-01-27 16:40:16 -08:00
Philip Rideout
042bfe2597 Partial fix for MSVC build. 2020-12-14 12:15:06 -08:00
Ben Doherty
97133f3591 Fix Windows iterator issue in Zip2Iterator and StructureOfArrays (#3322) 2020-12-14 12:14:53 -08:00
Philip Rideout
d06cc4390e filamat: minify struct fields.
This shrinks the arm64 so file by 24 KiB.
2020-12-14 11:19:43 -08:00
Philip Rideout
6047d3235f Remove VSM variant from Skybox material. 2020-12-14 11:19:31 -08:00
Philip Rideout
2cda6e35bd math: reduce template bloat for matrices
This does not change our API, it merely reduces the number of
non-inlined function instantiations appearing in non-optimized binaries.
2020-12-14 11:19:22 -08:00
Philip Rideout
8f8d51e17b Code review fixups for libibl_lite. 2020-12-14 11:19:16 -08:00
Philip Rideout
6919e3b274 libibl: use C callback for progress 2020-12-14 11:19:09 -08:00
Philip Rideout
10bf944410 Add libibl_lite. 2020-12-14 11:19:03 -08:00
Philip Rideout
9a2f6fdb53 Vulkan: change vkWaitForFences usage for SwiftShader.
When passing only 1 fence to vkWaitForFences, the `waitAll` argument
should not have any effect, but SwiftShader seems to skip the wait
when this argument is set to VK_FALSE.

More specifically, the failure to wait in `acquireWorkCommandBuffer`
causes the subsequent destruction of an in-use fence, which causes
a TSAN failure with Google's internal tests.

I am consulting with the SwiftShader team on a real fix, meanwhile
we can commit this easy workaround.

We have 5 usages of vkWaitForFences, one of which uses multiple fences
and should have used VK_TRUE anyway.
2020-11-20 09:37:54 -08:00
Philip Rideout
761977d385 Filament should always bind an IBL texture.
This prevents a SwiftShader crash and/or a slew of "no texture bound"
warnings that would appear when the client provides an IBL without
providing reflections texture, which should be a valid thing to do.

Note that it is okay to declare a sampler in GLSL that never gets bound,
as long as it is never sampled from. Since we always sample from the
IBL specular texture, we should always bind something to it.
2020-11-19 13:28:19 -08:00
Ben Doherty
21248f15b5 Fix, matc crash when building mobile materials (#3296) 2020-11-16 14:37:55 -08:00
Benjamin Doherty
4f32817f6d Bump version to 1.9.10 2020-11-16 12:37:49 -08:00
Benjamin Doherty
cc9e05e711 Merge branch 'rc/1.9.9' into release 2020-11-16 12:34:13 -08:00
Benjamin Doherty
419d68d4db Update RELEASE_NOTES for 1.9.9 2020-11-16 12:17:02 -08:00
Philip Rideout
8450232448 Improve the "unbound texture" warnings.
With Vulkan, this warning would sometimes be a false positive. It could
trigger for internal samplers like `ssao` and `structure`, even though
they were not declared in SPIR-V.

With OpenGL, this warning would never be a false positive because it has
the luxury of calling `glGetUniformLocation`.

This adds a private attribute to our samplers called `strict` that
indicates whether or not a sampler should always have a bound texture.
For now the only strict samplers are the custom ones declared in the
user's material.

At some point I think we should consider adding `spirv-reflect` to our
tree to help with problems like this.
2020-11-12 16:21:13 -08:00
Philip Rideout
cc51726590 Vulkan: improve robustness by providing 1x1 fallback. 2020-11-12 10:36:13 -08:00
Philip Rideout
318e22af51 Fix clear behavior with RenderTarget API.
This fixes a bug seen with client applications that use ClearOptions
instead of Skybox, and one or more offscreen RenderTarget objects.
These apps would see junk pixels because Filament would only clear the
first render target in the frame.

The fix is to factor some the flag-setting logic in `beginFrame()` into
a private method, and call this method from `render()` each time
the user-level RenderTarget has been changed.

I wrote a simple C++ demo to reproduce the issue and to verify that
this fix works.
2020-11-11 09:25:36 -08:00
Philip Rideout
68ac87dc24 NOOP backend should not care about GLSL vs SPIRV.
This fixes errors that would occur when using the NOOP backend with
materials that were built without OpenGL support.
2020-11-10 15:44:01 -08:00
Benjamin Doherty
acb8f00075 Bump version to 1.9.9 2020-11-09 09:32:55 -08:00
Benjamin Doherty
06d9183aaa Merge branch 'rc/1.9.8' into release 2020-11-09 09:28:58 -08:00
Benjamin Doherty
75af25419d Update RELEASE_NOTES for 1.9.8 2020-11-09 09:26:27 -08:00
Benjamin Doherty
f6b90d2a31 Bump version to 1.9.8 2020-11-09 09:21:36 -08:00
Philip Rideout
a3822f4af0 Fix FENCE_WAIT_FOR_EVER in Linux.
The number of infinite nanoseconds was negative because we asked
chrono for a signed integer, so "wait forever" really meant "do not
wait at all".
2020-11-02 16:12:53 -08:00
Benjamin Doherty
bcdad769ff Merge branch 'rc/1.9.7' into release 2020-11-02 11:03:57 -07:00
Benjamin Doherty
be4fb4fdbb Update RELEASE_NOTES for 1.9.7 2020-11-02 10:59:19 -07:00
Benjamin Doherty
65394f6301 Bump version to 1.9.7 2020-10-26 11:34:20 -06:00
Benjamin Doherty
b0beee03bc Merge branch 'rc/1.9.6' into release 2020-10-26 11:29:45 -06:00
Benjamin Doherty
fe1de41b8e Update RELEASE_NOTES for 1.9.6 2020-10-26 11:25:31 -06:00
Benjamin Doherty
a37b431e87 Bump version to 1.9.6 2020-10-19 11:55:13 -06:00
Benjamin Doherty
98107016b9 Merge branch 'rc/1.9.5' into release 2020-10-19 11:51:30 -06:00
Benjamin Doherty
8bccfc2863 Update RELEASE_NOTES for 1.9.5 2020-10-19 11:49:05 -06:00
Benjamin Doherty
f54a0a3452 Fix CocoaPod version 2020-10-13 15:15:02 -06:00
Benjamin Doherty
6778ab0624 Fix CocoaPod version 2020-10-13 15:09:59 -06:00
Benjamin Doherty
269d636785 Bump version to 1.9.5 2020-10-12 12:03:29 -06:00
Benjamin Doherty
39862c91ce Merge branch 'rc/1.9.4' into release 2020-10-12 11:56:24 -06:00
Benjamin Doherty
523f4026b4 Update RELEASE_NOTES for 1.9.4 2020-10-12 11:52:01 -06:00
Benjamin Doherty
a6bf162431 Bump version to 1.9.4 2020-10-07 16:06:23 -06:00
Benjamin Doherty
826e8d181c Merge branch 'rc/1.9.3' into release 2020-10-05 11:36:16 -06:00
Benjamin Doherty
16dfadbba0 Update RELEASE_NOTES for 1.9.3 2020-10-05 11:29:36 -06:00
Benjamin Doherty
5cbb97551f Bump version to 1.9.3 2020-09-28 11:40:31 -06:00
Benjamin Doherty
defee767c3 Merge branch 'rc/1.9.2' into release 2020-09-28 11:27:47 -06:00
Benjamin Doherty
9560318521 Update RELEASE_NOTES for 1.9.2 2020-09-28 11:26:21 -06:00
Benjamin Doherty
ef09feb048 Bump version to 1.9.2 2020-09-21 11:16:53 -06:00
Benjamin Doherty
39f323fe09 Merge branch 'rc/1.9.1' into release 2020-09-21 11:00:07 -06:00
Benjamin Doherty
11b95304ea Merge branch 'release' into rc/1.9.1 2020-09-21 10:59:52 -06:00
Benjamin Doherty
b7c30a7916 Update RELEASE_NOTES for 1.9.1 2020-09-21 10:59:06 -06:00
Benjamin Doherty
4cae48fc77 Bump version to 1.9.1 2020-09-14 11:51:36 -07:00
Benjamin Doherty
d1a93f0557 Update release notes for 1.9.0 2020-09-14 10:54:28 -07:00
Benjamin Doherty
b93059fad7 Bump version to 1.9.0 2020-09-08 10:47:23 -07:00
792 changed files with 35442 additions and 29385 deletions

7
.dir-locals.el Normal file
View File

@@ -0,0 +1,7 @@
;;; Directory Local Variables -*- no-byte-compile: t -*-
;;; For more information see (info "(emacs) Directory Variables")
((c++-mode . ((c-file-style . "filament")
(apheleia-inhibit . t)))
(c-mode . ((c-file-style . "filament")
(apheleia-inhibit . t))))

11
.editorconfig Normal file
View File

@@ -0,0 +1,11 @@
root = true
[*]
end_of_line = lf
insert_final_newline = true
charset = utf-8
[*.{c,cpp,h,inc,kt,java,js,md}]
indent_style = space
indent_size = 4
max_line_length = 100

View File

@@ -10,7 +10,7 @@ on:
jobs:
build-android:
name: build-android
runs-on: macos-latest
runs-on: macos-14
steps:
- uses: actions/checkout@v3.3.0
@@ -29,10 +29,6 @@ jobs:
with:
name: filamat-android-full
path: out/filamat-android-release.aar
- uses: actions/upload-artifact@v1.0.0
with:
name: filamat-android-lite
path: out/filamat-android-lite-release.aar
- uses: actions/upload-artifact@v1.0.0
with:
name: gltfio-android-release

30
.github/workflows/cocopods-deploy.yml vendored Normal file
View File

@@ -0,0 +1,30 @@
name: CocoaPods Deploy
# This must be run after the iOS release job has finished, and the iOS release
# asset has been uploaded to Github.
on:
workflow_dispatch:
inputs:
release_tag:
description: 'Release tag to deploy (e.g., v1.42.2)'
required: true
default: 'v1.42.2'
jobs:
cocoapods-deploy:
name: cocoapods-deploy
runs-on: macos-14
steps:
- name: Check out iOS/CocoaPods directory
uses: Bhacaz/checkout-files@49fc3050859046bf4f4873678d46099985640e89
with:
files: ios/CocoaPods
token: ${{ secrets.GITHUB_TOKEN }}
branch: ${{ github.event.inputs.release_tag }}
- name: Move podspec to root
run: mv ios/CocoaPods/*.podspec .
- name: Install CocoaPods
run: gem install cocoapods
- uses: michaelhenry/deploy-to-cocoapods-github-action@745686ab065f90596e0d5cfcf97bb2416d94262e
env:
COCOAPODS_TRUNK_TOKEN: ${{ secrets.COCOAPODS_TRUNK_TOKEN }}

View File

@@ -10,7 +10,7 @@ on:
jobs:
build-ios:
name: build-ios
runs-on: macos-latest
runs-on: macos-14
steps:
- uses: actions/checkout@v3.3.0

View File

@@ -10,7 +10,7 @@ on:
jobs:
build-mac:
name: build-mac
runs-on: macos-latest
runs-on: macos-14
steps:
- uses: actions/checkout@v3.3.0

33
.github/workflows/npm-deploy.yml vendored Normal file
View File

@@ -0,0 +1,33 @@
name: Npm Deploy
on:
workflow_dispatch:
inputs:
release_tag:
description: 'Release tag to deploy (e.g., v1.42.2)'
required: true
default: 'v1.42.2'
jobs:
npm-deploy:
name: npm-deploy
runs-on: macos-14
steps:
- uses: actions/checkout@v3.3.0
with:
ref: ${{ github.event.inputs.release_tag }}
# Setup .npmrc file to publish to npm
- uses: actions/setup-node@v3
with:
node-version: '18.x'
registry-url: 'https://registry.npmjs.org'
- name: Run build script
run: |
cd build/web && printf "y" | ./build.sh release
- name: Deploy to npm
run: |
cd out/cmake-webgl-release/web/filament-js
npm publish
env:
NODE_AUTH_TOKEN: ${{ secrets.NPM_TOKEN }}

View File

@@ -15,7 +15,7 @@ jobs:
strategy:
matrix:
os: [macos-latest, ubuntu-22.04]
os: [macos-14, ubuntu-22.04]
steps:
- uses: actions/checkout@v3.3.0
@@ -40,7 +40,7 @@ jobs:
build-android:
name: build-android
runs-on: macos-latest
runs-on: macos-14
steps:
- uses: actions/checkout@v3.3.0
@@ -54,7 +54,7 @@ jobs:
build-ios:
name: build-iOS
runs-on: macos-latest
runs-on: macos-14
steps:
- uses: actions/checkout@v3.3.0
@@ -67,7 +67,7 @@ jobs:
build-web:
name: build-web
runs-on: macos-latest
runs-on: macos-14
steps:
- uses: actions/checkout@v3.3.0

View File

@@ -31,7 +31,7 @@ jobs:
strategy:
matrix:
os: [macos-latest, ubuntu-22.04]
os: [macos-14, ubuntu-22.04]
steps:
- name: Decide Git ref
@@ -65,7 +65,7 @@ jobs:
build-web:
name: build-web
runs-on: macos-latest
runs-on: macos-14
if: github.event_name == 'release' || github.event.inputs.platform == 'web'
steps:
@@ -98,7 +98,7 @@ jobs:
build-android:
name: build-android
runs-on: macos-latest
runs-on: macos-14
if: github.event_name == 'release' || github.event.inputs.platform == 'android'
steps:
@@ -124,13 +124,12 @@ jobs:
cd ../..
mv out/filament-android-release.aar out/filament-${TAG}-android.aar
mv out/filamat-android-release.aar out/filamat-${TAG}-android.aar
mv out/filamat-android-lite-release.aar out/filamat-${TAG}-lite-android.aar
mv out/gltfio-android-release.aar out/gltfio-${TAG}-android.aar
mv out/filament-utils-android-release.aar out/filament-utils-${TAG}-android.aar
- name: Sign sample-gltf-viewer
run: |
echo "${APK_KEYSTORE_BASE64}" > filament.jks.base64
base64 --decode filament.jks.base64 > filament.jks
base64 --decode -i filament.jks.base64 > filament.jks
BUILD_TOOLS_VERSION=$(ls ${ANDROID_HOME}/build-tools | sort -V | tail -n 1)
APKSIGNER=${ANDROID_HOME}/build-tools/${BUILD_TOOLS_VERSION}/apksigner
IN_FILE="out/sample-gltf-viewer-release.apk"
@@ -153,7 +152,7 @@ jobs:
build-ios:
name: build-ios
runs-on: macos-latest
runs-on: macos-14
if: github.event_name == 'release' || github.event.inputs.platform == 'ios'
steps:
@@ -206,8 +205,7 @@ jobs:
TAG: ${{ steps.git_ref.outputs.tag }}
run: |
build\windows\build-github.bat release
cd ..\..
move out\filament-windows.tgz out\filament-%TAG%-windows.tgz
move out\filament-windows.tgz out\filament-$Env:TAG-windows.tgz
shell: cmd
- uses: actions/github-script@v6
env:

View File

@@ -10,7 +10,7 @@ on:
jobs:
build-web:
name: build-web
runs-on: macos-latest
runs-on: macos-14
steps:
- uses: actions/checkout@v3.3.0

2
.gitignore vendored
View File

@@ -16,3 +16,5 @@ settings.json
test*.png
test*.json
results
/compile_commands.json
/.cache

View File

@@ -40,25 +40,27 @@ inside the Filament source tree.
To trigger an incremental debug build:
```
$ ./build.sh debug
```shell
./build.sh debug
```
To trigger an incremental release build:
```
$ ./build.sh release
```shell
./build.sh release
```
To trigger both incremental debug and release builds:
```
$ ./build.sh debug release
```shell
./build.sh debug release
```
If build fails for some reasons, it may leave the `out/` directory in a broken state. You can
force a clean build by adding the `-c` flag in that case.
To install the libraries and executables in `out/debug/` and `out/release/`, add the `-i` flag.
You can force a clean build by adding the `-c` flag. The script offers more features described
by executing `build.sh -h`.
The script offers more features described by executing `build.sh -h`.
### Filament-specific CMake Options
@@ -76,9 +78,9 @@ The following CMake options are boolean options specific to Filament:
To turn an option on or off:
```
$ cd <cmake-build-directory>
$ cmake . -DOPTION=ON # Replace OPTION with the option name, set to ON / OFF
```shell
cd <cmake-build-directory>
cmake . -DOPTION=ON # Replace OPTION with the option name, set to ON / OFF
```
Options can also be set with the CMake GUI.
@@ -102,38 +104,38 @@ script.
If you'd like to run `cmake` directly rather than using the build script, it can be invoked as
follows, with some caveats that are explained further down.
```
$ mkdir out/cmake-release
$ cd out/cmake-release
$ cmake -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=../release/filament ../..
```shell
mkdir out/cmake-release
cd out/cmake-release
cmake -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=../release/filament ../..
```
Your Linux distribution might default to `gcc` instead of `clang`, if that's the case invoke
`cmake` with the following command:
```
$ mkdir out/cmake-release
$ cd out/cmake-release
```shell
mkdir out/cmake-release
cd out/cmake-release
# Or use a specific version of clang, for instance /usr/bin/clang-14
$ CC=/usr/bin/clang CXX=/usr/bin/clang++ CXXFLAGS=-stdlib=libc++ \
cmake -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=../release/filament ../..
CC=/usr/bin/clang CXX=/usr/bin/clang++ CXXFLAGS=-stdlib=libc++ \
cmake -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=../release/filament ../..
```
You can also export the `CC` and `CXX` environment variables to always point to `clang`. Another
solution is to use `update-alternatives` to both change the default compiler, and point to a
specific version of clang:
```
$ update-alternatives --install /usr/bin/clang clang /usr/bin/clang-14 100
$ update-alternatives --install /usr/bin/clang++ clang++ /usr/bin/clang++-14 100
$ update-alternatives --install /usr/bin/cc cc /usr/bin/clang 100
$ update-alternatives --install /usr/bin/c++ c++ /usr/bin/clang++ 100
```shell
update-alternatives --install /usr/bin/clang clang /usr/bin/clang-14 100
update-alternatives --install /usr/bin/clang++ clang++ /usr/bin/clang++-14 100
update-alternatives --install /usr/bin/cc cc /usr/bin/clang 100
update-alternatives --install /usr/bin/c++ c++ /usr/bin/clang++ 100
```
Finally, invoke `ninja`:
```
$ ninja
```shell
ninja
```
This will build Filament, its tests and samples, and various host tools.
@@ -143,8 +145,8 @@ This will build Filament, its tests and samples, and various host tools.
To compile Filament you must have the most recent version of Xcode installed and you need to
make sure the command line tools are setup by running:
```
$ xcode-select --install
```shell
xcode-select --install
```
If you wish to run the Vulkan backend instead of the default Metal backend, you must install
@@ -152,11 +154,11 @@ the LunarG SDK, enable "System Global Components", and reboot your machine.
Then run `cmake` and `ninja` to trigger a build:
```
$ mkdir out/cmake-release
$ cd out/cmake-release
$ cmake -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=../release/filament ../..
$ ninja
```shell
mkdir out/cmake-release
cd out/cmake-release
cmake -G Ninja -DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=../release/filament ../..
ninja
```
### iOS
@@ -164,20 +166,20 @@ $ ninja
The easiest way to build Filament for iOS is to use `build.sh` and the
`-p ios` flag. For instance to build the debug target:
```
$ ./build.sh -p ios debug
```shell
./build.sh -p ios debug
```
See [ios/samples/README.md](./ios/samples/README.md) for more information.
### Windows
#### Building on Windows with Visual Studio 2019
#### Building on Windows with Visual Studio 2019 or later
Install the following components:
- [Visual Studio 2019](https://www.visualstudio.com/downloads)
- [Windows 10 SDK](https://developer.microsoft.com/en-us/windows/downloads/windows-10-sdk)
- [Visual Studio 2019 or later](https://www.visualstudio.com/downloads)
- [Windows SDK](https://developer.microsoft.com/en-us/windows/downloads/windows-sdk/)
- [Python 3.7](https://www.python.org/ftp/python/3.7.0/python-3.7.0.exe)
- [CMake 3.14 or later](https://github.com/Kitware/CMake/releases/download/v3.14.7/cmake-3.14.7-win64-x64.msi)
@@ -191,10 +193,10 @@ using `fsutil.exe file queryCaseSensitiveInfo`.
Next, open `x64 Native Tools Command Prompt for VS 2019`, create a working directory, and run
CMake in it:
```
> mkdir out
> cd out
> cmake ..
```bat
mkdir out
cd out
cmake ..
```
Open the generated solution file `TNT.sln` in Visual Studio.
@@ -204,15 +206,15 @@ target in the _Solution Explorer_ and choose _Build_ to build a specific target.
For example, build the `material_sandbox` sample and run it from the `out` directory with:
```
> samples\Debug\material_sandbox.exe ..\assets\models\monkey\monkey.obj
```bat
samples\Debug\material_sandbox.exe ..\assets\models\monkey\monkey.obj
```
You can also use CMake to invoke the build without opening Visual Studio. For example, from the
`out` folder run the following command.
```
> cmake --build . --target gltf_viewer --config Release
```bat
cmake --build . --target gltf_viewer --config Release
```
### Android
@@ -237,8 +239,8 @@ To build Android on Windows machines, see [android/Windows.md](android/Windows.m
The easiest way to build Filament for Android is to use `build.sh` and the
`-p android` flag. For instance to build the release target:
```
$ ./build.sh -p android release
```shell
./build.sh -p android release
```
Run `build.sh -h` for more information.
@@ -248,23 +250,23 @@ Run `build.sh -h` for more information.
Invoke CMake in a build directory of your choice, inside of filament's directory. The commands
below show how to build Filament for ARM 64-bit (`aarch64`).
```
$ mkdir out/android-build-release-aarch64
$ cd out/android-build-release-aarch64
$ cmake -G Ninja -DCMAKE_TOOLCHAIN_FILE=../../build/toolchain-aarch64-linux-android.cmake \
-DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=../android-release/filament ../..
```shell
mkdir out/android-build-release-aarch64
cd out/android-build-release-aarch64
cmake -G Ninja -DCMAKE_TOOLCHAIN_FILE=../../build/toolchain-aarch64-linux-android.cmake \
-DCMAKE_BUILD_TYPE=Release -DCMAKE_INSTALL_PREFIX=../android-release/filament ../..
```
And then invoke `ninja`:
```
$ ninja install
```shell
ninja install
```
or
```
$ ninja install/strip
```shell
ninja install/strip
```
This will generate Filament's Android binaries in `out/android-release`. This location is important
@@ -296,8 +298,8 @@ AAR.
Alternatively you can build the AAR from the command line by executing the following in the
`android/` directory:
```
$ ./gradlew -Pcom.google.android.filament.dist-dir=../../out/android-release/filament assembleRelease
```shell
./gradlew -Pcom.google.android.filament.dist-dir=../../out/android-release/filament assembleRelease
```
The `-Pcom.google.android.filament.dist-dir` can be used to specify a different installation
@@ -311,7 +313,7 @@ sure to add the newly created module as a dependency to your application.
If you do not wish to include all supported ABIs, make sure to create the appropriate flavors in
your Gradle build file. For example:
```
```gradle
flavorDimensions 'cpuArch'
productFlavors {
arm8 {
@@ -353,7 +355,7 @@ started, follow the instructions for building Filament on your platform ([macOS]
Next, you need to install the Emscripten SDK. The following instructions show how to install the
same version that our continuous builds use.
```
```shell
cd <your chosen parent folder for the emscripten SDK>
curl -L https://github.com/emscripten-core/emsdk/archive/refs/tags/3.1.15.zip > emsdk.zip
unzip emsdk.zip ; mv emsdk-* emsdk ; cd emsdk
@@ -364,7 +366,7 @@ source ./emsdk_env.sh
After this you can invoke the [easy build](#easy-build) script as follows:
```
```shell
export EMSDK=<your chosen home for the emscripten SDK>
./build.sh -p webgl release
```
@@ -374,7 +376,7 @@ creates a `samples` folder that can be used as the root of a simple static web s
cannot open the HTML directly from the filesystem due to CORS. We recommend using the emrun tool
to create a quick localhost server:
```
```shell
emrun out/cmake-webgl-release/web/samples --no_browser --port 8000
```
@@ -395,7 +397,7 @@ Some of the samples accept FBX/OBJ meshes while others rely on the `filamesh` fi
generate a `filamesh ` file from an FBX/OBJ asset, run the `filamesh` tool
(`./tools/filamesh/filamesh` in your build directory):
```
```shell
filamesh ./assets/models/monkey/monkey.obj monkey.filamesh
```
@@ -405,7 +407,7 @@ files for the IBL (which are PNGs containing `R11F_G11F_B10F` data) or a path to
containing two `.ktx` files (one for the IBL itself, one for the skybox). To generate an IBL
simply use this command:
```
```shell
cmgen -f ktx -x ./ibls/ my_ibl.exr
```
@@ -427,9 +429,9 @@ value is the desired roughness between 0 and 1.
To generate the documentation you must first install `doxygen` and `graphviz`, then run the
following commands:
```
$ cd filament/filament
$ doxygen docs/doxygen/filament.doxygen
```shell
cd filament/filament
doxygen docs/doxygen/filament.doxygen
```
Finally simply open `docs/html/index.html` in your web browser.
@@ -439,7 +441,7 @@ Finally simply open `docs/html/index.html` in your web browser.
To try out Filament's Vulkan support with SwiftShader, first build SwiftShader and set the
`SWIFTSHADER_LD_LIBRARY_PATH` variable to the folder that contains `libvk_swiftshader.dylib`:
```
```shell
git clone https://github.com/google/swiftshader.git
cd swiftshader/build
cmake .. && make -j
@@ -454,7 +456,7 @@ Continuous testing turnaround can be quite slow if you need to build SwiftShader
provide an Ubuntu-based Docker image that has it already built. The Docker image also includes
everything necessary for building Filament. You can fetch and run the image as follows:
```
```shell
docker pull ghcr.io/filament-assets/swiftshader
docker run -it ghcr.io/filament-assets/swiftshader
```

View File

@@ -3,6 +3,14 @@
# ==================================================================================================
cmake_minimum_required(VERSION 3.19)
# ==================================================================================================
# Toolchain configuration
# ==================================================================================================
if (APPLE AND NOT IOS)
# This must be set before project() is called
set(CMAKE_OSX_DEPLOYMENT_TARGET 10.15 CACHE STRING "")
endif()
# ==================================================================================================
# Project declaration
# ==================================================================================================
@@ -33,6 +41,10 @@ option(FILAMENT_LINUX_IS_MOBILE "Treat Linux as Mobile" OFF)
option(FILAMENT_ENABLE_ASAN_UBSAN "Enable Address and Undefined Behavior Sanitizers" OFF)
option(FILAMENT_ENABLE_TSAN "Enable Thread Sanitizer" OFF)
option(FILAMENT_ENABLE_FEATURE_LEVEL_0 "Enable Feature Level 0" ON)
set(FILAMENT_NDK_VERSION "" CACHE STRING
"Android NDK version or version prefix to be used when building for Android."
)
@@ -57,6 +69,9 @@ set(FILAMENT_METAL_HANDLE_ARENA_SIZE_IN_MB "8" CACHE STRING
"Size of the Metal handle arena, default 8."
)
# Enable exceptions by default in spirv-cross.
set(SPIRV_CROSS_EXCEPTIONS_TO_ASSERTIONS OFF)
# ==================================================================================================
# CMake policies
# ==================================================================================================
@@ -217,6 +232,21 @@ if (WIN32)
# we don't need them on CI.
string(REPLACE "/INCREMENTAL" "/INCREMENTAL:NO" ${LinkerFlag} ${${LinkerFlag}})
endforeach()
# We turn off compile-time optimizations for CI, as options that speed up the compile-time
# (e.g. /MP) might increase memory usage, leading to instabilities on limited CI machines.
option(FILAMENT_SHORTEN_MSVC_COMPILATION "Shorten compile-time in Visual Studio" OFF)
else()
option(FILAMENT_SHORTEN_MSVC_COMPILATION "Shorten compile-time in Visual Studio" ON)
endif()
if (MSVC)
if (FILAMENT_SHORTEN_MSVC_COMPILATION)
# enable multi-processor compilation
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} /MP")
# disable run-time STL checks to improve tools (e.g. matc) performance
set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} /D_ITERATOR_DEBUG_LEVEL=0")
endif()
endif()
endif()
@@ -329,6 +359,7 @@ endif()
if (CYGWIN)
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -fno-exceptions -fno-rtti")
set(SPIRV_CROSS_EXCEPTIONS_TO_ASSERTIONS ON)
endif()
if (MSVC)
@@ -365,6 +396,7 @@ endif()
# saved by -fno-exception and 10 KiB saved by -fno-rtti).
if (ANDROID OR IOS OR WEBGL)
set(CMAKE_CXX_FLAGS_RELEASE "${CMAKE_CXX_FLAGS_RELEASE} -fno-exceptions -fno-rtti")
set(SPIRV_CROSS_EXCEPTIONS_TO_ASSERTIONS ON)
if (ANDROID OR WEBGL)
# Omitting unwind info prevents the generation of readable stack traces in crash reports on iOS
@@ -376,6 +408,7 @@ endif()
# std::visit, which is not supported on iOS 11.0 when exceptions are enabled.
if (IOS)
set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -fno-exceptions")
set(SPIRV_CROSS_EXCEPTIONS_TO_ASSERTIONS ON)
endif()
# With WebGL, we disable RTTI even for debug builds because we pass emscripten::val back and forth
@@ -395,7 +428,13 @@ endif()
if (FILAMENT_ENABLE_ASAN_UBSAN)
set(EXTRA_SANITIZE_OPTIONS "-fsanitize=address -fsanitize=undefined")
endif()
if (FILAMENT_ENABLE_TSAN)
set(EXTRA_SANITIZE_OPTIONS "-fsanitize=thread")
endif()
if (ANDROID)
# keep STL debug infos (mimics what the NDK does)
set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -fno-limit-debug-info")
endif()
if (NOT MSVC AND NOT WEBGL)
set(CMAKE_CXX_FLAGS_DEBUG "${CMAKE_CXX_FLAGS_DEBUG} -fstack-protector")
endif()
@@ -515,6 +554,11 @@ if (FILAMENT_SUPPORTS_METAL)
set(MATC_API_FLAGS ${MATC_API_FLAGS} -a metal)
endif()
# Disable ESSL 1.0 code generation.
if (NOT FILAMENT_ENABLE_FEATURE_LEVEL_0)
set(MATC_API_FLAGS ${MATC_API_FLAGS} -1)
endif()
# Enable debug info (preserves names in SPIR-V)
if (FILAMENT_ENABLE_MATDBG)
set(MATC_OPT_FLAGS ${MATC_OPT_FLAGS} -d)
@@ -590,9 +634,9 @@ function(combine_static_libs TARGET OUTPUT DEPS)
# Loop through the dependent libraries and query their location on disk.
set(DEPS_FILES )
foreach(DEPENDENCY ${DEPS})
if(TARGET ${DEPENDENCY})
if (TARGET ${DEPENDENCY})
get_property(dep_type TARGET ${DEPENDENCY} PROPERTY TYPE)
if(dep_type STREQUAL "STATIC_LIBRARY")
if (dep_type STREQUAL "STATIC_LIBRARY")
list(APPEND DEPS_FILES "$<TARGET_FILE:${DEPENDENCY}>")
endif()
endif()
@@ -662,7 +706,7 @@ function(get_resgen_vars ARCHIVE_DIR ARCHIVE_NAME)
set(RESGEN_OUTPUTS "${OUTPUTS}" PARENT_SCOPE)
set(RESGEN_FLAGS -qx ${ARCHIVE_DIR} -p ${ARCHIVE_NAME} PARENT_SCOPE)
set(RESGEN_SOURCE "${ARCHIVE_DIR}/${ARCHIVE_NAME}${ASM_SUFFIX}.S" PARENT_SCOPE)
set(RESGEN_SOURCE_FLAGS "-I${ARCHIVE_DIR} ${ASM_ARCH_FLAG}" PARENT_SCOPE)
set(RESGEN_SOURCE_FLAGS "-I'${ARCHIVE_DIR}' ${ASM_ARCH_FLAG}" PARENT_SCOPE)
endif()
endfunction()
@@ -677,7 +721,6 @@ add_subdirectory(${LIBRARIES}/filabridge)
add_subdirectory(${LIBRARIES}/filaflat)
add_subdirectory(${LIBRARIES}/filagui)
add_subdirectory(${LIBRARIES}/filameshio)
add_subdirectory(${LIBRARIES}/geometry)
add_subdirectory(${LIBRARIES}/gltfio)
add_subdirectory(${LIBRARIES}/ibl)
add_subdirectory(${LIBRARIES}/iblprefilter)
@@ -705,6 +748,9 @@ add_subdirectory(${EXTERNAL}/jsmn/tnt)
add_subdirectory(${EXTERNAL}/stb/tnt)
add_subdirectory(${EXTERNAL}/getopt)
# Note that this has to be placed after mikktspace in order for combine_static_libs to work.
add_subdirectory(${LIBRARIES}/geometry)
if (FILAMENT_BUILD_FILAMAT OR IS_HOST_PLATFORM)
# spirv-tools must come before filamat, as filamat relies on the presence of the
# spirv-tools_SOURCE_DIR variable.
@@ -722,6 +768,8 @@ endif()
if (FILAMENT_SUPPORTS_VULKAN)
add_subdirectory(${LIBRARIES}/bluevk)
add_subdirectory(${EXTERNAL}/vkmemalloc/tnt)
set(SPIRV_HEADERS_SKIP_EXAMPLES ON)
add_subdirectory(${EXTERNAL}/spirv-headers)
endif()
set(FILAMENT_SAMPLES_BINARY_DIR ${PROJECT_BINARY_DIR}/samples)

View File

@@ -27,7 +27,7 @@ again.
## Code Style
See [CodeStyle.md](/CODE_STYLE.md)
See [CODE_STYLE.md](/CODE_STYLE.md)
## Code reviews

View File

@@ -7,9 +7,3 @@ for next branch cut* header.
appropriate header in [RELEASE_NOTES.md](./RELEASE_NOTES.md).
## Release notes for next branch cut
- Fix possible NPE when updating fog options from Java/Kotlin
- The `emissive` property was not applied properly to `MASKED` materials, and could cause
dark fringes to appear (recompile materials)
- Allow glTF materials with transmission/volume extensions to choose their alpha mode
instead of forcing `MASKED`

View File

@@ -31,7 +31,7 @@ repositories {
}
dependencies {
implementation 'com.google.android.filament:filament-android:1.42.0'
implementation 'com.google.android.filament:filament-android:1.50.3'
}
```
@@ -50,8 +50,8 @@ Here are all the libraries available in the group `com.google.android.filament`:
iOS projects can use CocoaPods to install the latest release:
```
pod 'Filament', '~> 1.42.0'
```shell
pod 'Filament', '~> 1.50.3'
```
### Snapshots

View File

@@ -128,3 +128,15 @@ Navigate to [Filament's release
workflow](https://github.com/google/filament/actions/workflows/release.yml). Hit the _Run workflow_
dropdown. Modify _Platform to build_ and _Release tag to build_, then hit _Run workflow_. This will
initiate a new release run.
## 11. Kick off the npm and CocoaPods release jobs
Navigate to [Filament's npm deploy
workflow](https://github.com/google/filament/actions/workflows/npm-deploy.yml).
Hit the _Run workflow_ dropdown. Modify _Release tag to deploy_ to the tag corresponding to this
release (for example, v1.42.2).
Navigate to [Filament's CocoaPods deploy
workflow](https://github.com/google/filament/actions/workflows/cocopods-deploy.yml).
Hit the _Run workflow_ dropdown. Modify _Release tag to deploy_ to the tag corresponding to this
release (for example, v1.42.2).

View File

@@ -7,8 +7,108 @@ A new header is inserted each time a *tag* is created.
Instead, if you are authoring a PR for the main branch, add your release note to
[NEW_RELEASE_NOTES.md](./NEW_RELEASE_NOTES.md).
## v1.50.3
## v1.50.2
## v1.50.1
- Metal: fix some shader artifacts by disabling fast math optimizations.
- backend: remove `atan2` overload which had a typo and wasn't useful. Fixes b/320856413.
- utils: remove usages of `SpinLock`. Fixes b/321101014.
## v1.50.0
- engine: TAA now supports 4x upscaling [BETA] [⚠️ **New Material Version**]
## v1.49.3
- matc: Generate stereo variants for FL0 materials [⚠️ **Recompile materials**]
## v1.49.2
## v1.49.1
## v1.49.0
- matc: Fix ESSL 1.0 codegen when using external samplers [⚠️ **Recompile materials**]
## v1.48.0
- matc: New option `-1` to disable generation of ESSL 1.0 code in Feature Level 0 materials
- matc: Support optimizations for ESSL 1.0 code [⚠️ **Recompile materials**]
## v1.47.0
- engine: Support up to 4 side-by-side stereoscopic eyes, configurable at Engine creation time. See
`Engine::Config::stereoscopicEyeCount`. [⚠️ **Recompile Materials**]
## v1.46.0
- engine: Allow instantiating Engine at a given feature level via `Engine::Builder::featureLevel`
- matc: Enable `GL_OES_standard_derivatives` extension in ESSL 1.0 shaders
- matc: Fix code generation of double sided and masked materials in ESSL 1.0 shaders
- filagui: Add support for feature level 0
- matc: Add support for post-process materials in feature level 0
- engine: Add `Material::getFeatureLevel()`
- engine: Add missing `Material::getReflectionMode()` method in Java
- engine: Support basic usage of post-processing materials on feature level 0
- engine: Fix critical GLES 2.0 bugs
- engine: Add `FILAMENT_ENABLE_FEATURE_LEVEL_0` build-time option optionally allow building Filament
without FL0 support.
## v1.45.1
- engine: Added parameter for configuring JobSystem thread count
- engine: In Java, introduce Engine.Builder
- gltfio: fix ubershader index for transmission&volume material
- engine: New tone mapper: `AgXTonemapper`.
- matinfo: Add support for viewing ESSL 1.0 shaders
- engine: Add `Renderer::getClearOptions()` [b/243846268]
- engine: Fix stable shadows (again) when an IBL rotation is used
## v1.45.0
- materials: fix alpha masked materials when MSAA is turned on [⚠️ **Recompile materials**]
- materials: better support materials with custom depth [**Recompile Materials**]
- engine: fade shadows at shadowFar distance instead of hard cutoff [⚠️ **New Material Version**]
## v1.44.0
- engine: add support for skinning with more than four bones per vertex.
- engine: remove `BloomOptions::anamorphism` which wasn't working well in most cases [**API CHANGE**]
- engine: new API to return a Material's supported variants, C++ only (b/297456590)
- build: fix emscripten-1.3.46 build
- engine: materials built for feature level 0 can now also be loaded in higher feature levels [⚠️
**New Material Version**]
## v1.43.1
## v1.43.0
- gltfio: Fix possible change of scale sign when decomposing transform matrix for animation
- engine: Fixes "stable" shadows (see b/299310624)
## v1.42.2
- Fix possible NPE when updating fog options from Java/Kotlin
- The `emissive` property was not applied properly to `MASKED` materials, and could cause
dark fringes to appear (recompile materials)
- Allow glTF materials with transmission/volume extensions to choose their alpha mode
instead of forcing `MASKED`
- Fix a crash in gltfio when not using ubershaders
- Use flatmat for mat parameter in jsbinding
- Fix TextureFlags for sheenRoughnessMap when textures of sheenRoughnessMap and sheenColorMap is same
- Directional shadows can now be transformed (b/297095805)
## v1.42.1
- Fix potential `EXC_BAD_ACCESS` with Metal backend: b/297059776
- `setFrameCompletedCallback` now takes a `backend::CallbackHandler`.
## v1.42.0
- engine: add preliminary support for instanced stereoscopic rendering [⚠️ **Recompile materials**]

View File

@@ -81,14 +81,14 @@ buildscript {
ext.versions = [
'jdk': 17,
'minSdk': 19,
'targetSdk': 33,
'compileSdk': 33,
'kotlin': '1.9.0',
'kotlin_coroutines': '1.7.2',
'targetSdk': 34,
'compileSdk': 34,
'kotlin': '1.9.21',
'kotlin_coroutines': '1.7.3',
'buildTools': '34.0.0',
'ndk': '25.1.8937393',
'androidx_core': '1.10.1',
'androidx_annotations': '1.6.0'
'ndk': '26.1.10909125',
'androidx_core': '1.12.0',
'androidx_annotations': '1.7.0'
]
ext.deps = [
@@ -104,7 +104,7 @@ buildscript {
]
dependencies {
classpath 'com.android.tools.build:gradle:8.1.0'
classpath 'com.android.tools.build:gradle:8.2.0'
classpath "org.jetbrains.kotlin:kotlin-gradle-plugin:${versions.kotlin}"
}

View File

@@ -6,9 +6,6 @@ option(FILAMENT_ENABLE_MATDBG "Enables Material debugger" OFF)
set(FILAMENT_DIR ${FILAMENT_DIST_DIR})
set(FILAMAT_FLAVOR "filamat")
if(FILAMAT_LITE)
set(FILAMAT_FLAVOR "filamat_lite")
endif()
if (FILAMENT_SUPPORTS_VULKAN)
message("Library filamat ignores Vulkan settings")

View File

@@ -1,29 +1,8 @@
android {
namespace 'com.google.android.filament.filamat'
flavorDimensions "functionality"
productFlavors {
full {
dimension "functionality"
}
lite {
dimension "functionality"
externalNativeBuild {
cmake {
arguments.add("-DFILAMAT_LITE=ON")
}
}
}
}
publishing {
singleVariant("fullRelease") {
withSourcesJar()
withJavadocJar()
}
singleVariant("liteRelease") {
singleVariant("release") {
withSourcesJar()
withJavadocJar()
}
@@ -39,14 +18,9 @@ apply from: rootProject.file('gradle/gradle-mvn-push.gradle')
afterEvaluate { project ->
publishing {
publications {
fullRelease(MavenPublication) {
release(MavenPublication) {
artifactId = POM_ARTIFACT_ID_FULL
from components.fullRelease
}
liteRelease(MavenPublication) {
artifactId = POM_ARTIFACT_ID_LITE
from components.liteRelease
from components.release
}
}
}

View File

@@ -25,15 +25,8 @@
using namespace filament;
using namespace utils;
extern "C" JNIEXPORT jlong JNICALL
Java_com_google_android_filament_Engine_nCreateEngine(JNIEnv*, jclass, jlong backend,
jlong sharedContext) {
return (jlong) Engine::create((Engine::Backend) backend, nullptr, (void*) sharedContext);
}
extern "C" JNIEXPORT void JNICALL
Java_com_google_android_filament_Engine_nDestroyEngine(JNIEnv*, jclass,
jlong nativeEngine) {
Java_com_google_android_filament_Engine_nDestroyEngine(JNIEnv*, jclass, jlong nativeEngine) {
Engine* engine = (Engine*) nativeEngine;
Engine::destroy(&engine);
}
@@ -391,6 +384,13 @@ Java_com_google_android_filament_Engine_nFlushAndWait(JNIEnv*, jclass,
engine->flushAndWait();
}
extern "C" JNIEXPORT void JNICALL
Java_com_google_android_filament_Engine_nFlush(JNIEnv*, jclass,
jlong nativeEngine) {
Engine* engine = (Engine*) nativeEngine;
engine->flush();
}
// Managers...
extern "C" JNIEXPORT jlong JNICALL
@@ -435,6 +435,13 @@ Java_com_google_android_filament_Engine_nIsAutomaticInstancingEnabled(JNIEnv*, j
return (jboolean)engine->isAutomaticInstancingEnabled();
}
extern "C" JNIEXPORT jlong JNICALL
Java_com_google_android_filament_Engine_nGetMaxStereoscopicEyes(JNIEnv*, jclass, jlong nativeEngine) {
Engine* engine = (Engine*) nativeEngine;
return (jlong) engine->getMaxStereoscopicEyes();
}
extern "C" JNIEXPORT jint JNICALL
Java_com_google_android_filament_Engine_nGetSupportedFeatureLevel(JNIEnv *, jclass,
jlong nativeEngine) {
@@ -454,4 +461,60 @@ Java_com_google_android_filament_Engine_nGetActiveFeatureLevel(JNIEnv *, jclass,
jlong nativeEngine) {
Engine* engine = (Engine*) nativeEngine;
return (jint)engine->getActiveFeatureLevel();
}
}
extern "C" JNIEXPORT jlong JNICALL Java_com_google_android_filament_Engine_nCreateBuilder(JNIEnv*,
jclass) {
Engine::Builder* builder = new Engine::Builder{};
return (jlong) builder;
}
extern "C" JNIEXPORT void JNICALL Java_com_google_android_filament_Engine_nDestroyBuilder(JNIEnv*,
jclass, jlong nativeBuilder) {
Engine::Builder* builder = (Engine::Builder*) nativeBuilder;
delete builder;
}
extern "C" JNIEXPORT void JNICALL Java_com_google_android_filament_Engine_nSetBuilderBackend(
JNIEnv*, jclass, jlong nativeBuilder, jlong backend) {
Engine::Builder* builder = (Engine::Builder*) nativeBuilder;
builder->backend((Engine::Backend) backend);
}
extern "C" JNIEXPORT void JNICALL Java_com_google_android_filament_Engine_nSetBuilderConfig(JNIEnv*,
jclass, jlong nativeBuilder, jlong commandBufferSizeMB, jlong perRenderPassArenaSizeMB,
jlong driverHandleArenaSizeMB, jlong minCommandBufferSizeMB, jlong perFrameCommandsSizeMB,
jlong jobSystemThreadCount, jlong stereoscopicEyeCount,
jlong resourceAllocatorCacheSizeMB, jlong resourceAllocatorCacheMaxAge) {
Engine::Builder* builder = (Engine::Builder*) nativeBuilder;
Engine::Config config = {
.commandBufferSizeMB = (uint32_t) commandBufferSizeMB,
.perRenderPassArenaSizeMB = (uint32_t) perRenderPassArenaSizeMB,
.driverHandleArenaSizeMB = (uint32_t) driverHandleArenaSizeMB,
.minCommandBufferSizeMB = (uint32_t) minCommandBufferSizeMB,
.perFrameCommandsSizeMB = (uint32_t) perFrameCommandsSizeMB,
.jobSystemThreadCount = (uint32_t) jobSystemThreadCount,
.stereoscopicEyeCount = (uint8_t) stereoscopicEyeCount,
.resourceAllocatorCacheSizeMB = (uint32_t) resourceAllocatorCacheSizeMB,
.resourceAllocatorCacheMaxAge = (uint8_t) resourceAllocatorCacheMaxAge,
};
builder->config(&config);
}
extern "C" JNIEXPORT void JNICALL Java_com_google_android_filament_Engine_nSetBuilderFeatureLevel(
JNIEnv*, jclass, jlong nativeBuilder, jint ordinal) {
Engine::Builder* builder = (Engine::Builder*) nativeBuilder;
builder->featureLevel((Engine::FeatureLevel)ordinal);
}
extern "C" JNIEXPORT void JNICALL Java_com_google_android_filament_Engine_nSetBuilderSharedContext(
JNIEnv*, jclass, jlong nativeBuilder, jlong sharedContext) {
Engine::Builder* builder = (Engine::Builder*) nativeBuilder;
builder->sharedContext((void*) sharedContext);
}
extern "C" JNIEXPORT jlong JNICALL
Java_com_google_android_filament_Engine_nBuilderBuild(JNIEnv*, jclass, jlong nativeBuilder) {
Engine::Builder* builder = (Engine::Builder*) nativeBuilder;
return (jlong) builder->build();
}

View File

@@ -38,7 +38,7 @@ Java_com_google_android_filament_EntityManager_nCreateArray(JNIEnv* env, jclass,
// (which it is), but still.
em->create((size_t) n, reinterpret_cast<Entity *>(entities));
env->ReleaseIntArrayElements(entities_, entities, 0);
env->ReleaseIntArrayElements(entities_, entities, JNI_ABORT);
}
extern "C" JNIEXPORT jint JNICALL

View File

@@ -79,7 +79,8 @@ Java_com_google_android_filament_LightManager_nBuilderShadowOptions(JNIEnv* env,
jfloat shadowFarHint, jboolean stable, jboolean lispsm,
jfloat polygonOffsetConstant, jfloat polygonOffsetSlope,
jboolean screenSpaceContactShadows, jint stepCount,
jfloat maxShadowDistance, jboolean elvsm, jfloat blurWidth, jfloat shadowBulbRadius) {
jfloat maxShadowDistance, jboolean elvsm, jfloat blurWidth, jfloat shadowBulbRadius,
jfloatArray transform) {
LightManager::Builder *builder = (LightManager::Builder *) nativeBuilder;
LightManager::ShadowOptions shadowOptions {
.mapSize = (uint32_t)mapSize,
@@ -102,12 +103,18 @@ Java_com_google_android_filament_LightManager_nBuilderShadowOptions(JNIEnv* env,
},
.shadowBulbRadius = shadowBulbRadius
};
jfloat *nativeSplits = env->GetFloatArrayElements(splitPositions, NULL);
const jsize splitCount = std::min((jsize) 3, env->GetArrayLength(splitPositions));
for (jsize i = 0; i < splitCount; i++) {
shadowOptions.cascadeSplitPositions[i] = nativeSplits[i];
}
std::copy_n(nativeSplits, splitCount, shadowOptions.cascadeSplitPositions);
env->ReleaseFloatArrayElements(splitPositions, nativeSplits, 0);
jfloat* nativeTransform = env->GetFloatArrayElements(transform, NULL);
std::copy_n(nativeTransform,
std::min(4, env->GetArrayLength(transform)),
shadowOptions.transform.xyzw.v);
env->ReleaseFloatArrayElements(transform, nativeTransform, 0);
builder->shadowOptions(shadowOptions);
}

View File

@@ -19,6 +19,7 @@
#include <filament/Material.h>
#include "common/NioUtils.h"
#include "common/CallbackUtils.h"
using namespace filament;
@@ -105,6 +106,22 @@ Java_com_google_android_filament_Material_nGetRefractionType(JNIEnv*, jclass,
return (jint) material->getRefractionType();
}
extern "C"
JNIEXPORT jint JNICALL
Java_com_google_android_filament_Material_nGetReflectionMode(JNIEnv*, jclass,
jlong nativeMaterial) {
Material* material = (Material*) nativeMaterial;
return (jint) material->getReflectionMode();
}
extern "C"
JNIEXPORT jint JNICALL
Java_com_google_android_filament_Material_nGetFeatureLevel(JNIEnv*, jclass,
jlong nativeMaterial) {
Material* material = (Material*) nativeMaterial;
return (jint) material->getFeatureLevel();
}
extern "C"
JNIEXPORT jint JNICALL
Java_com_google_android_filament_Material_nGetVertexDomain(JNIEnv*, jclass,
@@ -255,3 +272,17 @@ Java_com_google_android_filament_Material_nHasParameter(JNIEnv* env, jclass,
env->ReleaseStringUTFChars(name_, name);
return (jboolean) hasParameter;
}
extern "C"
JNIEXPORT void JNICALL
Java_com_google_android_filament_Material_nCompile(JNIEnv *env, jclass clazz,
jlong nativeMaterial, jint priority, jint variants, jobject handler, jobject runnable) {
Material* material = (Material*) nativeMaterial;
JniCallback* jniCallback = JniCallback::make(env, handler, runnable);
material->compile(
(Material::CompilerPriorityQueue) priority,
(UserVariantFilterBit) variants,
jniCallback->getHandler(), [jniCallback](Material*){
JniCallback::postToJavaAndDestroy(jniCallback);
});
}

View File

@@ -205,7 +205,7 @@ Java_com_google_android_filament_MaterialInstance_nSetIntParameterArray(JNIEnv *
break;
}
env->ReleaseIntArrayElements(v_, v, 0);
env->ReleaseIntArrayElements(v_, v, JNI_ABORT);
env->ReleaseStringUTFChars(name_, name);
}

View File

@@ -71,6 +71,13 @@ Java_com_google_android_filament_Scene_nRemoveEntities(JNIEnv *env, jclass type,
env->ReleaseIntArrayElements(entities, (jint*) nativeEntities, JNI_ABORT);
}
extern "C" JNIEXPORT jint JNICALL
Java_com_google_android_filament_Scene_nGetEntityCount(JNIEnv *env, jclass type,
jlong nativeScene) {
Scene* scene = (Scene*) nativeScene;
return (jint) scene->getEntityCount();
}
extern "C" JNIEXPORT jint JNICALL
Java_com_google_android_filament_Scene_nGetRenderableCount(JNIEnv *env, jclass type,
jlong nativeScene) {
@@ -91,3 +98,22 @@ Java_com_google_android_filament_Scene_nHasEntity(JNIEnv *env, jclass type, jlon
Entity entity = Entity::import(entityId);
return (jboolean) scene->hasEntity(entity);
}
extern "C"
JNIEXPORT jboolean JNICALL
Java_com_google_android_filament_Scene_nGetEntities(JNIEnv *env, jclass ,
jlong nativeScene, jintArray outArray, jint length) {
Scene const* const scene = (Scene*) nativeScene;
if (length < scene->getEntityCount()) {
// should not happen because we already checked on the java side
return JNI_FALSE;
}
jint *out = (jint *) env->GetIntArrayElements(outArray, nullptr);
scene->forEach([out, length, i = 0](Entity entity)mutable {
if (i < length) { // this is just paranoia here
out[i++] = (jint) entity.getId();
}
});
env->ReleaseIntArrayElements(outArray, (jint*) out, 0);
return JNI_TRUE;
}

View File

@@ -47,6 +47,11 @@ Java_com_google_android_filament_ToneMapper_nCreateFilmicToneMapper(JNIEnv*, jcl
return (jlong) new FilmicToneMapper();
}
extern "C" JNIEXPORT jlong JNICALL
Java_com_google_android_filament_ToneMapper_nCreateAgxToneMapper(JNIEnv*, jclass, jint look) {
return (jlong) new AgxToneMapper(AgxToneMapper::AgxLook(look));
}
extern "C" JNIEXPORT jlong JNICALL
Java_com_google_android_filament_ToneMapper_nCreateGenericToneMapper(JNIEnv*, jclass,
jfloat contrast, jfloat midGrayIn, jfloat midGrayOut, jfloat hdrMax) {

View File

@@ -282,7 +282,7 @@ Java_com_google_android_filament_View_nSetSSCTOptions(JNIEnv *, jclass, jlong na
extern "C" JNIEXPORT void JNICALL
Java_com_google_android_filament_View_nSetBloomOptions(JNIEnv*, jclass,
jlong nativeView, jlong nativeTexture,
jfloat dirtStrength, jfloat strength, jint resolution, jfloat anamorphism, jint levels,
jfloat dirtStrength, jfloat strength, jint resolution, jint levels,
jint blendMode, jboolean threshold, jboolean enabled, jfloat highlight,
jboolean lensFlare, jboolean starburst, jfloat chromaticAberration, jint ghostCount,
jfloat ghostSpacing, jfloat ghostThreshold, jfloat haloThickness, jfloat haloRadius,
@@ -294,7 +294,6 @@ Java_com_google_android_filament_View_nSetBloomOptions(JNIEnv*, jclass,
.dirtStrength = dirtStrength,
.strength = strength,
.resolution = (uint32_t)resolution,
.anamorphism = anamorphism,
.levels = (uint8_t)levels,
.blendMode = (View::BloomOptions::BlendMode)blendMode,
.threshold = (bool)threshold,
@@ -481,6 +480,17 @@ Java_com_google_android_filament_View_nIsStencilBufferEnabled(JNIEnv *, jclass,
return view->isStencilBufferEnabled();
}
extern "C"
JNIEXPORT void JNICALL
Java_com_google_android_filament_View_nSetStereoscopicOptions(JNIEnv *, jclass, jlong nativeView,
jboolean enabled) {
View* view = (View*) nativeView;
View::StereoscopicOptions options {
.enabled = (bool) enabled
};
view->setStereoscopicOptions(options);
}
extern "C"
JNIEXPORT void JNICALL
Java_com_google_android_filament_View_nSetGuardBandOptions(JNIEnv *, jclass,

View File

@@ -111,6 +111,8 @@ public class Engine {
private long mNativeObject;
private Config mConfig;
@NonNull private final TransformManager mTransformManager;
@NonNull private final LightManager mLightManager;
@NonNull private final RenderableManager mRenderableManager;
@@ -150,16 +152,232 @@ public class Engine {
FEATURE_LEVEL_0,
/** OpenGL ES 3.0 features (default) */
FEATURE_LEVEL_1,
/** OpenGL ES 3.1 features + 16 textures units + cubemap arrays */
FEATURE_LEVEL_2,
/** OpenGL ES 3.1 features + 31 textures units + cubemap arrays */
FEATURE_LEVEL_2
FEATURE_LEVEL_3,
};
private Engine(long nativeEngine) {
/**
* Constructs <code>Engine</code> objects using a builder pattern.
*/
public static class Builder {
@SuppressWarnings({"FieldCanBeLocal", "UnusedDeclaration"})
private final BuilderFinalizer mFinalizer;
private final long mNativeBuilder;
private Config mConfig;
public Builder() {
mNativeBuilder = nCreateBuilder();
mFinalizer = new BuilderFinalizer(mNativeBuilder);
}
/**
* Sets the {@link Backend} for the Engine.
*
* @param backend Driver backend to use
* @return A reference to this Builder for chaining calls.
*/
public Builder backend(Backend backend) {
nSetBuilderBackend(mNativeBuilder, backend.ordinal());
return this;
}
/**
* Sets a sharedContext for the Engine.
*
* @param sharedContext A platform-dependant OpenGL context used as a shared context
* when creating filament's internal context. On Android this parameter
* <b>must be</b> an instance of {@link android.opengl.EGLContext}.
* @return A reference to this Builder for chaining calls.
*/
public Builder sharedContext(Object sharedContext) {
if (Platform.get().validateSharedContext(sharedContext)) {
nSetBuilderSharedContext(mNativeBuilder,
Platform.get().getSharedContextNativeHandle(sharedContext));
return this;
}
throw new IllegalArgumentException("Invalid shared context " + sharedContext);
}
/**
* Configure the Engine with custom parameters.
*
* @param config A {@link Config} object
* @return A reference to this Builder for chaining calls.
*/
public Builder config(Config config) {
mConfig = config;
nSetBuilderConfig(mNativeBuilder, config.commandBufferSizeMB,
config.perRenderPassArenaSizeMB, config.driverHandleArenaSizeMB,
config.minCommandBufferSizeMB, config.perFrameCommandsSizeMB,
config.jobSystemThreadCount, config.stereoscopicEyeCount,
config.resourceAllocatorCacheSizeMB, config.resourceAllocatorCacheMaxAge);
return this;
}
/**
* Sets the initial featureLevel for the Engine.
*
* @param featureLevel The feature level at which initialize Filament.
* @return A reference to this Builder for chaining calls.
*/
public Builder featureLevel(FeatureLevel featureLevel) {
nSetBuilderFeatureLevel(mNativeBuilder, featureLevel.ordinal());
return this;
}
/**
* Creates an instance of Engine
*
* @return A newly created <code>Engine</code>, or <code>null</code> if the GPU driver couldn't
* be initialized, for instance if it doesn't support the right version of OpenGL or
* OpenGL ES.
*
* @exception IllegalStateException can be thrown if there isn't enough memory to
* allocate the command buffer.
*/
public Engine build() {
long nativeEngine = nBuilderBuild(mNativeBuilder);
if (nativeEngine == 0) throw new IllegalStateException("Couldn't create Engine");
return new Engine(nativeEngine, mConfig);
}
private static class BuilderFinalizer {
private final long mNativeObject;
BuilderFinalizer(long nativeObject) {
mNativeObject = nativeObject;
}
@Override
public void finalize() {
try {
super.finalize();
} catch (Throwable t) { // Ignore
} finally {
nDestroyBuilder(mNativeObject);
}
}
}
}
/**
* Parameters for customizing the initialization of {@link Engine}.
*/
public static class Config {
// #defines in Engine.h
private static final long FILAMENT_PER_RENDER_PASS_ARENA_SIZE_IN_MB = 3;
private static final long FILAMENT_PER_FRAME_COMMANDS_SIZE_IN_MB = 2;
private static final long FILAMENT_MIN_COMMAND_BUFFERS_SIZE_IN_MB = 1;
private static final long FILAMENT_COMMAND_BUFFER_SIZE_IN_MB =
FILAMENT_MIN_COMMAND_BUFFERS_SIZE_IN_MB * 3;
/**
* Size in MiB of the low-level command buffer arena.
*
* Each new command buffer is allocated from here. If this buffer is too small the program
* might terminate or rendering errors might occur.
*
* This is typically set to minCommandBufferSizeMB * 3, so that up to 3 frames can be
* batched-up at once.
*
* This value affects the application's memory usage.
*/
public long commandBufferSizeMB = FILAMENT_COMMAND_BUFFER_SIZE_IN_MB;
/**
* Size in MiB of the per-frame data arena.
*
* This is the main arena used for allocations when preparing a frame.
* e.g.: Froxel data and high-level commands are allocated from this arena.
*
* If this size is too small, the program will abort on debug builds and have undefined
* behavior otherwise.
*
* This value affects the application's memory usage.
*/
public long perRenderPassArenaSizeMB = FILAMENT_PER_RENDER_PASS_ARENA_SIZE_IN_MB;
/**
* Size in MiB of the backend's handle arena.
*
* Backends will fallback to slower heap-based allocations when running out of space and
* log this condition.
*
* If 0, then the default value for the given platform is used
*
* This value affects the application's memory usage.
*/
public long driverHandleArenaSizeMB = 0;
/**
* Minimum size in MiB of a low-level command buffer.
*
* This is how much space is guaranteed to be available for low-level commands when a new
* buffer is allocated. If this is too small, the engine might have to stall to wait for
* more space to become available, this situation is logged.
*
* This value does not affect the application's memory usage.
*/
public long minCommandBufferSizeMB = FILAMENT_MIN_COMMAND_BUFFERS_SIZE_IN_MB;
/**
* Size in MiB of the per-frame high level command buffer.
*
* This buffer is related to the number of draw calls achievable within a frame, if it is
* too small, the program will abort on debug builds and have undefined behavior otherwise.
*
* It is allocated from the 'per-render-pass arena' above. Make sure that at least 1 MiB is
* left in the per-render-pass arena when deciding the size of this buffer.
*
* This value does not affect the application's memory usage.
*/
public long perFrameCommandsSizeMB = FILAMENT_PER_FRAME_COMMANDS_SIZE_IN_MB;
/**
* Number of threads to use in Engine's JobSystem.
*
* Engine uses a utils::JobSystem to carry out paralleization of Engine workloads. This
* value sets the number of threads allocated for JobSystem. Configuring this value can be
* helpful in CPU-constrained environments where too many threads can cause contention of
* CPU and reduce performance.
*
* The default value is 0, which implies that the Engine will use a heuristic to determine
* the number of threads to use.
*/
public long jobSystemThreadCount = 0;
/**
* The number of eyes to render when stereoscopic rendering is enabled. Supported values are
* between 1 and Engine#getMaxStereoscopicEyes() (inclusive).
*
* @see View#setStereoscopicOptions
* @see Engine#getMaxStereoscopicEyes
*/
public long stereoscopicEyeCount = 2;
/*
* Size in MiB of the frame graph texture cache. This should be adjusted based on the
* size of used render targets (typically the screen).
*/
public long resourceAllocatorCacheSizeMB = 64;
/*
* This value determines for how many frames are texture entries kept in the cache.
* The default value of 30 corresponds to about half a second at 60 fps.
*/
public long resourceAllocatorCacheMaxAge = 30;
}
private Engine(long nativeEngine, Config config) {
mNativeObject = nativeEngine;
mTransformManager = new TransformManager(nGetTransformManager(nativeEngine));
mLightManager = new LightManager(nGetLightManager(nativeEngine));
mRenderableManager = new RenderableManager(nGetRenderableManager(nativeEngine));
mEntityManager = new EntityManager(nGetEntityManager(nativeEngine));
mConfig = config;
}
/**
@@ -177,9 +395,7 @@ public class Engine {
*/
@NonNull
public static Engine create() {
long nativeEngine = nCreateEngine(0, 0);
if (nativeEngine == 0) throw new IllegalStateException("Couldn't create Engine");
return new Engine(nativeEngine);
return new Builder().build();
}
/**
@@ -199,9 +415,9 @@ public class Engine {
*/
@NonNull
public static Engine create(@NonNull Backend backend) {
long nativeEngine = nCreateEngine(backend.ordinal(), 0);
if (nativeEngine == 0) throw new IllegalStateException("Couldn't create Engine");
return new Engine(nativeEngine);
return new Builder()
.backend(backend)
.build();
}
/**
@@ -223,13 +439,9 @@ public class Engine {
*/
@NonNull
public static Engine create(@NonNull Object sharedContext) {
if (Platform.get().validateSharedContext(sharedContext)) {
long nativeEngine = nCreateEngine(0,
Platform.get().getSharedContextNativeHandle(sharedContext));
if (nativeEngine == 0) throw new IllegalStateException("Couldn't create Engine");
return new Engine(nativeEngine);
}
throw new IllegalArgumentException("Invalid shared context " + sharedContext);
return new Builder()
.sharedContext(sharedContext)
.build();
}
/**
@@ -296,17 +508,23 @@ public class Engine {
}
/**
* Activate all features of a given feature level. By default FeatureLevel::FEATURE_LEVEL_1 is
* active. The selected feature level must not be higher than the value returned by
* getActiveFeatureLevel() and it's not possible lower the active feature level.
* Activate all features of a given feature level. If an explicit feature level is not specified
* at Engine initialization time via {@link Builder#featureLevel}, the default feature level is
* {@link FeatureLevel#FEATURE_LEVEL_0} on devices not compatible with GLES 3.0; otherwise, the
* default is {@link FeatureLevel::FEATURE_LEVEL_1}. The selected feature level must not be
* higher than the value returned by {@link #getActiveFeatureLevel} and it's not possible lower
* the active feature level. Additionally, it is not possible to modify the feature level at all
* if the Engine was initialized at {@link FeatureLevel#FEATURE_LEVEL_0}.
*
* @param featureLevel the feature level to activate. If featureLevel is lower than
* getActiveFeatureLevel(), the current (higher) feature level is kept.
* If featureLevel is higher than getSupportedFeatureLevel(), an exception
* is thrown, or the program is terminated if exceptions are disabled.
* @param featureLevel the feature level to activate. If featureLevel is lower than {@link
* #getActiveFeatureLevel}, the current (higher) feature level is kept. If
* featureLevel is higher than {@link #getSupportedFeatureLevel}, or if the
* engine was initialized at feature level 0, an exception is thrown, or the
* program is terminated if exceptions are disabled.
*
* @return the active feature level.
*
* @see Builder#featureLevel
* @see #getSupportedFeatureLevel
* @see #getActiveFeatureLevel
*/
@@ -352,6 +570,37 @@ public class Engine {
return nIsAutomaticInstancingEnabled(getNativeObject());
}
/**
* Retrieves the configuration settings of this {@link Engine}.
*
* This method returns the configuration object that was supplied to the Engine's {@link
* Builder#config} method during the creation of this Engine. If the {@link Builder::config}
* method was not explicitly called (or called with null), this method returns the default
* configuration settings.
*
* @return a {@link Config} object with this Engine's configuration
* @see Builder#config
*/
@NonNull
public Config getConfig() {
if (mConfig == null) {
mConfig = new Config();
}
return mConfig;
}
/**
* Returns the maximum number of stereoscopic eyes supported by Filament. The actual number of
* eyes rendered is set at Engine creation time with the {@link
* Engine#Config#stereoscopicEyeCount} setting.
*
* @return the max number of stereoscopic eyes supported
* @see Engine#Config#stereoscopicEyeCount
*/
public long getMaxStereoscopicEyes() {
return nGetMaxStereoscopicEyes(getNativeObject());
}
// SwapChain
@@ -888,6 +1137,18 @@ public class Engine {
nFlushAndWait(getNativeObject());
}
/**
* Kicks the hardware thread (e.g. the OpenGL, Vulkan or Metal thread) but does not wait
* for commands to be either executed or the hardware finished.
*
* <p>This is typically used after creating a lot of objects to start draining the command
* queue which has a limited size.</p>
*/
public void flush() {
nFlush(getNativeObject());
}
@UsedByReflection("TextureHelper.java")
public long getNativeObject() {
if (mNativeObject == 0) {
@@ -914,7 +1175,6 @@ public class Engine {
}
}
private static native long nCreateEngine(long backend, long sharedContext);
private static native void nDestroyEngine(long nativeEngine);
private static native long nGetBackend(long nativeEngine);
private static native long nCreateSwapChain(long nativeEngine, Object nativeWindow, long flags);
@@ -961,6 +1221,7 @@ public class Engine {
private static native boolean nIsValidSwapChain(long nativeEngine, long nativeSwapChain);
private static native void nDestroyEntity(long nativeEngine, int entity);
private static native void nFlushAndWait(long nativeEngine);
private static native void nFlush(long nativeEngine);
private static native long nGetTransformManager(long nativeEngine);
private static native long nGetLightManager(long nativeEngine);
private static native long nGetRenderableManager(long nativeEngine);
@@ -968,7 +1229,20 @@ public class Engine {
private static native long nGetEntityManager(long nativeEngine);
private static native void nSetAutomaticInstancingEnabled(long nativeEngine, boolean enable);
private static native boolean nIsAutomaticInstancingEnabled(long nativeEngine);
private static native long nGetMaxStereoscopicEyes(long nativeEngine);
private static native int nGetSupportedFeatureLevel(long nativeEngine);
private static native int nSetActiveFeatureLevel(long nativeEngine, int ordinal);
private static native int nGetActiveFeatureLevel(long nativeEngine);
private static native long nCreateBuilder();
private static native void nDestroyBuilder(long nativeBuilder);
private static native void nSetBuilderBackend(long nativeBuilder, long backend);
private static native void nSetBuilderConfig(long nativeBuilder, long commandBufferSizeMB,
long perRenderPassArenaSizeMB, long driverHandleArenaSizeMB,
long minCommandBufferSizeMB, long perFrameCommandsSizeMB, long jobSystemThreadCount,
long stereoscopicEyeCount,
long resourceAllocatorCacheSizeMB, long resourceAllocatorCacheMaxAge);
private static native void nSetBuilderFeatureLevel(long nativeBuilder, int ordinal);
private static native void nSetBuilderSharedContext(long nativeBuilder, long sharedContext);
private static native long nBuilderBuild(long nativeBuilder);
}

View File

@@ -258,6 +258,7 @@ public class LightManager {
* shadows that are too far and wouldn't contribute to the scene much, improving
* performance and quality. This value is always positive.
* Use 0.0f to use the camera far distance.
* This only affect directional lights.
*/
public float shadowFar = 0.0f;
@@ -368,6 +369,17 @@ public class LightManager {
* enabled. (2cm by default).
*/
public float shadowBulbRadius = 0.02f;
/**
* Transforms the shadow direction. Must be a unit quaternion.
* The default is identity.
* Ignored if the light type isn't directional. For artistic use. Use with caution.
* The quaternion is stored as the imaginary part in the first 3 elements and the real
* part in the last element of the transform array.
*/
@NonNull
@Size(min = 4, max = 4)
public float[] transform = { 0.0f, 0.0f, 0.0f, 1.0f };
}
public static class ShadowCascades {
@@ -506,7 +518,7 @@ public class LightManager {
options.polygonOffsetConstant, options.polygonOffsetSlope,
options.screenSpaceContactShadows,
options.stepCount, options.maxShadowDistance,
options.elvsm, options.blurWidth, options.shadowBulbRadius);
options.elvsm, options.blurWidth, options.shadowBulbRadius, options.transform);
return this;
}
@@ -1169,7 +1181,7 @@ public class LightManager {
boolean stable, boolean lispsm,
float polygonOffsetConstant, float polygonOffsetSlope,
boolean screenSpaceContactShadows, int stepCount, float maxShadowDistance,
boolean elvsm, float blurWidth, float shadowBulbRadius);
boolean elvsm, float blurWidth, float shadowBulbRadius, float[] transform);
private static native void nBuilderCastLight(long nativeBuilder, boolean enabled);
private static native void nBuilderPosition(long nativeBuilder, float x, float y, float z);
private static native void nBuilderDirection(long nativeBuilder, float x, float y, float z);

View File

@@ -18,9 +18,11 @@ package com.google.android.filament;
import androidx.annotation.IntRange;
import androidx.annotation.NonNull;
import androidx.annotation.Nullable;
import androidx.annotation.Size;
import com.google.android.filament.proguard.UsedByNative;
import com.google.android.filament.Engine.FeatureLevel;
import java.nio.Buffer;
import java.util.ArrayList;
@@ -46,6 +48,8 @@ public class Material {
static final BlendingMode[] sBlendingModeValues = BlendingMode.values();
static final RefractionMode[] sRefractionModeValues = RefractionMode.values();
static final RefractionType[] sRefractionTypeValues = RefractionType.values();
static final ReflectionMode[] sReflectionModeValues = ReflectionMode.values();
static final FeatureLevel[] sFeatureLevelValues = FeatureLevel.values();
static final VertexDomain[] sVertexDomainValues = VertexDomain.values();
static final CullingMode[] sCullingModeValues = CullingMode.values();
static final VertexBuffer.VertexAttribute[] sVertexAttributeValues =
@@ -181,6 +185,18 @@ public class Material {
THIN
}
/**
* Supported reflection modes
*
* @see
* <a href="https://google.github.io/filament/Materials.html#materialdefinitions/materialblock/lighting:reflections">
* Lighting: reflections</a>
*/
public enum ReflectionMode {
DEFAULT,
SCREEN_SPACE
}
/**
* Supported types of vertex domains
*
@@ -223,6 +239,31 @@ public class Material {
FRONT_AND_BACK
}
public enum CompilerPriorityQueue {
HIGH,
LOW
}
public static class UserVariantFilterBit {
/** Directional lighting */
public static int DIRECTIONAL_LIGHTING = 0x01;
/** Dynamic lighting */
public static int DYNAMIC_LIGHTING = 0x02;
/** Shadow receiver */
public static int SHADOW_RECEIVER = 0x04;
/** Skinning */
public static int SKINNING = 0x08;
/** Fog */
public static int FOG = 0x10;
/** Variance shadow maps */
public static int VSM = 0x20;
/** Screen-space reflections */
public static int SSR = 0x40;
/** Instanced stereo rendering */
public static int STE = 0x80;
public static int ALL = 0xFF;
}
@UsedByNative("Material.cpp")
public static class Parameter {
private static final Type[] sTypeValues = Type.values();
@@ -337,6 +378,55 @@ public class Material {
}
}
/**
* Asynchronously ensures that a subset of this Material's variants are compiled. After issuing
* several compile() calls in a row, it is recommended to call {@link Engine#flush}
* such that the backend can start the compilation work as soon as possible.
* The provided callback is guaranteed to be called on the main thread after all specified
* variants of the material are compiled. This can take hundreds of milliseconds.
*<p>
* If all the material's variants are already compiled, the callback will be scheduled as
* soon as possible, but this might take a few dozen millisecond, corresponding to how
* many previous frames are enqueued in the backend. This also varies by backend. Therefore,
* it is recommended to only call this method once per material shortly after creation.
*</p>
*<p>
* If the same variant is scheduled for compilation multiple times, the first scheduling
* takes precedence; later scheduling are ignored.
*</p>
*<p>
* caveat: A consequence is that if a variant is scheduled on the low priority queue and later
* scheduled again on the high priority queue, the later scheduling is ignored.
* Therefore, the second callback could be called before the variant is compiled.
* However, the first callback, if specified, will trigger as expected.
*</p>
*<p>
* The callback is guaranteed to be called. If the engine is destroyed while some material
* variants are still compiling or in the queue, these will be discarded and the corresponding
* callback will be called. In that case however the Material pointer passed to the callback
* is guaranteed to be invalid (either because it's been destroyed by the user already, or,
* because it's been cleaned-up by the Engine).
*</p>
*<p>
* {@link UserVariantFilterBit#ALL} should be used with caution. Only variants that an application
* needs should be included in the variants argument. For example, the STE variant is only used
* for stereoscopic rendering. If an application is not planning to render in stereo, this bit
* should be turned off to avoid unnecessary material compilations.
*</p>
* @param priority Which priority queue to use, LOW or HIGH.
* @param variants Variants to include to the compile command.
* @param handler An {@link java.util.concurrent.Executor Executor}. On Android this can also be a {@link android.os.Handler Handler}.
* @param callback callback called on the main thread when the compilation is done on
* by backend.
*/
public void compile(@NonNull CompilerPriorityQueue priority,
int variants,
@Nullable Object handler,
@Nullable Runnable callback) {
nCompile(getNativeObject(), priority.ordinal(), variants, handler, callback);
}
/**
* Creates a new instance of this material. Material instances should be freed using
* {@link Engine#destroyMaterialInstance(MaterialInstance)}.
@@ -437,6 +527,28 @@ public class Material {
return EnumCache.sRefractionTypeValues[nGetRefractionType(getNativeObject())];
}
/**
* Returns the reflection mode of this material.
*
* @see
* <a href="https://google.github.io/filament/Materials.html#materialdefinitions/materialblock/lighting:reflections">
* Lighting: reflections</a>
*/
public ReflectionMode getReflectionMode() {
return EnumCache.sReflectionModeValues[nGetReflectionMode(getNativeObject())];
}
/**
* Returns the minimum required feature level for this material.
*
* @see
* <a href="https://google.github.io/filament/Materials.html#materialdefinitions/materialblock/general:featurelevel">
* General: featureLevel</a>
*/
public FeatureLevel getFeatureLevel() {
return EnumCache.sFeatureLevelValues[nGetFeatureLevel(getNativeObject())];
}
/**
* Returns the vertex domain of this material.
*
@@ -916,6 +1028,7 @@ public class Material {
private static native long nCreateInstanceWithName(long nativeMaterial, @NonNull String name);
private static native long nGetDefaultInstance(long nativeMaterial);
private static native void nCompile(long nativeMaterial, int priority, int variants, Object handler, Runnable runnable);
private static native String nGetName(long nativeMaterial);
private static native int nGetShading(long nativeMaterial);
private static native int nGetInterpolation(long nativeMaterial);
@@ -932,6 +1045,8 @@ public class Material {
private static native float nGetSpecularAntiAliasingThreshold(long nativeMaterial);
private static native int nGetRefractionMode(long nativeMaterial);
private static native int nGetRefractionType(long nativeMaterial);
private static native int nGetReflectionMode(long nativeMaterial);
private static native int nGetFeatureLevel(long nativeMaterial);
private static native int nGetParameterCount(long nativeMaterial);

View File

@@ -16,6 +16,7 @@
package com.google.android.filament;
import androidx.annotation.NonNull;
import androidx.annotation.Nullable;
/**
@@ -146,18 +147,29 @@ public class Scene {
}
/**
* Returns the number of {@link RenderableManager} components in the <code>Scene</code>.
* Returns the total number of Entities in the <code>Scene</code>, whether alive or not.
*
* @return number of {@link RenderableManager} components in the <code>Scene</code>..
* @return the total number of Entities in the <code>Scene</code>.
*/
public int getEntityCount() {
return nGetEntityCount(getNativeObject());
}
/**
* Returns the number of active (alive) {@link RenderableManager} components in the
* <code>Scene</code>.
*
* @return number of {@link RenderableManager} components in the <code>Scene</code>.
*/
public int getRenderableCount() {
return nGetRenderableCount(getNativeObject());
}
/**
* Returns the number of {@link LightManager} components in the <code>Scene</code>.
* Returns the number of active (alive) {@link LightManager} components in the
* <code>Scene</code>.
*
* @return number of {@link LightManager} components in the <code>Scene</code>..
* @return number of {@link LightManager} components in the <code>Scene</code>.
*/
public int getLightCount() {
return nGetLightCount(getNativeObject());
@@ -179,6 +191,52 @@ public class Scene {
return mNativeObject;
}
/**
* Returns the list of all entities in the Scene. If outArray is provided and large enough,
* it is used to store the list and returned, otherwise a new array is allocated and returned.
* @param outArray an array to store the list of entities in the scene.
* @return outArray if it was used or a newly allocated array.
* @see #getEntityCount
*/
public int[] getEntities(@Nullable int[] outArray) {
int c = getEntityCount();
if (outArray == null || outArray.length < c) {
outArray = new int[c];
}
boolean success = nGetEntities(getNativeObject(), outArray, outArray.length);
if (!success) {
throw new IllegalStateException("Error retriving Scene's entities");
}
return outArray;
}
/**
* Returns the list of all entities in the Scene in a newly allocated array.
* @return an array containing the list of all entities in the scene.
* @see #getEntityCount
*/
public int[] getEntities() {
return getEntities(null);
}
public interface EntityProcessor {
void process(@Entity int entity);
}
/**
* Invokes user functor on each entity in the scene.
*
* It is not allowed to add or remove an entity from the scene within the functor.
*
* @param entityProcessor User provided functor called for each entity in the scene
*/
public void forEach(@NonNull EntityProcessor entityProcessor) {
int[] entities = getEntities(null);
for (int entity : entities) {
entityProcessor.process(entity);
}
}
void clearNativeObject() {
mNativeObject = 0;
}
@@ -189,7 +247,9 @@ public class Scene {
private static native void nAddEntities(long nativeScene, int[] entities);
private static native void nRemove(long nativeScene, int entity);
private static native void nRemoveEntities(long nativeScene, int[] entities);
private static native int nGetEntityCount(long nativeScene);
private static native int nGetRenderableCount(long nativeScene);
private static native int nGetLightCount(long nativeScene);
private static native boolean nHasEntity(long nativeScene, int entity);
private static native boolean nGetEntities(long nativeScene, int[] outArray, int length);
}

View File

@@ -103,6 +103,31 @@ public class SwapChain {
*/
public static final long CONFIG_SRGB_COLORSPACE = 0x10;
/**
* Indicates that this SwapChain should allocate a stencil buffer in addition to a depth buffer.
*
* This flag is necessary when using View::setStencilBufferEnabled and rendering directly into
* the SwapChain (when post-processing is disabled).
*
* The specific format of the stencil buffer depends on platform support. The following pixel
* formats are tried, in order of preference:
*
* Depth only (without CONFIG_HAS_STENCIL_BUFFER):
* - DEPTH32F
* - DEPTH24
*
* Depth + stencil (with CONFIG_HAS_STENCIL_BUFFER):
* - DEPTH32F_STENCIL8
* - DEPTH24F_STENCIL8
*
* Note that enabling the stencil buffer may hinder depth precision and should only be used if
* necessary.
*
* @see View#setStencilBufferEnabled
* @see View#setPostProcessingEnabled
*/
public static final long CONFIG_HAS_STENCIL_BUFFER = 0x20;
SwapChain(long nativeSwapChain, Object surface) {
mNativeObject = nativeSwapChain;
mSurface = surface;

View File

@@ -849,6 +849,10 @@ public class Texture {
public static final int SAMPLEABLE = 0x10;
/** Texture can be used as a subpass input */
public static final int SUBPASS_INPUT = 0x20;
/** Texture can be used the source of a blit() */
public static final int BLIT_SRC = 0x40;
/** Texture can be used the destination of a blit() */
public static final int BLIT_DST = 0x80;
/** by default textures are <code>UPLOADABLE</code> and <code>SAMPLEABLE</code>*/
public static final int DEFAULT = UPLOADABLE | SAMPLEABLE;
}

View File

@@ -100,6 +100,45 @@ public class ToneMapper {
}
}
/**
* AgX tone mapping operator.
*/
public static class Agx extends ToneMapper {
public enum AgxLook {
/**
* Base contrast with no look applied
*/
NONE,
/**
* A punchy and more chroma laden look for sRGB displays
*/
PUNCHY,
/**
* A golden tinted, slightly washed look for BT.1886 displays
*/
GOLDEN
}
/**
* Builds a new AgX tone mapper with no look applied.
*/
public Agx() {
this(AgxLook.NONE);
}
/**
* Builds a new AgX tone mapper.
*
* @param look: an optional creative adjustment to contrast and saturation
*/
public Agx(AgxLook look) {
super(nCreateAgxToneMapper(look.ordinal()));
}
}
/**
* Generic tone mapping operator that gives control over the tone mapping
* curve. This operator can be used to control the aesthetics of the final
@@ -194,6 +233,7 @@ public class ToneMapper {
private static native long nCreateACESToneMapper();
private static native long nCreateACESLegacyToneMapper();
private static native long nCreateFilmicToneMapper();
private static native long nCreateAgxToneMapper(int look);
private static native long nCreateGenericToneMapper(
float contrast, float midGrayIn, float midGrayOut, float hdrMax);

View File

@@ -75,6 +75,7 @@ public class View {
private AmbientOcclusionOptions mAmbientOcclusionOptions;
private BloomOptions mBloomOptions;
private FogOptions mFogOptions;
private StereoscopicOptions mStereoscopicOptions;
private RenderTarget mRenderTarget;
private BlendMode mBlendMode;
private DepthOfFieldOptions mDepthOfFieldOptions;
@@ -904,7 +905,7 @@ public class View {
mBloomOptions = options;
nSetBloomOptions(getNativeObject(), options.dirt != null ? options.dirt.getNativeObject() : 0,
options.dirtStrength, options.strength, options.resolution,
options.anamorphism, options.levels, options.blendMode.ordinal(),
options.levels, options.blendMode.ordinal(),
options.threshold, options.enabled, options.highlight,
options.lensFlare, options.starburst, options.chromaticAberration,
options.ghostCount, options.ghostSpacing, options.ghostThreshold,
@@ -1032,7 +1033,8 @@ public class View {
* </p>
*
* <p>
* Post-processing must be enabled in order to use the stencil buffer.
* If post-processing is disabled, then the SwapChain must have the CONFIG_HAS_STENCIL_BUFFER
* flag set in order to use the stencil buffer.
* </p>
*
* <p>
@@ -1054,6 +1056,51 @@ public class View {
return nIsStencilBufferEnabled(getNativeObject());
}
/**
* Sets the stereoscopic rendering options for this view.
*
* <p>
* Currently, only one type of stereoscopic rendering is supported: side-by-side.
* Side-by-side stereo rendering splits the viewport into two halves: a left and right half.
* Eye 0 will render to the left half, while Eye 1 will render into the right half.
* </p>
*
* <p>
* Currently, the following features are not supported with stereoscopic rendering:
* - post-processing
* - shadowing
* - punctual lights
* </p>
*
* <p>
* Stereo rendering depends on device and platform support. To check if stereo rendering is
* supported, use {@link Engine#isStereoSupported()}. If stereo rendering is not supported, then
* the stereoscopic options have no effect.
* </p>
*
* @param options The stereoscopic options to use on this view
* @see #getStereoscopicOptions
*/
public void setStereoscopicOptions(@NonNull StereoscopicOptions options) {
mStereoscopicOptions = options;
nSetStereoscopicOptions(getNativeObject(), options.enabled);
}
/**
* Gets the stereoscopic options.
*
* @return options Stereoscopic options currently set.
* @see #setStereoscopicOptions
*/
@NonNull
public StereoscopicOptions getStereoscopicOptions() {
if (mStereoscopicOptions == null) {
mStereoscopicOptions = new StereoscopicOptions();
}
return mStereoscopicOptions;
}
/**
* A class containing the result of a picking query
*/
@@ -1216,9 +1263,10 @@ public class View {
private static native int nGetAmbientOcclusion(long nativeView);
private static native void nSetAmbientOcclusionOptions(long nativeView, float radius, float bias, float power, float resolution, float intensity, float bilateralThreshold, int quality, int lowPassFilter, int upsampling, boolean enabled, boolean bentNormals, float minHorizonAngleRad);
private static native void nSetSSCTOptions(long nativeView, float ssctLightConeRad, float ssctStartTraceDistance, float ssctContactDistanceMax, float ssctIntensity, float v, float v1, float v2, float ssctDepthBias, float ssctDepthSlopeBias, int ssctSampleCount, int ssctRayCount, boolean ssctEnabled);
private static native void nSetBloomOptions(long nativeView, long dirtNativeObject, float dirtStrength, float strength, int resolution, float anamorphism, int levels, int blendMode, boolean threshold, boolean enabled, float highlight,
private static native void nSetBloomOptions(long nativeView, long dirtNativeObject, float dirtStrength, float strength, int resolution, int levels, int blendMode, boolean threshold, boolean enabled, float highlight,
boolean lensFlare, boolean starburst, float chromaticAberration, int ghostCount, float ghostSpacing, float ghostThreshold, float haloThickness, float haloRadius, float haloThreshold);
private static native void nSetFogOptions(long nativeView, float distance, float maximumOpacity, float height, float heightFalloff, float cutOffDistance, float v, float v1, float v2, float density, float inScatteringStart, float inScatteringSize, boolean fogColorFromIbl, long skyColorNativeObject, boolean enabled);
private static native void nSetStereoscopicOptions(long nativeView, boolean enabled);
private static native void nSetBlendMode(long nativeView, int blendMode);
private static native void nSetDepthOfFieldOptions(long nativeView, float cocScale, float maxApertureDiameter, boolean enabled, int filter,
boolean nativeResolution, int foregroundRingCount, int backgroundRingCount, int fastGatherRingCount, int maxForegroundCOC, int maxBackgroundCOC);
@@ -1353,8 +1401,6 @@ public class View {
* blendMode: Whether the bloom effect is purely additive (false) or mixed with the original
* image (true).
*
* anamorphism: Bloom's aspect ratio (x/y), for artistic purposes.
*
* threshold: When enabled, a threshold at 1.0 is applied on the source image, this is
* useful for artistic reasons and is usually needed when a dirt texture is used.
*
@@ -1393,10 +1439,6 @@ public class View {
* resolution of vertical axis (2^levels to 2048)
*/
public int resolution = 384;
/**
* bloom x/y aspect-ratio (1/32 to 32)
*/
public float anamorphism = 1.0f;
/**
* number of blur levels (1 to 11)
*/
@@ -1418,6 +1460,17 @@ public class View {
* limit highlights to this value before bloom [10, +inf]
*/
public float highlight = 1000.0f;
/**
* Bloom quality level.
* LOW (default): use a more optimized down-sampling filter, however there can be artifacts
* with dynamic resolution, this can be alleviated by using the homogenous mode.
* MEDIUM: Good balance between quality and performance.
* HIGH: In this mode the bloom resolution is automatically increased to avoid artifacts.
* This mode can be significantly slower on mobile, especially at high resolution.
* This mode greatly improves the anamorphic bloom.
*/
@NonNull
public QualityLevel quality = QualityLevel.LOW;
/**
* enable screen-space lens flare
*/
@@ -1584,6 +1637,10 @@ public class View {
* circle of confusion scale factor (amount of blur)
*/
public float cocScale = 1.0f;
/**
* width/height aspect ratio of the circle of confusion (simulate anamorphic lenses)
*/
public float cocAspectRatio = 1.0f;
/**
* maximum aperture diameter in meters (zero to disable rotation)
*/
@@ -1799,7 +1856,7 @@ public class View {
}
/**
* Options for Temporal Multi-Sample Anti-aliasing (MSAA)
* Options for Multi-Sample Anti-aliasing (MSAA)
* @see setMultiSampleAntiAliasingOptions()
*/
public static class MultiSampleAntiAliasingOptions {
@@ -1823,21 +1880,111 @@ public class View {
/**
* Options for Temporal Anti-aliasing (TAA)
* Most TAA parameters are extremely costly to change, as they will trigger the TAA post-process
* shaders to be recompiled. These options should be changed or set during initialization.
* `filterWidth`, `feedback` and `jitterPattern`, however, can be changed at any time.
*
* `feedback` of 0.1 effectively accumulates a maximum of 19 samples in steady state.
* see "A Survey of Temporal Antialiasing Techniques" by Lei Yang and all for more information.
*
* @see setTemporalAntiAliasingOptions()
*/
public static class TemporalAntiAliasingOptions {
public enum BoxType {
/**
* use an AABB neighborhood
*/
AABB,
/**
* use the variance of the neighborhood (not recommended)
*/
VARIANCE,
/**
* use both AABB and variance
*/
AABB_VARIANCE,
}
public enum BoxClipping {
/**
* Accurate box clipping
*/
ACCURATE,
/**
* clamping
*/
CLAMP,
/**
* no rejections (use for debugging)
*/
NONE,
}
public enum JitterPattern {
RGSS_X4,
UNIFORM_HELIX_X4,
HALTON_23_X8,
HALTON_23_X16,
HALTON_23_X32,
}
/**
* reconstruction filter width typically between 0 (sharper, aliased) and 1 (smoother)
* reconstruction filter width typically between 0.2 (sharper, aliased) and 1.5 (smoother)
*/
public float filterWidth = 1.0f;
/**
* history feedback, between 0 (maximum temporal AA) and 1 (no temporal AA).
*/
public float feedback = 0.04f;
public float feedback = 0.12f;
/**
* texturing lod bias (typically -1 or -2)
*/
public float lodBias = -1.0f;
/**
* post-TAA sharpen, especially useful when upscaling is true.
*/
public float sharpness = 0.0f;
/**
* enables or disables temporal anti-aliasing
*/
public boolean enabled = false;
/**
* 4x TAA upscaling. Disables Dynamic Resolution. [BETA]
*/
public boolean upscaling = false;
/**
* whether to filter the history buffer
*/
public boolean filterHistory = true;
/**
* whether to apply the reconstruction filter to the input
*/
public boolean filterInput = true;
/**
* whether to use the YcoCg color-space for history rejection
*/
public boolean useYCoCg = false;
/**
* type of color gamut box
*/
@NonNull
public TemporalAntiAliasingOptions.BoxType boxType = TemporalAntiAliasingOptions.BoxType.AABB;
/**
* clipping algorithm
*/
@NonNull
public TemporalAntiAliasingOptions.BoxClipping boxClipping = TemporalAntiAliasingOptions.BoxClipping.ACCURATE;
@NonNull
public TemporalAntiAliasingOptions.JitterPattern jitterPattern = TemporalAntiAliasingOptions.JitterPattern.HALTON_23_X16;
public float varianceGamma = 1.0f;
/**
* adjust the feedback dynamically to reduce flickering
*/
public boolean preventFlickering = false;
/**
* whether to apply history reprojection (debug option)
*/
public boolean historyReprojection = true;
}
/**
@@ -1924,6 +2071,7 @@ public class View {
* PCF with soft shadows and contact hardening
*/
PCSS,
PCFd,
}
/**

View File

@@ -182,28 +182,85 @@ public class UiHelper {
void detach();
}
private static class SurfaceViewHandler implements RenderSurface {
private final SurfaceView mSurfaceView;
private class SurfaceViewHandler implements RenderSurface, SurfaceHolder.Callback {
@NonNull private final SurfaceView mSurfaceView;
SurfaceViewHandler(SurfaceView surface) {
mSurfaceView = surface;
SurfaceViewHandler(@NonNull SurfaceView surfaceView) {
mSurfaceView = surfaceView;
@NonNull SurfaceHolder holder = surfaceView.getHolder();
holder.addCallback(this);
if (mDesiredWidth > 0 && mDesiredHeight > 0) {
holder.setFixedSize(mDesiredWidth, mDesiredHeight);
}
// in case the SurfaceView's surface already existed
final Surface surface = holder.getSurface();
if (surface != null && surface.isValid()) {
surfaceCreated(holder);
// there is no way to retrieve the actual PixelFormat, since it is not used
// in the callback, we can use whatever we want.
surfaceChanged(holder, PixelFormat.RGBA_8888,
holder.getSurfaceFrame().width(), holder.getSurfaceFrame().height());
}
}
@Override
public void resize(int width, int height) {
mSurfaceView.getHolder().setFixedSize(width, height);
@NonNull SurfaceHolder holder = mSurfaceView.getHolder();
holder.setFixedSize(width, height);
}
@Override
public void detach() {
@NonNull SurfaceHolder holder = mSurfaceView.getHolder();
holder.removeCallback(this);
}
@Override
public void surfaceCreated(@NonNull SurfaceHolder holder) {
if (LOGGING) Log.d(LOG_TAG, "surfaceCreated()");
createSwapChain(holder.getSurface());
}
@Override
public void surfaceChanged(
@NonNull SurfaceHolder holder, int format, int width, int height) {
// Note: this is always called at least once after surfaceCreated()
if (LOGGING) Log.d(LOG_TAG, "surfaceChanged(" + width + ", " + height + ")");
if (mRenderCallback != null) {
mRenderCallback.onResized(width, height);
}
}
@Override
public void surfaceDestroyed(@NonNull SurfaceHolder holder) {
if (LOGGING) Log.d(LOG_TAG, "surfaceDestroyed()");
destroySwapChain();
}
}
private static class SurfaceHolderHandler implements RenderSurface {
private class SurfaceHolderHandler implements RenderSurface, SurfaceHolder.Callback {
private final SurfaceHolder mSurfaceHolder;
SurfaceHolderHandler(SurfaceHolder surface) {
mSurfaceHolder = surface;
SurfaceHolderHandler(@NonNull SurfaceHolder holder) {
mSurfaceHolder = holder;
holder.addCallback(this);
if (mDesiredWidth > 0 && mDesiredHeight > 0) {
holder.setFixedSize(mDesiredWidth, mDesiredHeight);
}
// in case the SurfaceHolder's surface already existed
final Surface surface = holder.getSurface();
if (surface != null && surface.isValid()) {
surfaceCreated(holder);
// there is no way to retrieve the actual PixelFormat, since it is not used
// in the callback, we can use whatever we want.
surfaceChanged(holder, PixelFormat.RGBA_8888,
holder.getSurfaceFrame().width(), holder.getSurfaceFrame().height());
}
}
@Override
@@ -213,30 +270,127 @@ public class UiHelper {
@Override
public void detach() {
mSurfaceHolder.removeCallback(this);
}
@Override
public void surfaceCreated(@NonNull SurfaceHolder holder) {
if (LOGGING) Log.d(LOG_TAG, "surfaceCreated()");
createSwapChain(holder.getSurface());
}
@Override
public void surfaceChanged(@NonNull SurfaceHolder holder, int format, int width, int height) {
// Note: this is always called at least once after surfaceCreated()
if (LOGGING) Log.d(LOG_TAG, "surfaceChanged(" + width + ", " + height + ")");
if (mRenderCallback != null) {
mRenderCallback.onResized(width, height);
}
}
@Override
public void surfaceDestroyed(@NonNull SurfaceHolder surfaceHolder) {
if (LOGGING) Log.d(LOG_TAG, "surfaceDestroyed()");
destroySwapChain();
}
}
private class TextureViewHandler implements RenderSurface {
private class TextureViewHandler implements RenderSurface, TextureView.SurfaceTextureListener {
private final TextureView mTextureView;
private Surface mSurface;
TextureViewHandler(TextureView surface) { mTextureView = surface; }
TextureViewHandler(@NonNull TextureView view) {
mTextureView = view;
mTextureView.setSurfaceTextureListener(this);
// in case the View's SurfaceTexture already existed
if (view.isAvailable()) {
SurfaceTexture surfaceTexture = view.getSurfaceTexture();
if (surfaceTexture != null) {
this.onSurfaceTextureAvailable(surfaceTexture,
mDesiredWidth, mDesiredHeight);
}
}
}
@Override
public void resize(int width, int height) {
if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.ICE_CREAM_SANDWICH_MR1) {
mTextureView.getSurfaceTexture().setDefaultBufferSize(width, height);
final SurfaceTexture surfaceTexture = mTextureView.getSurfaceTexture();
if (surfaceTexture != null) {
surfaceTexture.setDefaultBufferSize(width, height);
}
}
if (mRenderCallback != null) {
// the call above won't cause TextureView.onSurfaceTextureSizeChanged()
mRenderCallback.onResized(width, height);
}
// the call above won't cause TextureView.onSurfaceTextureSizeChanged()
mRenderCallback.onResized(width, height);
}
@Override
public void detach() {
setSurface(null);
mTextureView.setSurfaceTextureListener(null);
}
void setSurface(Surface surface) {
@Override
public void onSurfaceTextureAvailable(
@NonNull SurfaceTexture surfaceTexture, int width, int height) {
if (LOGGING) Log.d(LOG_TAG, "onSurfaceTextureAvailable()");
if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.ICE_CREAM_SANDWICH_MR1) {
if (mDesiredWidth > 0 && mDesiredHeight > 0) {
surfaceTexture.setDefaultBufferSize(mDesiredWidth, mDesiredHeight);
}
}
final Surface surface = new Surface(surfaceTexture);
setSurface(surface);
createSwapChain(surface);
if (mRenderCallback != null) {
// Call this the first time because onSurfaceTextureSizeChanged()
// isn't called at initialization time
mRenderCallback.onResized(width, height);
}
}
@Override
public void onSurfaceTextureSizeChanged(
@NonNull SurfaceTexture surfaceTexture, int width, int height) {
if (LOGGING) Log.d(LOG_TAG, "onSurfaceTextureSizeChanged()");
if (mRenderCallback != null) {
if (mDesiredWidth > 0 && mDesiredHeight > 0) {
surfaceTexture.setDefaultBufferSize(mDesiredWidth, mDesiredHeight);
mRenderCallback.onResized(mDesiredWidth, mDesiredHeight);
} else {
mRenderCallback.onResized(width, height);
}
// We must recreate the SwapChain to guarantee that it sees the new size.
// More precisely, for an EGL client, the EGLSurface must be recreated. For
// a Vulkan client, the SwapChain must be recreated. Calling
// onNativeWindowChanged() will accomplish that.
// This requirement comes from SurfaceTexture.setDefaultBufferSize()
// documentation.
final Surface surface = getSurface();
if (surface != null) {
mRenderCallback.onNativeWindowChanged(surface);
}
}
}
@Override
public boolean onSurfaceTextureDestroyed(@NonNull SurfaceTexture surfaceTexture) {
if (LOGGING) Log.d(LOG_TAG, "onSurfaceTextureDestroyed()");
setSurface(null);
destroySwapChain();
return true;
}
@Override
public void onSurfaceTextureUpdated(@NonNull SurfaceTexture surface) { }
private void setSurface(@Nullable Surface surface) {
if (surface == null) {
if (mSurface != null) {
mSurface.release();
@@ -245,7 +399,7 @@ public class UiHelper {
mSurface = surface;
}
public Surface getSurface() {
private Surface getSurface() {
return mSurface;
}
}
@@ -291,6 +445,9 @@ public class UiHelper {
* {@link #attachTo(TextureView)}, or {@link #attachTo(SurfaceHolder)}.
*/
public void detach() {
if (mRenderSurface != null) {
mRenderSurface.detach();
}
destroySwapChain();
mNativeWindow = null;
mRenderSurface = null;
@@ -298,7 +455,6 @@ public class UiHelper {
/**
* Checks whether we are ready to render into the attached surface.
*
* Using OpenGL ES when this returns true, will result in drawing commands being lost,
* HOWEVER, GLES state will be preserved. This is useful to initialize the engine.
*
@@ -343,7 +499,6 @@ public class UiHelper {
/**
* Controls whether the render target (SurfaceView or TextureView) is opaque or not.
* The render target is considered opaque by default.
*
* Must be called before calling {@link #attachTo(SurfaceView)}, {@link #attachTo(TextureView)},
* or {@link #attachTo(SurfaceHolder)}.
*
@@ -366,10 +521,8 @@ public class UiHelper {
* positioned above other surfaces but below the activity's surface. This property
* only has an effect when used in combination with {@link #setOpaque(boolean) setOpaque(false)}
* and does not affect TextureView targets.
*
* Must be called before calling {@link #attachTo(SurfaceView)}
* or {@link #attachTo(TextureView)}.
*
* Has no effect when using {@link #attachTo(SurfaceHolder)}.
*
* @param overlay Indicates whether the render target should be rendered below the activity's
@@ -390,7 +543,6 @@ public class UiHelper {
/**
* Associate UiHelper with a SurfaceView.
*
* As soon as SurfaceView is ready (i.e. has a Surface), we'll create the
* EGL resources needed, and call user callbacks if needed.
*/
@@ -405,171 +557,32 @@ public class UiHelper {
view.setZOrderOnTop(translucent);
}
int format = isOpaque() ? PixelFormat.OPAQUE : PixelFormat.TRANSLUCENT;
view.getHolder().setFormat(format);
view.getHolder().setFormat(isOpaque() ? PixelFormat.OPAQUE : PixelFormat.TRANSLUCENT);
mRenderSurface = new SurfaceViewHandler(view);
final SurfaceHolder.Callback callback = new SurfaceHolder.Callback() {
@Override
public void surfaceCreated(SurfaceHolder holder) {
if (LOGGING) Log.d(LOG_TAG, "surfaceCreated()");
createSwapChain(holder.getSurface());
}
@Override
public void surfaceChanged(
SurfaceHolder holder, int format, int width, int height) {
// Note: this is always called at least once after surfaceCreated()
if (LOGGING) Log.d(LOG_TAG, "surfaceChanged(" + width + ", " + height + ")");
mRenderCallback.onResized(width, height);
}
@Override
public void surfaceDestroyed(SurfaceHolder holder) {
if (LOGGING) Log.d(LOG_TAG, "surfaceDestroyed()");
destroySwapChain();
}
};
SurfaceHolder holder = view.getHolder();
holder.addCallback(callback);
if (mDesiredWidth > 0 && mDesiredHeight > 0) {
holder.setFixedSize(mDesiredWidth, mDesiredHeight);
}
// in case the SurfaceView's surface already existed
final Surface surface = holder.getSurface();
if (surface != null && surface.isValid()) {
callback.surfaceCreated(holder);
callback.surfaceChanged(holder, format,
holder.getSurfaceFrame().width(), holder.getSurfaceFrame().height());
}
}
}
/**
* Associate UiHelper with a TextureView.
*
* As soon as TextureView is ready (i.e. has a buffer), we'll create the
* EGL resources needed, and call user callbacks if needed.
*/
public void attachTo(@NonNull TextureView view) {
if (attach(view)) {
view.setOpaque(isOpaque());
mRenderSurface = new TextureViewHandler(view);
TextureView.SurfaceTextureListener listener = new TextureView.SurfaceTextureListener() {
@Override
public void onSurfaceTextureAvailable(
SurfaceTexture surfaceTexture, int width, int height) {
if (LOGGING) Log.d(LOG_TAG, "onSurfaceTextureAvailable()");
if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.ICE_CREAM_SANDWICH_MR1) {
if (mDesiredWidth > 0 && mDesiredHeight > 0) {
surfaceTexture.setDefaultBufferSize(mDesiredWidth, mDesiredHeight);
}
}
Surface surface = new Surface(surfaceTexture);
TextureViewHandler textureViewHandler = (TextureViewHandler) mRenderSurface;
textureViewHandler.setSurface(surface);
createSwapChain(surface);
// Call this the first time because onSurfaceTextureSizeChanged()
// isn't called at initialization time
mRenderCallback.onResized(width, height);
}
@Override
public void onSurfaceTextureSizeChanged(
SurfaceTexture surfaceTexture, int width, int height) {
if (LOGGING) Log.d(LOG_TAG, "onSurfaceTextureSizeChanged()");
if (mDesiredWidth > 0 && mDesiredHeight > 0) {
surfaceTexture.setDefaultBufferSize(mDesiredWidth, mDesiredHeight);
mRenderCallback.onResized(mDesiredWidth, mDesiredHeight);
} else {
mRenderCallback.onResized(width, height);
}
// We must recreate the SwapChain to guarantee that it sees the new size.
// More precisely, for an EGL client, the EGLSurface must be recreated. For
// a Vulkan client, the SwapChain must be recreated. Calling
// onNativeWindowChanged() will accomplish that.
// This requirement comes from SurfaceTexture.setDefaultBufferSize()
// documentation.
TextureViewHandler textureViewHandler = (TextureViewHandler) mRenderSurface;
mRenderCallback.onNativeWindowChanged(textureViewHandler.getSurface());
}
@Override
public boolean onSurfaceTextureDestroyed(SurfaceTexture surfaceTexture) {
if (LOGGING) Log.d(LOG_TAG, "onSurfaceTextureDestroyed()");
destroySwapChain();
return true;
}
@Override
public void onSurfaceTextureUpdated(SurfaceTexture surface) { }
};
view.setSurfaceTextureListener(listener);
// in case the View's SurfaceTexture already existed
if (view.isAvailable()) {
SurfaceTexture surfaceTexture = view.getSurfaceTexture();
listener.onSurfaceTextureAvailable(surfaceTexture, mDesiredWidth, mDesiredHeight);
}
}
}
/**
* Associate UiHelper with a SurfaceHolder.
*
* As soon as a Surface is created, we'll create the
* EGL resources needed, and call user callbacks if needed.
*/
public void attachTo(@NonNull SurfaceHolder holder) {
if (attach(holder)) {
int format = isOpaque() ? PixelFormat.OPAQUE : PixelFormat.TRANSLUCENT;
holder.setFormat(format);
holder.setFormat(isOpaque() ? PixelFormat.OPAQUE : PixelFormat.TRANSLUCENT);
mRenderSurface = new SurfaceHolderHandler(holder);
final SurfaceHolder.Callback callback = new SurfaceHolder.Callback() {
@Override
public void surfaceCreated(SurfaceHolder surfaceHolder) {
if (LOGGING) Log.d(LOG_TAG, "surfaceCreated()");
createSwapChain(holder.getSurface());
}
@Override
public void surfaceChanged(SurfaceHolder holder, int format, int width, int height) {
// Note: this is always called at least once after surfaceCreated()
if (LOGGING) Log.d(LOG_TAG, "surfaceChanged(" + width + ", " + height + ")");
mRenderCallback.onResized(width, height);
}
@Override
public void surfaceDestroyed(SurfaceHolder surfaceHolder) {
if (LOGGING) Log.d(LOG_TAG, "surfaceDestroyed()");
destroySwapChain();
}
};
holder.addCallback(callback);
if (mDesiredWidth > 0 && mDesiredHeight > 0) {
holder.setFixedSize(mDesiredWidth, mDesiredHeight);
}
// in case the SurfaceHolder's surface already existed
final Surface surface = holder.getSurface();
if (surface != null && surface.isValid()) {
callback.surfaceCreated(holder);
callback.surfaceChanged(holder, format,
holder.getSurfaceFrame().width(), holder.getSurfaceFrame().height());
}
}
}
@@ -580,6 +593,10 @@ public class UiHelper {
// nothing to do
return false;
}
if (mRenderSurface != null) {
mRenderSurface.detach();
mRenderSurface = null;
}
destroySwapChain();
}
mNativeWindow = nativeWindow;
@@ -587,15 +604,16 @@ public class UiHelper {
}
private void createSwapChain(@NonNull Surface surface) {
mRenderCallback.onNativeWindowChanged(surface);
if (mRenderCallback != null) {
mRenderCallback.onNativeWindowChanged(surface);
}
mHasSwapChain = true;
}
private void destroySwapChain() {
if (mRenderSurface != null) {
mRenderSurface.detach();
if (mRenderCallback != null) {
mRenderCallback.onDetachedFromSurface();
}
mRenderCallback.onDetachedFromSurface();
mHasSwapChain = false;
}
}

View File

@@ -12,9 +12,6 @@ android {
}
}
defaultConfig {
missingDimensionStrategy 'functionality', 'full'
}
packagingOptions {
// No need to package up the following shared libs, which arise as a side effect of our
// externalNativeBuild dependencies. When clients pick and choose from project-level gradle

View File

@@ -17,6 +17,8 @@
// Operators +, *, / based on http://half.sourceforge.net/ by Christian Rau
// and licensed under MIT
@file:Suppress("NOTHING_TO_INLINE")
package com.google.android.filament.utils
import com.google.android.filament.utils.Half.Companion.POSITIVE_INFINITY

View File

@@ -52,6 +52,7 @@ set(GLTFIO_SRCS
${GLTFIO_DIR}/include/gltfio/FilamentInstance.h
${GLTFIO_DIR}/include/gltfio/MaterialProvider.h
${GLTFIO_DIR}/include/gltfio/NodeManager.h
${GLTFIO_DIR}/include/gltfio/TrsTransformManager.h
${GLTFIO_DIR}/include/gltfio/ResourceLoader.h
${GLTFIO_DIR}/include/gltfio/TextureProvider.h
${GLTFIO_DIR}/include/gltfio/math.h
@@ -69,10 +70,12 @@ set(GLTFIO_SRCS
${GLTFIO_DIR}/src/FilamentAsset.cpp
${GLTFIO_DIR}/src/FilamentInstance.cpp
${GLTFIO_DIR}/src/FNodeManager.h
${GLTFIO_DIR}/src/FTrsTransformManager.h
${GLTFIO_DIR}/src/GltfEnums.h
${GLTFIO_DIR}/src/Ktx2Provider.cpp
${GLTFIO_DIR}/src/MaterialProvider.cpp
${GLTFIO_DIR}/src/NodeManager.cpp
${GLTFIO_DIR}/src/TrsTransformManager.cpp
${GLTFIO_DIR}/src/ResourceLoader.cpp
${GLTFIO_DIR}/src/StbProvider.cpp
${GLTFIO_DIR}/src/TangentsJob.cpp

View File

@@ -1,12 +1,6 @@
android {
namespace 'com.google.android.filament.gltfio'
flavorDimensions "functionality"
productFlavors {
full {
dimension "functionality"
}
}
packagingOptions {
// No need to package up the following shared libs, which arise as a side effect of our
// externalNativeBuild dependencies. When clients pick and choose from project-level gradle
@@ -18,7 +12,7 @@ android {
}
publishing {
singleVariant("fullRelease") {
singleVariant("release") {
withSourcesJar()
withJavadocJar()
}
@@ -36,9 +30,9 @@ apply from: rootProject.file('gradle/gradle-mvn-push.gradle')
afterEvaluate { project ->
publishing {
publications {
fullRelease(MavenPublication) {
release(MavenPublication) {
artifactId = POM_ARTIFACT_ID_FULL
from components.fullRelease
from components.release
}
}
}

View File

@@ -144,7 +144,7 @@ public class FilamentInstance {
*
* Ignored if variantIndex is out of bounds.
*/
void applyMaterialVariant(@IntRange(from = 0) int variantIndex) {
public void applyMaterialVariant(@IntRange(from = 0) int variantIndex) {
nApplyMaterialVariant(mNativeObject, variantIndex);
}

View File

@@ -1,5 +1,5 @@
GROUP=com.google.android.filament
VERSION_NAME=1.42.0
VERSION_NAME=1.50.3
POM_DESCRIPTION=Real-time physically based rendering engine for Android.

View File

@@ -1,6 +1,6 @@
#Wed Nov 17 10:40:18 PST 2021
distributionBase=GRADLE_USER_HOME
distributionPath=wrapper/dists
distributionUrl=https\://services.gradle.org/distributions/gradle-8.0-bin.zip
distributionUrl=https\://services.gradle.org/distributions/gradle-8.2-bin.zip
zipStoreBase=GRADLE_USER_HOME
zipStorePath=wrapper/dists

View File

@@ -87,9 +87,9 @@ compile Filament's native library and Filament's AAR for this project. The easie
is to install all the required dependencies and to run the following commands at the root of the
source tree:
```
$ ./build.sh -p desktop -i release
$ ./build.sh -p android release
```shell
./build.sh -p desktop -i release
./build.sh -p android release
```
This will build all the native components and the AAR required by this sample application.
@@ -100,8 +100,8 @@ distribution/install directory for desktop (produced by make/ninja install). Thi
contain `bin/matc` and `bin/cmgen`.
Example:
```
$ ./gradlew -Pfilament_tools_dir=../../dist-release assembleDebug
```shell
./gradlew -Pfilament_tools_dir=../../dist-release assembleDebug
```
## Important: SDK location
@@ -110,14 +110,24 @@ Either ensure your `ANDROID_HOME` environment variable is set or make sure the r
contains a `local.properties` file with the `sdk.dir` property pointing to your installation of
the Android SDK.
## Android Studio
## Compiling
### Android Studio
You must use the latest stable release of Android Studio. To open the project, point Studio to the
`android` folder. After opening the project and syncing to gradle, select the sample of your choice
using the drop-down widget in the toolbar.
## Compiling
To compile and run each sample make sure you have selected the appropriate build variant
(arm7, arm8, x86 or x86_64). If you are not sure you can simply select the "universal"
variant which includes all the other ones.
### Command Line
From the `android` directory in the project root:
```shell
./gradlew :samples:sample-hello-triangle:installDebug
```
Replace `sample-hello-triangle` with your preferred project.

View File

@@ -32,7 +32,6 @@ android {
applicationId "com.google.android.filament.gltf"
minSdkVersion 19
targetSdkVersion versions.targetSdk
missingDimensionStrategy 'functionality', 'full'
}
// NOTE: This is a workaround required because the AGP task collectReleaseDependencies

View File

@@ -26,6 +26,7 @@ import android.widget.TextView
import android.widget.Toast
import com.google.android.filament.Fence
import com.google.android.filament.IndirectLight
import com.google.android.filament.Material
import com.google.android.filament.Skybox
import com.google.android.filament.View
import com.google.android.filament.View.OnPickCallback
@@ -392,6 +393,36 @@ class MainActivity : Activity() {
Log.i(TAG, "The Filament backend took $total ms to load the model geometry.")
modelViewer.engine.destroyFence(it)
loadStartFence = null
val materials = mutableSetOf<Material>()
val rcm = modelViewer.engine.renderableManager
modelViewer.scene.forEach {
val entity = it
if (rcm.hasComponent(entity)) {
val ri = rcm.getInstance(entity)
val c = rcm.getPrimitiveCount(ri)
for (i in 0 until c) {
val mi = rcm.getMaterialInstanceAt(ri, i)
val ma = mi.material
materials.add(ma)
}
}
}
materials.forEach {
it.compile(
Material.CompilerPriorityQueue.HIGH,
Material.UserVariantFilterBit.DIRECTIONAL_LIGHTING or
Material.UserVariantFilterBit.DYNAMIC_LIGHTING or
Material.UserVariantFilterBit.SHADOW_RECEIVER,
null, null)
it.compile(
Material.CompilerPriorityQueue.LOW,
Material.UserVariantFilterBit.FOG or
Material.UserVariantFilterBit.SKINNING or
Material.UserVariantFilterBit.SSR or
Material.UserVariantFilterBit.VSM,
null, null)
}
}
}

View File

@@ -20,6 +20,8 @@ import android.animation.ValueAnimator
import android.app.Activity
import android.opengl.Matrix
import android.os.Bundle
import android.os.Handler
import android.os.Looper
import android.view.Choreographer
import android.view.Surface
import android.view.SurfaceView
@@ -110,7 +112,7 @@ class MainActivity : Activity() {
}
private fun setupFilament() {
engine = Engine.create()
engine = Engine.Builder().featureLevel(Engine.FeatureLevel.FEATURE_LEVEL_0).build()
renderer = engine.createRenderer()
scene = engine.createScene()
view = engine.createView()
@@ -120,13 +122,8 @@ class MainActivity : Activity() {
private fun setupView() {
scene.skybox = Skybox.Builder().color(0.035f, 0.035f, 0.035f, 1.0f).build(engine)
if (engine.activeFeatureLevel == Engine.FeatureLevel.FEATURE_LEVEL_0) {
// post-processing is not supported at feature level 0
view.isPostProcessingEnabled = false
} else {
// NOTE: Try to disable post-processing (tone-mapping, etc.) to see the difference
// view.isPostProcessingEnabled = false
}
// post-processing is not supported at feature level 0
view.isPostProcessingEnabled = false
// Tell the view which camera we want to use
view.camera = camera
@@ -160,12 +157,16 @@ class MainActivity : Activity() {
}
private fun loadMaterial() {
var name = "materials/baked_color.filamat"
if (engine.activeFeatureLevel == Engine.FeatureLevel.FEATURE_LEVEL_0) {
name = "materials/baked_color_es2.filamat"
}
readUncompressedAsset(name).let {
readUncompressedAsset("materials/baked_color.filamat").let {
material = Material.Builder().payload(it, it.remaining()).build(engine)
material.compile(
Material.CompilerPriorityQueue.HIGH,
Material.UserVariantFilterBit.ALL,
Handler(Looper.getMainLooper())) {
android.util.Log.i("hellotriangle",
"Material " + material.name + " compiled.")
}
engine.flush()
}
}

View File

@@ -17,7 +17,8 @@ material {
],
// This material disables all lighting
shadingModel : unlit
shadingModel : unlit,
featureLevel : 0
}
fragment {

View File

@@ -1,33 +0,0 @@
// Simple unlit material that uses the colors associated with each vertex.
//
// This source material must be compiled to a binary material using the matc tool.
// The command used to compile this material is:
// matc -p mobile -a opengl -o app/src/main/assets/baked_color.filamat app/src/materials/baked_color.mat
//
// See build.gradle for an example of how to compile materials automatically
// Please refer to the documentation for more information about matc and the materials system.
material {
name : baked_color,
// Lists the required vertex attributes
// Here we only need a color (RGBA)
requires : [
color
],
// This material disables all lighting
shadingModel : unlit,
featureLevel : 0
}
fragment {
void material(inout MaterialInputs material) {
// You must always call the prepareMaterial() function
prepareMaterial(material);
// We set the material's color to the color interpolated from
// the model's vertices
material.baseColor = getColor();
}
}

View File

@@ -32,13 +32,6 @@ android {
targetSdkVersion versions.targetSdk
}
// The filamat library has two variants: full and lite. Here we default to the "full" variant.
// Replace "full" with "lite" to use the filamat-lite variant, which has a smaller binary size
// but comes with certain restrictions. See "Filamat Lite" in libs/filamat/README.md.
defaultConfig {
missingDimensionStrategy 'functionality', 'full'
}
// NOTE: This is a workaround required because the AGP task collectReleaseDependencies
// is not configuration-cache friendly yet; this is only useful for Play publication
dependenciesInfo {

View File

@@ -33,7 +33,6 @@ android {
applicationId "com.google.android.filament.textured"
minSdkVersion versions.minSdk
targetSdkVersion versions.targetSdk
missingDimensionStrategy 'functionality', 'full'
}
// NOTE: This is a workaround required because the AGP task collectReleaseDependencies

104
build.sh
View File

@@ -191,6 +191,7 @@ function build_clean {
rm -Rf android/filamat-android/build android/filamat-android/.externalNativeBuild android/filamat-android/.cxx
rm -Rf android/gltfio-android/build android/gltfio-android/.externalNativeBuild android/gltfio-android/.cxx
rm -Rf android/filament-utils-android/build android/filament-utils-android/.externalNativeBuild android/filament-utils-android/.cxx
rm -f compile_commands.json
}
function build_clean_aggressive {
@@ -210,13 +211,10 @@ function build_desktop_target {
echo "Building ${lc_target} in out/cmake-${lc_target}..."
mkdir -p "out/cmake-${lc_target}"
cd "out/cmake-${lc_target}"
pushd "out/cmake-${lc_target}" > /dev/null
# On macOS, set the deployment target to 10.15.
local lc_name=$(echo "${UNAME}" | tr '[:upper:]' '[:lower:]')
if [[ "${lc_name}" == "darwin" ]]; then
local deployment_target="-DCMAKE_OSX_DEPLOYMENT_TARGET=10.15"
if [[ "${BUILD_UNIVERSAL_LIBRARIES}" == "true" ]]; then
local architectures="-DCMAKE_OSX_ARCHITECTURES=arm64;x86_64"
fi
@@ -233,9 +231,10 @@ function build_desktop_target {
${MATDBG_OPTION} \
${MATOPT_OPTION} \
${ASAN_UBSAN_OPTION} \
${deployment_target} \
${architectures} \
../..
ln -sf "out/cmake-${lc_target}/compile_commands.json" \
../../compile_commands.json
fi
${BUILD_COMMAND} ${build_targets}
@@ -247,12 +246,13 @@ function build_desktop_target {
if [[ -d "../${lc_target}/filament" ]]; then
if [[ "${ISSUE_ARCHIVES}" == "true" ]]; then
echo "Generating out/filament-${lc_target}-${LC_UNAME}.tgz..."
cd "../${lc_target}"
pushd "../${lc_target}" > /dev/null
tar -czvf "../filament-${lc_target}-${LC_UNAME}.tgz" filament
popd > /dev/null
fi
fi
cd ../..
popd > /dev/null
}
function build_desktop {
@@ -270,7 +270,7 @@ function build_webgl_with_target {
echo "Building WebGL ${lc_target}..."
mkdir -p "out/cmake-webgl-${lc_target}"
cd "out/cmake-webgl-${lc_target}"
pushd "out/cmake-webgl-${lc_target}" > /dev/null
if [[ ! "${BUILD_TARGETS}" ]]; then
BUILD_TARGETS=${BUILD_CUSTOM_TARGETS}
@@ -290,6 +290,8 @@ function build_webgl_with_target {
-DCMAKE_INSTALL_PREFIX="../webgl-${lc_target}/filament" \
-DWEBGL=1 \
../..
ln -sf "out/cmake-webgl-${lc_target}/compile_commands.json" \
../../compile_commands.json
${BUILD_COMMAND} ${BUILD_TARGETS}
)
fi
@@ -307,17 +309,17 @@ function build_webgl_with_target {
if [[ "${ISSUE_ARCHIVES}" == "true" ]]; then
echo "Generating out/filament-${lc_target}-web.tgz..."
cd web/filament-js
pushd web/filament-js > /dev/null
tar -cvf "../../../filament-${lc_target}-web.tar" filament.js
tar -rvf "../../../filament-${lc_target}-web.tar" filament.wasm
tar -rvf "../../../filament-${lc_target}-web.tar" filament.d.ts
cd -
popd > /dev/null
gzip -c "../filament-${lc_target}-web.tar" > "../filament-${lc_target}-web.tgz"
rm "../filament-${lc_target}-web.tar"
fi
fi
cd ../..
popd > /dev/null
}
function build_webgl {
@@ -348,7 +350,7 @@ function build_android_target {
echo "Building Android ${lc_target} (${arch})..."
mkdir -p "out/cmake-android-${lc_target}-${arch}"
cd "out/cmake-android-${lc_target}-${arch}"
pushd "out/cmake-android-${lc_target}-${arch}" > /dev/null
if [[ ! -d "CMakeFiles" ]] || [[ "${ISSUE_CMAKE_ALWAYS}" == "true" ]]; then
cmake \
@@ -362,12 +364,14 @@ function build_android_target {
${MATOPT_OPTION} \
${VULKAN_ANDROID_OPTION} \
../..
ln -sf "out/cmake-android-${lc_target}-${arch}/compile_commands.json" \
../../compile_commands.json
fi
# We must always install Android libraries to build the AAR
${BUILD_COMMAND} install
cd ../..
popd > /dev/null
}
function build_android_arch {
@@ -388,9 +392,9 @@ function archive_android {
if [[ -d "out/android-${lc_target}/filament" ]]; then
if [[ "${ISSUE_ARCHIVES}" == "true" ]]; then
echo "Generating out/filament-android-${lc_target}-${LC_UNAME}.tgz..."
cd "out/android-${lc_target}"
pushd "out/android-${lc_target}" > /dev/null
tar -czvf "../filament-android-${lc_target}-${LC_UNAME}.tgz" filament
cd ../..
popd > /dev/null
fi
fi
}
@@ -468,7 +472,7 @@ function build_android {
archive_android "Release"
fi
cd android
pushd android > /dev/null
if [[ "${ISSUE_DEBUG_BUILD}" == "true" ]]; then
./gradlew \
@@ -498,13 +502,13 @@ function build_android {
if [[ "${INSTALL_COMMAND}" ]]; then
echo "Installing out/filamat-android-debug.aar..."
cp filamat-android/build/outputs/aar/filamat-android-full-debug.aar ../out/filamat-android-debug.aar
cp filamat-android/build/outputs/aar/filamat-android-debug.aar ../out/filamat-android-debug.aar
echo "Installing out/filament-android-debug.aar..."
cp filament-android/build/outputs/aar/filament-android-debug.aar ../out/
echo "Installing out/gltfio-android-debug.aar..."
cp gltfio-android/build/outputs/aar/gltfio-android-full-debug.aar ../out/gltfio-android-debug.aar
cp gltfio-android/build/outputs/aar/gltfio-android-debug.aar ../out/gltfio-android-debug.aar
echo "Installing out/filament-utils-android-debug.aar..."
cp filament-utils-android/build/outputs/aar/filament-utils-android-debug.aar ../out/filament-utils-android-debug.aar
@@ -547,14 +551,13 @@ function build_android {
if [[ "${INSTALL_COMMAND}" ]]; then
echo "Installing out/filamat-android-release.aar..."
cp filamat-android/build/outputs/aar/filamat-android-lite-release.aar ../out/
cp filamat-android/build/outputs/aar/filamat-android-full-release.aar ../out/filamat-android-release.aar
cp filamat-android/build/outputs/aar/filamat-android-release.aar ../out/filamat-android-release.aar
echo "Installing out/filament-android-release.aar..."
cp filament-android/build/outputs/aar/filament-android-release.aar ../out/
echo "Installing out/gltfio-android-release.aar..."
cp gltfio-android/build/outputs/aar/gltfio-android-full-release.aar ../out/gltfio-android-release.aar
cp gltfio-android/build/outputs/aar/gltfio-android-release.aar ../out/gltfio-android-release.aar
echo "Installing out/filament-utils-android-release.aar..."
cp filament-utils-android/build/outputs/aar/filament-utils-android-release.aar ../out/filament-utils-android-release.aar
@@ -569,7 +572,7 @@ function build_android {
fi
fi
cd ..
popd > /dev/null
}
function build_ios_target {
@@ -580,7 +583,7 @@ function build_ios_target {
echo "Building iOS ${lc_target} (${arch}) for ${platform}..."
mkdir -p "out/cmake-ios-${lc_target}-${arch}"
cd "out/cmake-ios-${lc_target}-${arch}"
pushd "out/cmake-ios-${lc_target}-${arch}" > /dev/null
if [[ ! -d "CMakeFiles" ]] || [[ "${ISSUE_CMAKE_ALWAYS}" == "true" ]]; then
cmake \
@@ -595,6 +598,8 @@ function build_ios_target {
${MATDBG_OPTION} \
${MATOPT_OPTION} \
../..
ln -sf "out/cmake-ios-${lc_target}-${arch}/compile_commands.json" \
../../compile_commands.json
fi
${BUILD_COMMAND}
@@ -604,7 +609,7 @@ function build_ios_target {
${BUILD_COMMAND} ${INSTALL_COMMAND}
fi
cd ../..
popd > /dev/null
}
function archive_ios {
@@ -613,9 +618,9 @@ function archive_ios {
if [[ -d "out/ios-${lc_target}/filament" ]]; then
if [[ "${ISSUE_ARCHIVES}" == "true" ]]; then
echo "Generating out/filament-${lc_target}-ios.tgz..."
cd "out/ios-${lc_target}"
pushd "out/ios-${lc_target}" > /dev/null
tar -czvf "../filament-${lc_target}-ios.tgz" filament
cd ../..
popd > /dev/null
fi
fi
}
@@ -674,14 +679,14 @@ function build_web_docs {
mkdir -p out/web-docs
cp -f docs/web-docs-package.json out/web-docs/package.json
cd out/web-docs
pushd out/web-docs > /dev/null
npm install > /dev/null
# Generate documents
npx markdeep-rasterizer ../../docs/Filament.md.html ../../docs/Materials.md.html ../../docs/
cd ../..
popd > /dev/null
}
function validate_build_command {
@@ -754,6 +759,19 @@ function run_tests {
fi
}
function check_debug_release_build {
if [[ "${ISSUE_DEBUG_BUILD}" == "true" || \
"${ISSUE_RELEASE_BUILD}" == "true" || \
"${ISSUE_CLEAN}" == "true" || \
"${ISSUE_WEB_DOCS}" == "true" ]]; then
"$@";
else
echo "You must declare a debug or release target for $@ builds."
echo ""
exit 1
fi
}
# Beginning of the script
pushd "$(dirname "$0")" > /dev/null
@@ -762,7 +780,7 @@ while getopts ":hacCfgijmp:q:uvslwtedk:b" opt; do
case ${opt} in
h)
print_help
exit 1
exit 0
;;
a)
ISSUE_ARCHIVES=true
@@ -798,7 +816,7 @@ while getopts ":hacCfgijmp:q:uvslwtedk:b" opt; do
platforms=$(echo "${OPTARG}" | tr ',' '\n')
for platform in ${platforms}
do
case ${platform} in
case $(echo "${platform}" | tr '[:upper:]' '[:lower:]') in
desktop)
ISSUE_DESKTOP_BUILD=true
;;
@@ -817,6 +835,12 @@ while getopts ":hacCfgijmp:q:uvslwtedk:b" opt; do
ISSUE_DESKTOP_BUILD=true
ISSUE_WEBGL_BUILD=false
;;
*)
echo "Unknown platform ${platform}"
echo "Platform must be one of [desktop|android|ios|webgl|all]"
echo ""
exit 1
;;
esac
done
;;
@@ -829,7 +853,7 @@ while getopts ":hacCfgijmp:q:uvslwtedk:b" opt; do
abis=$(echo "${OPTARG}" | tr ',' '\n')
for abi in ${abis}
do
case ${abi} in
case $(echo "${abi}" | tr '[:upper:]' '[:lower:]') in
armeabi-v7a)
ABI_ARMEABI_V7A=true
;;
@@ -848,6 +872,12 @@ while getopts ":hacCfgijmp:q:uvslwtedk:b" opt; do
ABI_X86=true
ABI_X86_64=true
;;
*)
echo "Unknown abi ${abi}"
echo "ABI must be one of [armeabi-v7a|arm64-v8a|x86|x86_64|all]"
echo ""
exit 1
;;
esac
done
;;
@@ -911,9 +941,9 @@ fi
shift $((OPTIND - 1))
for arg; do
if [[ "${arg}" == "release" ]]; then
if [[ $(echo "${arg}" | tr '[:upper:]' '[:lower:]') == "release" ]]; then
ISSUE_RELEASE_BUILD=true
elif [[ "${arg}" == "debug" ]]; then
elif [[ $(echo "${arg}" | tr '[:upper:]' '[:lower:]') == "debug" ]]; then
ISSUE_DEBUG_BUILD=true
else
BUILD_CUSTOM_TARGETS="${BUILD_CUSTOM_TARGETS} ${arg}"
@@ -931,19 +961,19 @@ if [[ "${ISSUE_CLEAN_AGGRESSIVE}" == "true" ]]; then
fi
if [[ "${ISSUE_DESKTOP_BUILD}" == "true" ]]; then
build_desktop
check_debug_release_build build_desktop
fi
if [[ "${ISSUE_ANDROID_BUILD}" == "true" ]]; then
build_android
check_debug_release_build build_android
fi
if [[ "${ISSUE_IOS_BUILD}" == "true" ]]; then
build_ios
check_debug_release_build build_ios
fi
if [[ "${ISSUE_WEBGL_BUILD}" == "true" ]]; then
build_webgl
check_debug_release_build build_webgl
fi
if [[ "${ISSUE_WEB_DOCS}" == "true" ]]; then

View File

@@ -50,7 +50,7 @@ function replace {
FIND_STR="${1//\{\{VERSION\}\}/${VERSION_REGEX}}"
REPLACE_STR="${1//\{\{VERSION\}\}/${NEW_VERSION}}"
local FILE_NAME="$2"
if [ IS_DARWIN ]; then
if [ $IS_DARWIN == 1 ]; then
sed -i '' -E "s/${FIND_STR}/${REPLACE_STR}/" "${FILE_NAME}"
else
sed -i -E "s/${FIND_STR}/${REPLACE_STR}/" "${FILE_NAME}"

View File

@@ -5,7 +5,6 @@ libs/math/test_math
libs/image/test_image compare libs/image/tests/reference/
libs/utils/test_utils
libs/filamat/test_filamat
libs/filamat/test_filamat_lite
tools/matc/test_matc
tools/cmgen/test_cmgen compare
tools/glslminifier/test_glslminifier

View File

@@ -125,7 +125,13 @@ if [[ "${has_universal}" == "true" ]]; then
arch_output="${OUTPUT_PATH%.a}_${arch}.a"
arch_outputs+=("$arch_output")
combine_static_libs "$arch_output" $(find "$(pwd)/${archs_temp_dir}/${arch}" -iname '*.a')
archives=()
while IFS= read -r -d $'\0'; do
archives+=("$REPLY")
done < <(find "$(pwd)/${archs_temp_dir}/${arch}" -iname '*.a' -print0)
combine_static_libs "$arch_output" "$archives"
done
# Finally, combine the single-architecture archives into a universal binary.

File diff suppressed because one or more lines are too long

Binary file not shown.

View File

@@ -15,7 +15,7 @@
*/
// If you are bundling this with rollup, webpack, or esbuild, the following URL should be trimmed.
import { LitElement, html, css } from "https://unpkg.com/lit?module";
import { LitElement, html, css } from "https://unpkg.com/lit@2.8.0?module";
// This little utility checks if the Filament module is ready for action.
// If so, it immediately calls the given function. If not, it asks the Filament
@@ -287,12 +287,12 @@ class FilamentViewer extends LitElement {
// Dropping a glb file is simple because there are no external resources.
if (this.srcBlob && this.srcBlob.name.endsWith(".glb")) {
this.srcBlob.arrayBuffer().then(buffer => {
this.asset = this.loader.createAssetFromBinary(new Uint8Array(buffer));
this.asset = this.loader.createAsset(new Uint8Array(buffer));
const aabb = this.asset.getBoundingBox();
this.assetRoot = this.asset.getRoot();
this.unitCubeTransform = Filament.fitIntoUnitCube(aabb, zoffset);
this.asset.loadResources();
this.animator = this.asset.getAnimator();
this.animator = this.asset.getInstance().getAnimator();
this.animationStartTime = Date.now();
this._updateOverlay();
});
@@ -304,8 +304,6 @@ class FilamentViewer extends LitElement {
const config = {
normalizeSkinningWeights: true,
recomputeBoundingBoxes: false,
ignoreBindTransform: false,
asyncInterval: 30
};
@@ -320,22 +318,20 @@ class FilamentViewer extends LitElement {
resourceLoader.delete();
stbProvider.delete();
ktx2Provider.delete();
this.animator = this.asset.getAnimator();
this.animator = this.asset.getInstance().getAnimator();
this.animationStartTime = Date.now();
}
}, config.asyncInterval);
};
this.srcBlob.arrayBuffer().then(buffer => {
this.asset = this.loader.createAssetFromJson(new Uint8Array(buffer));
this.asset = this.loader.createAsset(new Uint8Array(buffer));
const aabb = this.asset.getBoundingBox();
this.assetRoot = this.asset.getRoot();
this.unitCubeTransform = Filament.fitIntoUnitCube(aabb, zoffset);
const resourceLoader = new Filament.gltfio$ResourceLoader(this.engine,
config.normalizeSkinningWeights,
config.recomputeBoundingBoxes,
config.ignoreBindTransform);
config.normalizeSkinningWeights);
const stbProvider = new Filament.gltfio$StbProvider(this.engine);
const ktx2Provider = new Filament.gltfio$Ktx2Provider(this.engine);
@@ -367,12 +363,7 @@ class FilamentViewer extends LitElement {
return response.arrayBuffer();
}).then(arrayBuffer => {
const modelData = new Uint8Array(arrayBuffer);
if (this.src.endsWith(".glb")) {
this.asset = this.loader.createAssetFromBinary(modelData);
} else {
this.asset = this.loader.createAssetFromJson(modelData);
}
this.asset = this.loader.createAsset(modelData);
const aabb = this.asset.getBoundingBox();
this.assetRoot = this.asset.getRoot();
this.unitCubeTransform = Filament.fitIntoUnitCube(aabb, zoffset);
@@ -380,7 +371,7 @@ class FilamentViewer extends LitElement {
const basePath = '' + new URL(this.src, document.location);
this.asset.loadResources(() => {
this.animator = this.asset.getAnimator();
this.animator = this.asset.getInstance().getAnimator();
this.animationStartTime = Date.now();
this._applyMaterialVariant();
}, null, basePath);
@@ -441,14 +432,15 @@ class FilamentViewer extends LitElement {
if (!this.hasAttribute("materialVariant")) {
return;
}
const names = this.asset.getMaterialVariantNames();
const instance = this.asset.getInstance();
const names = instance.getMaterialVariantNames();
const index = this.materialVariant;
if (index < 0 || index >= names.length) {
console.error(`Material variant ${index} does not exist in this asset.`);
return;
}
console.info(this.src, `Applying material variant: ${names[index]}`);
this.asset.applyMaterialVariant(index);
instance.applyMaterialVariant(index);
}
}

View File

@@ -43,7 +43,7 @@ filament-viewer::part(canvas) {
</p>
</main>
<script src="https://unpkg.com/filament@1.25.3/filament.js"></script>
<script src="https://unpkg.com/filament@1.44.0/filament.js"></script>
<script src="https://unpkg.com/gltumble"></script>
<script src="filament-viewer.js" type="module"></script>
</body>

File diff suppressed because one or more lines are too long

Binary file not shown.

Binary file not shown.

Binary file not shown.

View File

@@ -587,6 +587,7 @@
<li><em>overrides</em> Dictionary with one or more of the following properties: mapSize, shadowCascades, constantBias, normalBias, shadowFar, shadowNearHint, shadowFarHint, stable, polygonOffsetConstant, polygonOffsetSlope, \</li>
</ul>
</li>
<li><strong>engine.setStereoscopicOptions()</strong></li>
<li><strong>engine.setTemporalAntiAliasingOptions()</strong></li>
<li><strong>engine.setVignetteOptions()</strong></li>
</ul>
@@ -1462,6 +1463,8 @@ This defines the following functions:</p>
<li>STENCIL_ATTACHMENT</li>
<li>UPLOADABLE</li>
<li>SAMPLEABLE</li>
<li>BLIT_SRC</li>
<li>BLIT_DST</li>
<li>SUBPASS_INPUT</li>
</ul>
</div>

Binary file not shown.

Binary file not shown.

View File

@@ -22,27 +22,26 @@ converting <a href="https://github.com/google/filament/blob/main/assets/models/m
<p>Next, let&#x27;s create mipmapped KTX files using filament&#x27;s <code>mipgen</code> tool. We&#x27;ll create compressed and
non-compressed variants for each texture, since not all platforms support the same compression
formats. First copy over the PNG files from the <a href="https://github.com/google/filament/blob/main/assets/models/monkey">monkey folder</a>, then do:</p>
<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #3D7B7B; font-style: italic"># Create mipmaps for base color and two compressed variants.</span>
mipgen albedo.png albedo.ktx
mipgen --compression<span style="color: #666666">=</span>astc_fast_ldr_4x4 albedo.png albedo_astc.ktx
mipgen --compression<span style="color: #666666">=</span>s3tc_rgb_dxt1 albedo.png albedo_s3tc_srgb.ktx
<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #3D7B7B; font-style: italic"># Create mipmaps for base color</span>
mipgen albedo.png albedo.ktx2
mipgen --compression<span style="color: #666666">=</span>uastc albedo.png albedo.ktx2
<span style="color: #3D7B7B; font-style: italic"># Create mipmaps for the normal map and a compressed variant.</span>
mipgen --strip-alpha --kernel<span style="color: #666666">=</span>NORMALS --linear normal.png normal.ktx
mipgen --strip-alpha --kernel<span style="color: #666666">=</span>NORMALS --linear --compression<span style="color: #666666">=</span>etc_rgb8_normalxyz_40 <span style="color: #AA5D1F; font-weight: bold">\</span>
normal.png normal_etc.ktx
mipgen --strip-alpha --kernel<span style="color: #666666">=</span>NORMALS --linear --compression<span style="color: #666666">=</span>uastc_normals <span style="color: #AA5D1F; font-weight: bold">\</span>
normal.png normal.ktx2
<span style="color: #3D7B7B; font-style: italic"># Create mipmaps for the single-component roughness map and a compressed variant.</span>
mipgen --grayscale roughness.png roughness.ktx
mipgen --grayscale --compression<span style="color: #666666">=</span>etc_r11_numeric_40 roughness.png roughness_etc.ktx
mipgen --grayscale --compression<span style="color: #666666">=</span>uastc roughness.png roughness.ktx2
<span style="color: #3D7B7B; font-style: italic"># Create mipmaps for the single-component metallic map and a compressed variant.</span>
mipgen --grayscale metallic.png metallic.ktx
mipgen --grayscale --compression<span style="color: #666666">=</span>etc_r11_numeric_40 metallic.png metallic_etc.ktx
mipgen --grayscale --compression<span style="color: #666666">=</span>uastc metallic.png metallic.ktx2
<span style="color: #3D7B7B; font-style: italic"># Create mipmaps for the single-component occlusion map and a compressed variant.</span>
mipgen --grayscale ao.png ao.ktx
mipgen --grayscale --compression<span style="color: #666666">=</span>etc_r11_numeric_40 ao.png ao_etc.ktx
mipgen --grayscale --compression<span style="color: #666666">=</span>uastc ao.png ao.ktx2
</pre></div>
<p>For more information on mipgen&#x27;s arguments and supported formats, do <code>mipgen --help</code>.</p>
@@ -207,11 +206,11 @@ when the assets have finished downloading.</p>
recreate the skybox using a higher-resolution texture. As a last step we unhide the renderable that
was created in the app constructor.</p>
<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span>Filament.fetch([sky_large_url,<span style="color: #bbbbbb"> </span>albedo_url,<span style="color: #bbbbbb"> </span>roughness_url,<span style="color: #bbbbbb"> </span>metallic_url,<span style="color: #bbbbbb"> </span>normal_url,<span style="color: #bbbbbb"> </span>ao_url],<span style="color: #bbbbbb"> </span>()<span style="color: #bbbbbb"> </span>=&gt;<span style="color: #bbbbbb"> </span>{<span style="color: #bbbbbb"></span>
<span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">const</span><span style="color: #bbbbbb"> </span>albedo<span style="color: #bbbbbb"> </span><span style="color: #666666">=</span><span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">this</span>.engine.createTextureFromKtx1(albedo_url,<span style="color: #bbbbbb"> </span>{srgb<span style="color: #666666">:</span><span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">true</span>});<span style="color: #bbbbbb"></span>
<span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">const</span><span style="color: #bbbbbb"> </span>roughness<span style="color: #bbbbbb"> </span><span style="color: #666666">=</span><span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">this</span>.engine.createTextureFromKtx1(roughness_url);<span style="color: #bbbbbb"></span>
<span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">const</span><span style="color: #bbbbbb"> </span>metallic<span style="color: #bbbbbb"> </span><span style="color: #666666">=</span><span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">this</span>.engine.createTextureFromKtx1(metallic_url);<span style="color: #bbbbbb"></span>
<span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">const</span><span style="color: #bbbbbb"> </span>normal<span style="color: #bbbbbb"> </span><span style="color: #666666">=</span><span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">this</span>.engine.createTextureFromKtx1(normal_url);<span style="color: #bbbbbb"></span>
<span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">const</span><span style="color: #bbbbbb"> </span>ao<span style="color: #bbbbbb"> </span><span style="color: #666666">=</span><span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">this</span>.engine.createTextureFromKtx1(ao_url);<span style="color: #bbbbbb"></span>
<span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">const</span><span style="color: #bbbbbb"> </span>albedo<span style="color: #bbbbbb"> </span><span style="color: #666666">=</span><span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">this</span>.engine.createTextureFromKtx2(albedo_url,<span style="color: #bbbbbb"> </span>{srgb<span style="color: #666666">:</span><span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">true</span>});<span style="color: #bbbbbb"></span>
<span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">const</span><span style="color: #bbbbbb"> </span>roughness<span style="color: #bbbbbb"> </span><span style="color: #666666">=</span><span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">this</span>.engine.createTextureFromKtx2(roughness_url);<span style="color: #bbbbbb"></span>
<span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">const</span><span style="color: #bbbbbb"> </span>metallic<span style="color: #bbbbbb"> </span><span style="color: #666666">=</span><span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">this</span>.engine.createTextureFromKtx2(metallic_url);<span style="color: #bbbbbb"></span>
<span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">const</span><span style="color: #bbbbbb"> </span>normal<span style="color: #bbbbbb"> </span><span style="color: #666666">=</span><span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">this</span>.engine.createTextureFromKtx2(normal_url);<span style="color: #bbbbbb"></span>
<span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">const</span><span style="color: #bbbbbb"> </span>ao<span style="color: #bbbbbb"> </span><span style="color: #666666">=</span><span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">this</span>.engine.createTextureFromKtx2(ao_url);<span style="color: #bbbbbb"></span>
<span style="color: #bbbbbb"> </span><span style="color: #008000; font-weight: bold">const</span><span style="color: #bbbbbb"> </span>sampler<span style="color: #bbbbbb"> </span><span style="color: #666666">=</span><span style="color: #bbbbbb"> </span><span style="color: #AA22FF; font-weight: bold">new</span><span style="color: #bbbbbb"> </span>Filament.TextureSampler(<span style="color: #bbbbbb"></span>
<span style="color: #bbbbbb"> </span>Filament.MinFilter.LINEAR_MIPMAP_LINEAR,<span style="color: #bbbbbb"></span>

View File

@@ -33,13 +33,13 @@ class App {
});
Filament.fetch([sky_large_url, albedo_url, roughness_url, metallic_url, normal_url, ao_url],
() => {
const albedo = this.engine.createTextureFromKtx1(albedo_url, {
const albedo = this.engine.createTextureFromKtx2(albedo_url, {
srgb: true
});
const roughness = this.engine.createTextureFromKtx1(roughness_url);
const metallic = this.engine.createTextureFromKtx1(metallic_url);
const normal = this.engine.createTextureFromKtx1(normal_url);
const ao = this.engine.createTextureFromKtx1(ao_url);
const roughness = this.engine.createTextureFromKtx2(roughness_url);
const metallic = this.engine.createTextureFromKtx2(metallic_url);
const normal = this.engine.createTextureFromKtx2(normal_url);
const ao = this.engine.createTextureFromKtx2(ao_url);
const sampler = new Filament.TextureSampler(Filament.MinFilter.LINEAR_MIPMAP_LINEAR,
Filament.MagFilter.LINEAR, Filament.WrapMode.CLAMP_TO_EDGE);
this.matinstance.setTextureParameter('albedo', albedo, sampler);

View File

@@ -209,10 +209,13 @@ set(PRIVATE_HDRS
)
set(MATERIAL_SRCS
src/materials/fsr/fsr_easu.mat
src/materials/fsr/fsr_easu_mobile.mat
src/materials/fsr/fsr_easu_mobileF.mat
src/materials/fsr/fsr_rcas.mat
src/materials/antiAliasing/fxaa.mat
src/materials/antiAliasing/taa.mat
src/materials/blitLow.mat
src/materials/bloom/bloomDownsample.mat
src/materials/bloom/bloomDownsample2x.mat
src/materials/bloom/bloomDownsample9.mat
src/materials/bloom/bloomUpsample.mat
src/materials/colorGrading/colorGrading.mat
src/materials/colorGrading/colorGradingAsSubpass.mat
src/materials/colorGrading/customResolveAsSubpass.mat
@@ -220,30 +223,31 @@ set(MATERIAL_SRCS
src/materials/defaultMaterial.mat
src/materials/dof/dof.mat
src/materials/dof/dofCoc.mat
src/materials/dof/dofDownsample.mat
src/materials/dof/dofCombine.mat
src/materials/dof/dofDilate.mat
src/materials/dof/dofDownsample.mat
src/materials/dof/dofMedian.mat
src/materials/dof/dofMipmap.mat
src/materials/dof/dofTiles.mat
src/materials/dof/dofTilesSwizzle.mat
src/materials/dof/dofDilate.mat
src/materials/dof/dofMipmap.mat
src/materials/dof/dofMedian.mat
src/materials/flare/flare.mat
src/materials/blitLow.mat
src/materials/bloom/bloomDownsample.mat
src/materials/bloom/bloomUpsample.mat
src/materials/fsr/fsr_easu.mat
src/materials/fsr/fsr_easu_mobile.mat
src/materials/fsr/fsr_easu_mobileF.mat
src/materials/fsr/fsr_rcas.mat
src/materials/resolveDepth.mat
src/materials/separableGaussianBlur.mat
src/materials/skybox.mat
src/materials/shadowmap.mat
src/materials/ssao/bilateralBlur.mat
src/materials/ssao/bilateralBlurBentNormals.mat
src/materials/ssao/mipmapDepth.mat
src/materials/skybox.mat
src/materials/ssao/sao.mat
src/materials/ssao/saoBentNormals.mat
src/materials/separableGaussianBlur.mat
src/materials/antiAliasing/fxaa.mat
src/materials/antiAliasing/taa.mat
src/materials/vsmMipmap.mat
)
set(MATERIAL_ES2_SRCS
set(MATERIAL_FL0_SRCS
src/materials/defaultMaterial0.mat
src/materials/skybox0.mat
)
@@ -272,6 +276,11 @@ if (NOT DFG_LUT_SIZE)
endif()
message(STATUS "DFG LUT size set to ${DFG_LUT_SIZE}x${DFG_LUT_SIZE}")
# Whether to include FL0 materials.
if (FILAMENT_ENABLE_FEATURE_LEVEL_0)
add_definitions(-DFILAMENT_ENABLE_FEATURE_LEVEL_0)
endif()
# ==================================================================================================
# Definitions
# ==================================================================================================
@@ -313,15 +322,15 @@ foreach (mat_src ${MATERIAL_SRCS})
list(APPEND MATERIAL_BINS ${output_path})
endforeach()
if (IS_MOBILE_TARGET AND FILAMENT_SUPPORTS_OPENGL)
foreach (mat_src ${MATERIAL_ES2_SRCS})
if (FILAMENT_ENABLE_FEATURE_LEVEL_0)
foreach (mat_src ${MATERIAL_FL0_SRCS})
get_filename_component(localname "${mat_src}" NAME_WE)
get_filename_component(fullname "${mat_src}" ABSOLUTE)
set(output_path "${MATERIAL_DIR}/${localname}.filamat")
add_custom_command(
OUTPUT ${output_path}
COMMAND matc -a opengl -p ${MATC_TARGET} ${MATC_OPT_FLAGS} -o ${output_path} ${fullname}
COMMAND matc ${MATC_BASE_FLAGS} -o ${output_path} ${fullname}
MAIN_DEPENDENCY ${fullname}
DEPENDS matc
COMMENT "Compiling material ${mat_src} to ${output_path}"
@@ -555,7 +564,7 @@ if (MSVC)
set(FILAMENT_WARNINGS /W3)
else()
set(FILAMENT_WARNINGS
-Wall -Wextra -Wno-unused-parameter
-Wall -Wextra -Wno-unused-parameter -Wno-missing-field-initializers
-Wextra-semi -Wnewline-eof -Wdeprecated -Wundef
-Wgnu-conditional-omitted-operand
-Wweak-vtables -Wnon-virtual-dtor -Wclass-varargs -Wimplicit-fallthrough

View File

@@ -61,7 +61,7 @@ with the platform name, for example, `filament-20181009-linux.tgz`.
Create a file, `main.cpp`, in the same directory with the following contents:
```
```c++
#include <filament/FilamentAPI.h>
#include <filament/Engine.h>
@@ -91,7 +91,7 @@ Copy your platform's Makefile below into a `Makefile` inside the same directory.
### Linux
```
```make
FILAMENT_LIBS=-lfilament -lbackend -lbluegl -lbluevk -lfilabridge -lfilaflat -lutils -lgeometry -lsmol-v -lvkshaders -libl
CC=clang++
@@ -109,7 +109,7 @@ clean:
### macOS
```
```make
FILAMENT_LIBS=-lfilament -lbackend -lbluegl -lbluevk -lfilabridge -lfilaflat -lutils -lgeometry -lsmol-v -lvkshaders -libl
FRAMEWORKS=-framework Cocoa -framework Metal -framework CoreVideo
CC=clang++
@@ -137,7 +137,7 @@ be sure to also include `matdbg.lib` in `FILAMENT_LIBS`.
When building Filament from source, the `USE_STATIC_CRT` CMake option can be
used to change the run-time library version.
```
```make
FILAMENT_LIBS=filament.lib backend.lib bluegl.lib bluevk.lib filabridge.lib filaflat.lib \
utils.lib geometry.lib smol-v.lib ibl.lib vkshaders.lib
CC=cl.exe
@@ -171,12 +171,12 @@ and invoke `nmake` instead of `make`.
### Generating C++ documentation
To generate the documentation you must first install `doxygen` and `graphviz`, then run the
To generate the documentation you must first install `doxygen` and `graphviz`, then run the
following commands:
```
$ cd filament/filament
$ doxygen docs/doxygen/filament.doxygen
```shell
cd filament/filament
doxygen docs/doxygen/filament.doxygen
```
Finally simply open `docs/html/index.html` in your web browser.

View File

@@ -27,6 +27,7 @@ set(SRCS
src/BackendUtils.cpp
src/BlobCacheKey.cpp
src/Callable.cpp
src/CallbackManager.cpp
src/CircularBuffer.cpp
src/CommandBufferQueue.cpp
src/CommandStream.cpp
@@ -54,6 +55,7 @@ set(PRIVATE_HDRS
include/private/backend/HandleAllocator.h
include/private/backend/PlatformFactory.h
include/private/backend/SamplerGroup.h
src/CallbackManager.h
src/CommandStreamDispatcher.h
src/CompilerThreadPool.h
src/DataReshaper.h
@@ -67,8 +69,6 @@ set(PRIVATE_HDRS
if (FILAMENT_SUPPORTS_OPENGL AND NOT FILAMENT_USE_EXTERNAL_GLES3 AND NOT FILAMENT_USE_SWIFTSHADER)
list(APPEND SRCS
include/backend/platforms/OpenGLPlatform.h
src/opengl/CallbackManager.h
src/opengl/CallbackManager.cpp
src/opengl/gl_headers.cpp
src/opengl/gl_headers.h
src/opengl/GLUtils.cpp
@@ -133,6 +133,7 @@ if (FILAMENT_SUPPORTS_METAL)
src/metal/MetalExternalImage.mm
src/metal/MetalHandles.mm
src/metal/MetalPlatform.mm
src/metal/MetalShaderCompiler.mm
src/metal/MetalState.mm
src/metal/MetalTimerQuery.mm
src/metal/MetalUtils.mm
@@ -168,6 +169,8 @@ if (FILAMENT_SUPPORTS_VULKAN)
src/vulkan/platform/VulkanPlatform.cpp
src/vulkan/platform/VulkanPlatformSwapChainImpl.cpp
src/vulkan/platform/VulkanPlatformSwapChainImpl.h
src/vulkan/spirv/VulkanSpirvUtils.cpp
src/vulkan/spirv/VulkanSpirvUtils.h
src/vulkan/VulkanBlitter.cpp
src/vulkan/VulkanBlitter.h
src/vulkan/VulkanBuffer.cpp
@@ -312,6 +315,7 @@ endif()
if (FILAMENT_SUPPORTS_VULKAN)
target_link_libraries(${TARGET} PUBLIC bluevk vkmemalloc vkshaders smol-v)
target_link_libraries(${TARGET} PRIVATE SPIRV-Headers)
endif()
if (FILAMENT_SUPPORTS_METAL)
@@ -355,7 +359,7 @@ if (MSVC)
set(FILAMENT_WARNINGS /W3)
else()
set(FILAMENT_WARNINGS
-Wall -Wextra -Wno-unused-parameter
-Wall -Wextra -Wno-unused-parameter -Wno-missing-field-initializers
-Wextra-semi -Wnewline-eof -Wdeprecated -Wundef
-Wgnu-conditional-omitted-operand
-Wweak-vtables -Wnon-virtual-dtor -Wclass-varargs -Wimplicit-fallthrough
@@ -396,8 +400,8 @@ install(DIRECTORY ${PUBLIC_HDR_DIR}/backend DESTINATION include)
# ==================================================================================================
option(INSTALL_BACKEND_TEST "Install the backend test library so it can be consumed on iOS" OFF)
if (APPLE)
add_library(backend_test STATIC
if (APPLE OR LINUX)
set(BACKEND_TEST_SRC
test/BackendTest.cpp
test/ShaderGenerator.cpp
test/TrianglePrimitive.cpp
@@ -409,18 +413,26 @@ if (APPLE)
test/test_BufferUpdates.cpp
test/test_MRT.cpp
test/test_LoadImage.cpp
test/test_RenderExternalImage.cpp
test/test_StencilBuffer.cpp
test/test_Scissor.cpp
)
target_link_libraries(backend_test PRIVATE
test/test_MipLevels.cpp
)
set(BACKEND_TEST_LIBS
backend
getopt
gtest
imageio
filamat
SPIRV
spirv-cross-glsl)
endif()
if (APPLE)
# TODO: we should expand this test to Linux and other platforms.
list(APPEND BACKEND_TEST_SRC
test/test_RenderExternalImage.cpp)
add_library(backend_test STATIC ${BACKEND_TEST_SRC})
target_link_libraries(backend_test PRIVATE ${BACKEND_TEST_LIBS})
set(BACKEND_TEST_DEPS
OGLCompiler
@@ -434,12 +446,11 @@ if (APPLE)
glslang
spirv-cross-core
spirv-cross-glsl
spirv-cross-msl
)
spirv-cross-msl)
if (NOT IOS)
target_link_libraries(backend_test PRIVATE image imageio)
list(APPEND BACKEND_TEST_DEPS image imageio)
list(APPEND BACKEND_TEST_DEPS image)
endif()
set(BACKEND_TEST_COMBINED_OUTPUT "${CMAKE_CURRENT_BINARY_DIR}/libbackendtest_combined.a")
@@ -453,15 +464,21 @@ if (APPLE)
endif()
set_target_properties(backend_test PROPERTIES FOLDER Tests)
if (APPLE AND NOT IOS)
add_executable(backend_test_mac test/mac_runner.mm)
target_link_libraries(backend_test_mac PRIVATE "-framework Metal -framework AppKit -framework QuartzCore")
# Because each test case is a separate file, the -force_load flag is necessary to prevent the
# linker from removing "unused" symbols.
target_link_libraries(backend_test_mac PRIVATE -force_load backend_test)
set_target_properties(backend_test_mac PROPERTIES FOLDER Tests)
endif()
endif()
if (APPLE AND NOT IOS)
add_executable(backend_test_mac test/mac_runner.mm)
target_link_libraries(backend_test_mac PRIVATE "-framework Metal -framework AppKit -framework QuartzCore")
# Because each test case is a separate file, the -force_load flag is necessary to prevent the
# linker from removing "unused" symbols.
target_link_libraries(backend_test_mac PRIVATE -force_load backend_test)
set_target_properties(backend_test_mac PROPERTIES FOLDER Tests)
if (LINUX)
add_executable(backend_test_linux test/linux_runner.cpp ${BACKEND_TEST_SRC})
target_link_libraries(backend_test_linux PRIVATE ${BACKEND_TEST_LIBS})
set_target_properties(backend_test_linux PROPERTIES FOLDER Tests)
endif()
# ==================================================================================================

View File

@@ -23,7 +23,6 @@
#include <utils/ostream.h>
#include <stddef.h>
#include <stdint.h>
namespace filament::backend {
@@ -113,7 +112,7 @@ public:
/**
* Helper to create a BufferDescriptor that uses a KNOWN method pointer w/ object passed
* by pointer as the callback. e.g.:
* auto bd = BufferDescriptor::make(buffer, size, &Foo::method, foo);
* auto bd = BufferDescriptor::make<Foo, &Foo::method>(buffer, size, foo);
*
* @param buffer Memory address of the CPU buffer to reference
* @param size Size of the CPU buffer in bytes
@@ -121,12 +120,12 @@ public:
* @return a new BufferDescriptor
*/
template<typename T, void(T::*method)(void const*, size_t)>
static BufferDescriptor make(
void const* buffer, size_t size, T* data, CallbackHandler* handler = nullptr) noexcept {
static BufferDescriptor make(void const* buffer, size_t size, T* data,
CallbackHandler* handler = nullptr) noexcept {
return {
buffer, size,
handler, [](void* b, size_t s, void* u) {
(*static_cast<T**>(u)->*method)(b, s);
(static_cast<T*>(u)->*method)(b, s);
}, data
};
}
@@ -145,14 +144,14 @@ public:
* @return a new BufferDescriptor
*/
template<typename T>
static BufferDescriptor make(
void const* buffer, size_t size, T&& functor, CallbackHandler* handler = nullptr) noexcept {
static BufferDescriptor make(void const* buffer, size_t size, T&& functor,
CallbackHandler* handler = nullptr) noexcept {
return {
buffer, size,
handler, [](void* b, size_t s, void* u) {
T& that = *static_cast<T*>(u);
that(b, s);
delete &that;
T* const that = static_cast<T*>(u);
that->operator()(b, s);
delete that;
},
new T(std::forward<T>(functor))
};
@@ -201,7 +200,7 @@ public:
return mUser;
}
//! CPU mempry-buffer virtual address
//! CPU memory-buffer virtual address
void* buffer = nullptr;
//! CPU memory-buffer size in bytes

View File

@@ -17,8 +17,6 @@
#ifndef TNT_FILAMENT_BACKEND_CALLBACKHANDLER_H
#define TNT_FILAMENT_BACKEND_CALLBACKHANDLER_H
#include <utils/compiler.h>
namespace filament::backend {
/**

View File

@@ -28,7 +28,8 @@
#include <math/vec4.h>
#include <array> // FIXME: STL headers are not allowed in public headers
#include <array> // FIXME: STL headers are not allowed in public headers
#include <type_traits> // FIXME: STL headers are not allowed in public headers
#include <stddef.h>
#include <stdint.h>
@@ -77,6 +78,11 @@ static constexpr uint64_t SWAP_CHAIN_CONFIG_APPLE_CVPIXELBUFFER = 0x8;
*/
static constexpr uint64_t SWAP_CHAIN_CONFIG_SRGB_COLORSPACE = 0x10;
/**
* Indicates that the SwapChain should also contain a stencil component.
*/
static constexpr uint64_t SWAP_CHAIN_HAS_STENCIL_BUFFER = 0x20;
static constexpr size_t MAX_VERTEX_ATTRIBUTE_COUNT = 16; // This is guaranteed by OpenGL ES.
static constexpr size_t MAX_SAMPLER_COUNT = 62; // Maximum needed at feature level 3.
@@ -138,6 +144,30 @@ static constexpr const char* backendToString(Backend backend) {
}
}
/**
* Defines the shader language. Similar to the above backend enum, but the OpenGL backend can select
* between two shader languages: ESSL 1.0 and ESSL 3.0.
*/
enum class ShaderLanguage {
ESSL1 = 0,
ESSL3 = 1,
SPIRV = 2,
MSL = 3,
};
static constexpr const char* shaderLanguageToString(ShaderLanguage shaderLanguage) {
switch (shaderLanguage) {
case ShaderLanguage::ESSL1:
return "ESSL 1.0";
case ShaderLanguage::ESSL3:
return "ESSL 3.0";
case ShaderLanguage::SPIRV:
return "SPIR-V";
case ShaderLanguage::MSL:
return "MSL";
}
}
/**
* Bitmask for selecting render buffers
*/
@@ -198,6 +228,7 @@ struct Viewport {
int32_t top() const noexcept { return bottom + int32_t(height); }
};
/**
* Specifies the mapping of the near and far clipping plane to window coordinates.
*/
@@ -629,13 +660,15 @@ enum class TextureFormat : uint16_t {
//! Bitmask describing the intended Texture Usage
enum class TextureUsage : uint8_t {
NONE = 0x0,
COLOR_ATTACHMENT = 0x1, //!< Texture can be used as a color attachment
DEPTH_ATTACHMENT = 0x2, //!< Texture can be used as a depth attachment
STENCIL_ATTACHMENT = 0x4, //!< Texture can be used as a stencil attachment
UPLOADABLE = 0x8, //!< Data can be uploaded into this texture (default)
NONE = 0x00,
COLOR_ATTACHMENT = 0x01, //!< Texture can be used as a color attachment
DEPTH_ATTACHMENT = 0x02, //!< Texture can be used as a depth attachment
STENCIL_ATTACHMENT = 0x04, //!< Texture can be used as a stencil attachment
UPLOADABLE = 0x08, //!< Data can be uploaded into this texture (default)
SAMPLEABLE = 0x10, //!< Texture can be sampled (default)
SUBPASS_INPUT = 0x20, //!< Texture can be used as a subpass input
BLIT_SRC = 0x40, //!< Texture can be used the source of a blit()
BLIT_DST = 0x80, //!< Texture can be used the destination of a blit()
DEFAULT = UPLOADABLE | SAMPLEABLE //!< Default texture usage
};
@@ -663,6 +696,17 @@ static constexpr bool isDepthFormat(TextureFormat format) noexcept {
}
}
static constexpr bool isStencilFormat(TextureFormat format) noexcept {
switch (format) {
case TextureFormat::STENCIL8:
case TextureFormat::DEPTH24_STENCIL8:
case TextureFormat::DEPTH32F_STENCIL8:
return true;
default:
return false;
}
}
static constexpr bool isUnsignedIntFormat(TextureFormat format) {
switch (format) {
case TextureFormat::R8UI:
@@ -813,22 +857,22 @@ struct SamplerParams { // NOLINT
struct Hasher {
size_t operator()(SamplerParams p) const noexcept {
// we don't use std::hash<> here, so we don't have to include <functional>
return *reinterpret_cast<uint64_t const*>(reinterpret_cast<char const*>(&p));
return *reinterpret_cast<uint32_t const*>(reinterpret_cast<char const*>(&p));
}
};
struct EqualTo {
bool operator()(SamplerParams lhs, SamplerParams rhs) const noexcept {
auto* pLhs = reinterpret_cast<uint64_t const*>(reinterpret_cast<char const*>(&lhs));
auto* pRhs = reinterpret_cast<uint64_t const*>(reinterpret_cast<char const*>(&rhs));
auto* pLhs = reinterpret_cast<uint32_t const*>(reinterpret_cast<char const*>(&lhs));
auto* pRhs = reinterpret_cast<uint32_t const*>(reinterpret_cast<char const*>(&rhs));
return *pLhs == *pRhs;
}
};
struct LessThan {
bool operator()(SamplerParams lhs, SamplerParams rhs) const noexcept {
auto* pLhs = reinterpret_cast<uint64_t const*>(reinterpret_cast<char const*>(&lhs));
auto* pRhs = reinterpret_cast<uint64_t const*>(reinterpret_cast<char const*>(&rhs));
auto* pLhs = reinterpret_cast<uint32_t const*>(reinterpret_cast<char const*>(&lhs));
auto* pRhs = reinterpret_cast<uint32_t const*>(reinterpret_cast<char const*>(&rhs));
return *pLhs == *pRhs;
}
};
@@ -838,6 +882,7 @@ private:
return SamplerParams::LessThan{}(lhs, rhs);
}
};
static_assert(sizeof(SamplerParams) == 4);
// The limitation to 64-bits max comes from how we store a SamplerParams in our JNI code
// see android/.../TextureSampler.cpp
@@ -1150,8 +1195,8 @@ using FrameScheduledCallback = void(*)(PresentCallable callable, void* user);
enum class Workaround : uint16_t {
// The EASU pass must split because shader compiler flattens early-exit branch
SPLIT_EASU,
// Backend allows feedback loop with ancillary buffers (depth/stencil) as long as they're read-only for
// the whole render pass.
// Backend allows feedback loop with ancillary buffers (depth/stencil) as long as they're
// read-only for the whole render pass.
ALLOW_READ_ONLY_ANCILLARY_FEEDBACK_LOOP,
// for some uniform arrays, it's needed to do an initialization to avoid crash on adreno gpu
ADRENO_UNIFORM_ARRAY_CRASH,

View File

@@ -17,16 +17,14 @@
#ifndef TNT_FILAMENT_BACKEND_HANDLE_H
#define TNT_FILAMENT_BACKEND_HANDLE_H
#include <utils/compiler.h>
#if !defined(NDEBUG)
#include <utils/Log.h>
#include <utils/ostream.h>
#endif
#include <utils/debug.h>
#include <stdint.h>
#include <type_traits> // FIXME: STL headers are not allowed in public headers
#include <limits>
#include <type_traits>
#include <stdint.h>
namespace filament::backend {
@@ -54,7 +52,7 @@ struct HwVertexBuffer;
class HandleBase {
public:
using HandleId = uint32_t;
static constexpr const HandleId nullid = HandleId{ std::numeric_limits<HandleId>::max() };
static constexpr const HandleId nullid = HandleId{ UINT32_MAX };
constexpr HandleBase() noexcept: object(nullid) {}

View File

@@ -20,7 +20,7 @@
#include <backend/DriverEnums.h>
#include <backend/Handle.h>
#include <limits>
#include <utils/ostream.h>
#include <stdint.h>
@@ -33,10 +33,7 @@ struct PipelineState {
RasterState rasterState;
StencilState stencilState;
PolygonOffset polygonOffset;
Viewport scissor{ 0, 0,
(uint32_t)std::numeric_limits<int32_t>::max(),
(uint32_t)std::numeric_limits<int32_t>::max()
};
Viewport scissor{ 0, 0, (uint32_t)INT32_MAX, (uint32_t)INT32_MAX };
};
} // namespace filament::backend

View File

@@ -24,6 +24,7 @@
#include <utils/compiler.h>
#include <utils/debug.h>
#include <utils/ostream.h>
#include <stddef.h>
#include <stdint.h>
@@ -141,7 +142,7 @@ public:
CallbackHandler* handler = nullptr) noexcept {
return { buffer, size, format, type, alignment, left, top, stride,
handler, [](void* b, size_t s, void* u) {
(*static_cast<T**>(u)->*method)(b, s); }, data };
(static_cast<T*>(u)->*method)(b, s); }, data };
}
template<typename T, void(T::*method)(void const*, size_t)>
@@ -149,7 +150,7 @@ public:
PixelDataFormat format, PixelDataType type, T* data,
CallbackHandler* handler = nullptr) noexcept {
return { buffer, size, format, type, handler, [](void* b, size_t s, void* u) {
(*static_cast<T**>(u)->*method)(b, s); }, data };
(static_cast<T*>(u)->*method)(b, s); }, data };
}
template<typename T, void(T::*method)(void const*, size_t)>
@@ -157,7 +158,7 @@ public:
backend::CompressedPixelDataType format, uint32_t imageSize, T* data,
CallbackHandler* handler = nullptr) noexcept {
return { buffer, size, format, imageSize, handler, [](void* b, size_t s, void* u) {
(*static_cast<T**>(u)->*method)(b, s); }, data
(static_cast<T*>(u)->*method)(b, s); }, data
};
}
@@ -168,9 +169,9 @@ public:
CallbackHandler* handler = nullptr) noexcept {
return { buffer, size, format, type, alignment, left, top, stride,
handler, [](void* b, size_t s, void* u) {
T& that = *static_cast<T*>(u);
that(b, s);
delete &that;
T* const that = static_cast<T*>(u);
that->operator()(b, s);
delete that;
}, new T(std::forward<T>(functor))
};
}
@@ -181,9 +182,9 @@ public:
CallbackHandler* handler = nullptr) noexcept {
return { buffer, size, format, type,
handler, [](void* b, size_t s, void* u) {
T& that = *static_cast<T*>(u);
that(b, s);
delete &that;
T* const that = static_cast<T*>(u);
that->operator()(b, s);
delete that;
}, new T(std::forward<T>(functor))
};
}
@@ -194,9 +195,9 @@ public:
CallbackHandler* handler = nullptr) noexcept {
return { buffer, size, format, imageSize,
handler, [](void* b, size_t s, void* u) {
T& that = *static_cast<T*>(u);
that(b, s);
delete &that;
T* const that = static_cast<T*>(u);
that->operator()(b, s);
delete that;
}, new T(std::forward<T>(functor))
};
}

View File

@@ -19,11 +19,11 @@
#ifndef TNT_FILAMENT_BACKEND_PLATFORM_H
#define TNT_FILAMENT_BACKEND_PLATFORM_H
#include <backend/DriverEnums.h>
#include <utils/compiler.h>
#include <utils/Invocable.h>
#include <stddef.h>
namespace filament::backend {
class Driver;
@@ -41,11 +41,24 @@ public:
struct Stream {};
struct DriverConfig {
/*
* size of handle arena in bytes. Setting to 0 indicates default value is to be used.
/**
* Size of handle arena in bytes. Setting to 0 indicates default value is to be used.
* Driver clamps to valid values.
*/
size_t handleArenaSize = 0;
/**
* This number of most-recently destroyed textures will be tracked for use-after-free.
* Throws an exception when a texture is freed but still bound to a SamplerGroup and used in
* a draw call. 0 disables completely. Currently only respected by the Metal backend.
*/
size_t textureUseAfterFreePoolSize = 0;
/**
* Set to `true` to forcibly disable parallel shader compilation in the backend.
* Currently only honored by the GL backend.
*/
bool disableParallelShaderCompile = false;
};
Platform() noexcept;
@@ -71,7 +84,7 @@ public:
*
* @return nullptr on failure, or a pointer to the newly created driver.
*/
virtual backend::Driver* createDriver(void* sharedContext,
virtual backend::Driver* UTILS_NULLABLE createDriver(void* UTILS_NULLABLE sharedContext,
const DriverConfig& driverConfig) noexcept = 0;
/**
@@ -89,7 +102,8 @@ public:
* cache.
*/
using InsertBlobFunc = utils::Invocable<
void(const void* key, size_t keySize, const void* value, size_t valueSize)>;
void(const void* UTILS_NONNULL key, size_t keySize,
const void* UTILS_NONNULL value, size_t valueSize)>;
/*
* RetrieveBlobFunc is an Invocable to an application-provided function that a
@@ -97,7 +111,8 @@ public:
* cache.
*/
using RetrieveBlobFunc = utils::Invocable<
size_t(const void* key, size_t keySize, void* value, size_t valueSize)>;
size_t(const void* UTILS_NONNULL key, size_t keySize,
void* UTILS_NONNULL value, size_t valueSize)>;
/**
* Sets the callback functions that the backend can use to interact with caching functionality
@@ -107,6 +122,7 @@ public:
* Platform. The <insert> and <retrieve> Invocables may be called at any time and
* from any thread from the time at which setBlobFunc is called until the time that Platform
* is destroyed. Concurrent calls to these functions from different threads is also allowed.
* Either function can be null.
*
* @param insertBlob an Invocable that inserts a new value into the cache and associates
* it with the given key
@@ -116,9 +132,21 @@ public:
void setBlobFunc(InsertBlobFunc&& insertBlob, RetrieveBlobFunc&& retrieveBlob) noexcept;
/**
* @return true if setBlobFunc was called.
* @return true if insertBlob is valid.
*/
bool hasBlobFunc() const noexcept;
bool hasInsertBlobFunc() const noexcept;
/**
* @return true if retrieveBlob is valid.
*/
bool hasRetrieveBlobFunc() const noexcept;
/**
* @return true if either of insertBlob or retrieveBlob are valid.
*/
bool hasBlobFunc() const noexcept {
return hasInsertBlobFunc() || hasRetrieveBlobFunc();
}
/**
* To insert a new binary value into the cache and associate it with a given
@@ -137,7 +165,8 @@ public:
* @param value pointer to the beginning of the value data that is to be inserted
* @param valueSize specifies the size in byte of the data pointed to by <value>
*/
void insertBlob(const void* key, size_t keySize, const void* value, size_t valueSize);
void insertBlob(const void* UTILS_NONNULL key, size_t keySize,
const void* UTILS_NONNULL value, size_t valueSize);
/**
* To retrieve the binary value associated with a given key from the cache, a
@@ -156,7 +185,8 @@ public:
* @return If the cache contains a value associated with the given key then the
* size of that binary value in bytes is returned. Otherwise 0 is returned.
*/
size_t retrieveBlob(const void* key, size_t keySize, void* value, size_t valueSize);
size_t retrieveBlob(const void* UTILS_NONNULL key, size_t keySize,
void* UTILS_NONNULL value, size_t valueSize);
private:
InsertBlobFunc mInsertBlob;

View File

@@ -21,8 +21,7 @@
#include <utils/compiler.h>
namespace filament {
namespace backend {
namespace filament::backend {
/**
* A PresentCallable is a callable object that, when called, schedules a frame for presentation on
@@ -98,7 +97,6 @@ private:
*/
using FrameFinishedCallback UTILS_DEPRECATED = void(*)(PresentCallable callable, void* user);
} // namespace backend
} // namespace filament
} // namespace filament::backend
#endif // TNT_FILAMENT_BACKEND_PRESENTCALLABLE

View File

@@ -17,17 +17,19 @@
#ifndef TNT_FILAMENT_BACKEND_PRIVATE_PROGRAM_H
#define TNT_FILAMENT_BACKEND_PRIVATE_PROGRAM_H
#include <utils/compiler.h>
#include <utils/CString.h>
#include <utils/FixedCapacityVector.h>
#include <utils/Invocable.h>
#include <utils/Log.h>
#include <utils/ostream.h>
#include <backend/DriverEnums.h>
#include <array>
#include <variant>
#include <array> // FIXME: STL headers are not allowed in public headers
#include <utility> // FIXME: STL headers are not allowed in public headers
#include <variant> // FIXME: STL headers are not allowed in public headers
#include <stddef.h>
#include <stdint.h>
namespace filament::backend {
@@ -67,7 +69,7 @@ public:
Program& operator=(const Program& rhs) = delete;
Program(Program&& rhs) noexcept;
Program& operator=(Program&& rhs) noexcept;
Program& operator=(Program&& rhs) noexcept = delete;
~Program() noexcept;
@@ -104,8 +106,9 @@ public:
Sampler const* samplers, size_t count) noexcept;
struct SpecializationConstant {
using Type = std::variant<int32_t, float, bool>;
uint32_t id; // id set in glsl
std::variant<int32_t, float, bool> value; // value and type
Type value; // value and type
};
Program& specializationConstants(

View File

@@ -24,9 +24,6 @@
#include <utils/compiler.h>
#include <stddef.h>
#include <stdint.h>
namespace filament::backend {
struct UTILS_PUBLIC SamplerDescriptor {

View File

@@ -17,9 +17,11 @@
#ifndef TNT_FILAMENT_BACKEND_TARGETBUFFERINFO_H
#define TNT_FILAMENT_BACKEND_TARGETBUFFERINFO_H
#include <backend/DriverEnums.h>
#include <backend/Handle.h>
#include <utils/ostream.h>
#include <stddef.h>
#include <stdint.h>
namespace filament::backend {

View File

@@ -18,8 +18,13 @@
#define TNT_FILAMENT_BACKEND_PRIVATE_OPENGLPLATFORM_H
#include <backend/AcquiredImage.h>
#include <backend/DriverEnums.h>
#include <backend/Platform.h>
#include <utils/compiler.h>
#include <stdint.h>
namespace filament::backend {
class Driver;
@@ -38,8 +43,8 @@ protected:
* Derived classes can use this to instantiate the default OpenGLDriver backend.
* This is typically called from your implementation of createDriver()
*/
static Driver* createDefaultDriver(OpenGLPlatform* platform,
void* sharedContext, const DriverConfig& driverConfig);
static Driver* UTILS_NULLABLE createDefaultDriver(OpenGLPlatform* UTILS_NONNULL platform,
void* UTILS_NULLABLE sharedContext, const DriverConfig& driverConfig);
~OpenGLPlatform() noexcept override;
@@ -66,7 +71,8 @@ public:
* @return The driver's SwapChain object.
*
*/
virtual SwapChain* createSwapChain(void* nativeWindow, uint64_t flags) noexcept = 0;
virtual SwapChain* UTILS_NONNULL createSwapChain(
void* UTILS_NULLABLE nativeWindow, uint64_t flags) noexcept = 0;
/**
* Return whether createSwapChain supports the SWAP_CHAIN_CONFIG_SRGB_COLORSPACE flag.
@@ -87,13 +93,14 @@ public:
* TODO: we need a more generic way of passing construction parameters
* A void* might be enough.
*/
virtual SwapChain* createSwapChain(uint32_t width, uint32_t height, uint64_t flags) noexcept = 0;
virtual SwapChain* UTILS_NULLABLE createSwapChain(
uint32_t width, uint32_t height, uint64_t flags) noexcept = 0;
/**
* Called by the driver to destroys the SwapChain
* @param swapChain SwapChain to be destroyed.
*/
virtual void destroySwapChain(SwapChain* swapChain) noexcept = 0;
virtual void destroySwapChain(SwapChain* UTILS_NONNULL swapChain) noexcept = 0;
/**
* Returns the set of buffers that must be preserved up to the call to commit().
@@ -106,7 +113,7 @@ public:
* @return buffer that must be preserved
* @see commit()
*/
virtual TargetBufferFlags getPreservedFlags(SwapChain* swapChain) noexcept;
virtual TargetBufferFlags getPreservedFlags(SwapChain* UTILS_NONNULL swapChain) noexcept;
/**
* Called by the driver to establish the default FBO. The default implementation returns 0.
@@ -120,14 +127,16 @@ public:
* @param drawSwapChain SwapChain to draw to. It must be bound to the default FBO.
* @param readSwapChain SwapChain to read from (for operation like `glBlitFramebuffer`)
*/
virtual void makeCurrent(SwapChain* drawSwapChain, SwapChain* readSwapChain) noexcept = 0;
virtual void makeCurrent(
SwapChain* UTILS_NONNULL drawSwapChain,
SwapChain* UTILS_NONNULL readSwapChain) noexcept = 0;
/**
* Called by the driver once the current frame finishes drawing. Typically, this should present
* the drawSwapChain. This is for example where `eglMakeCurrent()` would be called.
* @param swapChain the SwapChain to present.
*/
virtual void commit(SwapChain* swapChain) noexcept = 0;
virtual void commit(SwapChain* UTILS_NONNULL swapChain) noexcept = 0;
/**
* Set the time the next committed buffer should be presented to the user at.
@@ -152,14 +161,14 @@ public:
*
* @return A Fence object. The default implementation returns nullptr.
*/
virtual Fence* createFence() noexcept;
virtual Fence* UTILS_NULLABLE createFence() noexcept;
/**
* Destroys a Fence object. The default implementation does nothing.
*
* @param fence Fence to destroy.
*/
virtual void destroyFence(Fence* fence) noexcept;
virtual void destroyFence(Fence* UTILS_NONNULL fence) noexcept;
/**
* Waits on a Fence.
@@ -169,7 +178,7 @@ public:
* @return Whether the fence signaled or timed out. See backend::FenceStatus.
* The default implementation always return backend::FenceStatus::ERROR.
*/
virtual backend::FenceStatus waitFence(Fence* fence, uint64_t timeout) noexcept;
virtual backend::FenceStatus waitFence(Fence* UTILS_NONNULL fence, uint64_t timeout) noexcept;
// --------------------------------------------------------------------------------------------
@@ -183,13 +192,13 @@ public:
* @param nativeStream The native stream, this parameter depends on the concrete implementation.
* @return A new Stream object.
*/
virtual Stream* createStream(void* nativeStream) noexcept;
virtual Stream* UTILS_NULLABLE createStream(void* UTILS_NULLABLE nativeStream) noexcept;
/**
* Destroys a Stream.
* @param stream Stream to destroy.
*/
virtual void destroyStream(Stream* stream) noexcept;
virtual void destroyStream(Stream* UTILS_NONNULL stream) noexcept;
/**
* The specified stream takes ownership of the texture (tname) object
@@ -199,20 +208,21 @@ public:
* @param stream Stream to take ownership of the texture
* @param tname GL texture id to "bind" to the Stream.
*/
virtual void attach(Stream* stream, intptr_t tname) noexcept;
virtual void attach(Stream* UTILS_NONNULL stream, intptr_t tname) noexcept;
/**
* Destroys the texture associated to the stream
* @param stream Stream to detach from its texture
*/
virtual void detach(Stream* stream) noexcept;
virtual void detach(Stream* UTILS_NONNULL stream) noexcept;
/**
* Updates the content of the texture attached to the stream.
* @param stream Stream to update
* @param timestamp Output parameter: Timestamp of the image bound to the texture.
*/
virtual void updateTexImage(Stream* stream, int64_t* timestamp) noexcept;
virtual void updateTexImage(Stream* UTILS_NONNULL stream,
int64_t* UTILS_NONNULL timestamp) noexcept;
// --------------------------------------------------------------------------------------------
@@ -225,13 +235,13 @@ public:
* implementation could just return { 0, GL_TEXTURE_2D } at this point. The actual
* values can be delayed until setExternalImage.
*/
virtual ExternalTexture *createExternalImageTexture() noexcept;
virtual ExternalTexture* UTILS_NULLABLE createExternalImageTexture() noexcept;
/**
* Destroys an external texture handle and associated data.
* @param texture a pointer to the handle to destroy.
*/
virtual void destroyExternalImage(ExternalTexture* texture) noexcept;
virtual void destroyExternalImage(ExternalTexture* UTILS_NONNULL texture) noexcept;
// called on the application thread to allow Filament to take ownership of the image
@@ -244,7 +254,7 @@ public:
* @param externalImage A token representing the platform's external image.
* @see destroyExternalImage
*/
virtual void retainExternalImage(void* externalImage) noexcept;
virtual void retainExternalImage(void* UTILS_NONNULL externalImage) noexcept;
/**
* Called to bind the platform-specific externalImage to an ExternalTexture.
@@ -258,7 +268,8 @@ public:
* @param texture an in/out pointer to ExternalTexture, id and target can be updated if necessary.
* @return true on success, false on error.
*/
virtual bool setExternalImage(void* externalImage, ExternalTexture* texture) noexcept;
virtual bool setExternalImage(void* UTILS_NONNULL externalImage,
ExternalTexture* UTILS_NONNULL texture) noexcept;
/**
* The method allows platforms to convert a user-supplied external image object into a new type

View File

@@ -17,11 +17,10 @@
#ifndef TNT_FILAMENT_BACKEND_OPENGL_OPENGL_PLATFORM_COCOA_GL_H
#define TNT_FILAMENT_BACKEND_OPENGL_OPENGL_PLATFORM_COCOA_GL_H
#include <stdint.h>
#include <backend/DriverEnums.h>
#include <backend/platforms/OpenGLPlatform.h>
#include <backend/DriverEnums.h>
#include <stdint.h>
namespace filament::backend {

View File

@@ -30,7 +30,7 @@ struct PlatformCocoaTouchGLImpl;
class PlatformCocoaTouchGL : public OpenGLPlatform {
public:
PlatformCocoaTouchGL();
~PlatformCocoaTouchGL() noexcept;
~PlatformCocoaTouchGL() noexcept override;
// --------------------------------------------------------------------------------------------
// Platform Interface

View File

@@ -17,18 +17,21 @@
#ifndef TNT_FILAMENT_BACKEND_OPENGL_OPENGL_PLATFORM_EGL_H
#define TNT_FILAMENT_BACKEND_OPENGL_OPENGL_PLATFORM_EGL_H
#include <stdint.h>
#include <backend/DriverEnums.h>
#include <backend/Platform.h>
#include <backend/platforms/OpenGLPlatform.h>
#include <EGL/egl.h>
#include <EGL/eglext.h>
#include <EGL/eglplatform.h>
#include <backend/platforms/OpenGLPlatform.h>
#include <backend/DriverEnums.h>
#include <initializer_list>
#include <utility>
#include <vector>
#include <stddef.h>
#include <stdint.h>
namespace filament::backend {
/**
@@ -42,6 +45,9 @@ public:
void createContext(bool shared) override;
void releaseContext() noexcept override;
// Return true if we're on an OpenGL platform (as opposed to OpenGL ES). false by default.
virtual bool isOpenGL() const noexcept;
protected:
// --------------------------------------------------------------------------------------------
@@ -126,6 +132,7 @@ protected:
EGLSurface mCurrentDrawSurface = EGL_NO_SURFACE;
EGLSurface mCurrentReadSurface = EGL_NO_SURFACE;
EGLSurface mEGLDummySurface = EGL_NO_SURFACE;
// mEGLConfig is valid only if ext.egl.KHR_no_config_context is false
EGLConfig mEGLConfig = EGL_NO_CONFIG_KHR;
Config mContextAttribs;
std::vector<EGLContext> mAdditionalContexts;
@@ -146,8 +153,8 @@ protected:
void initializeGlExtensions() noexcept;
private:
EGLConfig findSwapChainConfig(uint64_t flags) const;
protected:
EGLConfig findSwapChainConfig(uint64_t flags, bool window, bool pbuffer) const;
};
} // namespace filament::backend

View File

@@ -17,8 +17,14 @@
#ifndef TNT_FILAMENT_BACKEND_OPENGL_OPENGL_PLATFORM_EGL_ANDROID_H
#define TNT_FILAMENT_BACKEND_OPENGL_OPENGL_PLATFORM_EGL_ANDROID_H
#include <backend/AcquiredImage.h>
#include <backend/Platform.h>
#include <backend/platforms/OpenGLPlatform.h>
#include <backend/platforms/PlatformEGL.h>
#include <stddef.h>
#include <stdint.h>
namespace filament::backend {
class ExternalStreamManagerAndroid;

View File

@@ -30,6 +30,9 @@ public:
Driver* createDriver(void* sharedContext,
const Platform::DriverConfig& driverConfig) noexcept override;
protected:
bool isOpenGL() const noexcept override;
};
} // namespace filament

View File

@@ -20,12 +20,18 @@
#include <backend/Platform.h>
#include <bluevk/BlueVK.h>
#include <utils/CString.h>
#include <utils/FixedCapacityVector.h>
#include <utils/PrivateImplementation.h>
#include <string_view>
#include <tuple>
#include <unordered_set>
#include <stddef.h>
#include <stdint.h>
namespace filament::backend {
using SwapChain = Platform::SwapChain;
@@ -89,36 +95,44 @@ public:
// ----------------------------------------------------
// ---------- Platform Customization options ----------
/**
* The client preference can be stored within the struct. We allow for two specification of
* preference:
* 1) A substring to match against `VkPhysicalDeviceProperties.deviceName`.
* 2) Index of the device in the list as returned by vkEnumeratePhysicalDevices.
*/
struct GPUPreference {
std::string deviceName;
int8_t index = -1;
struct Customization {
/**
* The client can specify the GPU (i.e. VkDevice) for the platform. We allow the
* following preferences:
* 1) A substring to match against `VkPhysicalDeviceProperties.deviceName`. Empty string
* by default.
* 2) Index of the device in the list as returned by
* `vkEnumeratePhysicalDevices`. -1 by default to indicate no preference.
*/
struct GPUPreference {
utils::CString deviceName;
int8_t index = -1;
} gpu;
/**
* Whether the platform supports sRGB swapchain. Default is true.
*/
bool isSRGBSwapChainSupported = true;
/**
* When the platform window is resized, we will flush and wait on the command queues
* before recreating the swapchain. Default is true.
*/
bool flushAndWaitOnWindowResize = true;
};
/**
* Client can provide a preference over the GPU to use in the vulkan instance
* @return `GPUPreference` struct that indicates the client's preference
* Client can override to indicate customized behavior or parameter for their platform.
* @return `Customization` struct that indicates the client's platform
* customizations.
*/
virtual GPUPreference getPreferredGPU() noexcept {
virtual Customization getCustomization() const noexcept {
return {};
}
// -------- End platform customization options --------
// ----------------------------------------------------
/**
* Returns whether the platform supports sRGB swapchain. This is true by default, and the client
* needs to override this method to specify otherwise.
* @return Whether the platform supports sRGB swapchain.
*/
virtual bool isSRGBSwapChainSupported() const {
return true;
}
/**
* Get the images handles and format of the memory backing the swapchain. This should be called
* after createSwapChain() or after recreateIfResized().

View File

@@ -17,10 +17,7 @@
#ifndef TNT_FILAMENT_BACKEND_PRIVATE_CIRCULARBUFFER_H
#define TNT_FILAMENT_BACKEND_PRIVATE_CIRCULARBUFFER_H
#include <utils/compiler.h>
#include <stddef.h>
#include <stdint.h>
namespace filament::backend {

View File

@@ -19,12 +19,14 @@
#include "private/backend/CircularBuffer.h"
#include <utils/compiler.h>
#include <utils/Condition.h>
#include <utils/Mutex.h>
#include <vector>
#include <stddef.h>
#include <stdint.h>
namespace filament::backend {
/*

View File

@@ -32,11 +32,13 @@
#include <backend/TargetBufferInfo.h>
#include <utils/compiler.h>
#include <utils/debug.h>
#include <utils/ThreadUtils.h>
#include <cstddef>
#include <functional>
#include <tuple>
#include <type_traits>
#include <utility>
#ifndef NDEBUG
@@ -152,14 +154,14 @@ struct CommandType<void (Driver::*)(ARGS...)> {
}
// placement new declared as "throw" to avoid the compiler's null-check
inline void* operator new(std::size_t size, void* ptr) {
inline void* operator new(std::size_t, void* ptr) {
assert_invariant(ptr);
return ptr;
}
};
};
// convert an method of "class Driver" into a Command<> type
// convert a method of "class Driver" into a Command<> type
#define COMMAND_TYPE(method) CommandType<decltype(&Driver::method)>::Command<&Driver::method>
// ------------------------------------------------------------------------------------------------

View File

@@ -24,10 +24,12 @@
#include <backend/PipelineState.h>
#include <backend/TargetBufferInfo.h>
#include <utils/CString.h>
#include <utils/compiler.h>
#include <functional>
#include <stddef.h>
#include <stdint.h>
// Command debugging off. debugging virtuals are not called.

View File

@@ -219,7 +219,7 @@ DECL_DRIVER_API_R_N(backend::TextureHandle, importTexture,
backend::TextureUsage, usage)
DECL_DRIVER_API_R_N(backend::SamplerGroupHandle, createSamplerGroup,
uint32_t, size)
uint32_t, size, utils::FixedSizeString<32>, debugName)
DECL_DRIVER_API_R_N(backend::RenderPrimitiveHandle, createRenderPrimitive,
backend::VertexBufferHandle, vbh,
@@ -300,10 +300,10 @@ DECL_DRIVER_API_SYNCHRONOUS_0(bool, isAutoDepthResolveSupported)
DECL_DRIVER_API_SYNCHRONOUS_0(bool, isSRGBSwapChainSupported)
DECL_DRIVER_API_SYNCHRONOUS_0(bool, isStereoSupported)
DECL_DRIVER_API_SYNCHRONOUS_0(bool, isParallelShaderCompileSupported)
DECL_DRIVER_API_SYNCHRONOUS_0(bool, isDepthStencilResolveSupported)
DECL_DRIVER_API_SYNCHRONOUS_0(uint8_t, getMaxDrawBuffers)
DECL_DRIVER_API_SYNCHRONOUS_0(size_t, getMaxUniformBufferSize)
DECL_DRIVER_API_SYNCHRONOUS_0(math::float2, getClipSpaceParams)
DECL_DRIVER_API_SYNCHRONOUS_0(bool, canGenerateMipmaps)
DECL_DRIVER_API_SYNCHRONOUS_N(void, setupExternalImage, void*, image)
DECL_DRIVER_API_SYNCHRONOUS_N(bool, getTimerQueryValue, backend::TimerQueryHandle, query, uint64_t*, elapsedTime)
DECL_DRIVER_API_SYNCHRONOUS_N(bool, isWorkaroundNeeded, backend::Workaround, workaround)
@@ -468,7 +468,7 @@ DECL_DRIVER_API_N(readBufferSubData,
* --------------------
*/
DECL_DRIVER_API_N(blit,
DECL_DRIVER_API_N(blitDEPRECATED,
backend::TargetBufferFlags, buffers,
backend::RenderTargetHandle, dst,
backend::Viewport, dstRect,
@@ -476,6 +476,21 @@ DECL_DRIVER_API_N(blit,
backend::Viewport, srcRect,
backend::SamplerMagFilter, filter)
DECL_DRIVER_API_N(resolve,
backend::TextureHandle, dst,
uint8_t, dstLevel, uint8_t, dstLayer,
backend::TextureHandle, src,
uint8_t, srcLevel, uint8_t, srcLayer)
DECL_DRIVER_API_N(blit,
backend::TextureHandle, dst,
uint8_t, dstLevel, uint8_t, dstLayer,
math::uint2, dstOrigin,
backend::TextureHandle, src,
uint8_t, srcLevel, uint8_t, srcLayer,
math::uint2, srcOrigin,
math::uint2, size)
DECL_DRIVER_API_N(draw,
backend::PipelineState, state,
backend::RenderPrimitiveHandle, rph,

View File

@@ -22,9 +22,18 @@
#include <utils/Allocator.h>
#include <utils/Log.h>
#include <utils/compiler.h>
#include <utils/debug.h>
#include <utils/ostream.h>
#include <tsl/robin_map.h>
#include <exception>
#include <type_traits>
#include <unordered_map>
#include <stddef.h>
#include <stdint.h>
#if !defined(NDEBUG) && UTILS_HAS_RTTI
# define HANDLE_TYPE_SAFETY 1
#else
@@ -33,7 +42,7 @@
#define HandleAllocatorGL HandleAllocator<16, 64, 208>
#define HandleAllocatorVK HandleAllocator<16, 64, 880>
#define HandleAllocatorMTL HandleAllocator<16, 64, 576>
#define HandleAllocatorMTL HandleAllocator<16, 64, 584>
namespace filament::backend {
@@ -222,7 +231,7 @@ private:
explicit Allocator(const utils::AreaPolicy::HeapArea& area);
// this is in fact always called with a constexpr size argument
[[nodiscard]] inline void* alloc(size_t size, size_t alignment, size_t extra) noexcept {
[[nodiscard]] inline void* alloc(size_t size, size_t, size_t extra) noexcept {
void* p = nullptr;
if (size <= mPool0.getSize()) p = mPool0.alloc(size, 16, extra);
else if (size <= mPool1.getSize()) p = mPool1.alloc(size, 16, extra);
@@ -239,14 +248,16 @@ private:
}
};
// FIXME: We should be using a Spinlock here, at least on platforms where mutexes are not
// efficient (i.e. non-Linux). However, we've seen some hangs on that spinlock, which
// we don't understand well (b/308029108).
#ifndef NDEBUG
using HandleArena = utils::Arena<Allocator,
utils::LockingPolicy::SpinLock,
utils::LockingPolicy::Mutex,
utils::TrackingPolicy::DebugAndHighWatermark>;
#else
using HandleArena = utils::Arena<Allocator,
utils::LockingPolicy::SpinLock>;
utils::LockingPolicy::Mutex>;
#endif
// allocateHandle()/deallocateHandle() selects the pool to use at compile-time based on the
@@ -256,6 +267,7 @@ private:
HandleBase::HandleId allocateHandle() noexcept {
if constexpr (SIZE <= P0) { return allocateHandleInPool<P0>(); }
if constexpr (SIZE <= P1) { return allocateHandleInPool<P1>(); }
static_assert(SIZE <= P2);
return allocateHandleInPool<P2>();
}
@@ -266,6 +278,7 @@ private:
} else if constexpr (SIZE <= P1) {
deallocateHandleFromPool<P1>(id);
} else {
static_assert(SIZE <= P2);
deallocateHandleFromPool<P2>(id);
}
}

View File

@@ -19,9 +19,9 @@
#ifndef TNT_FILAMENT_BACKEND_PLATFORM_FACTORY_H
#define TNT_FILAMENT_BACKEND_PLATFORM_FACTORY_H
#include "backend/DriverEnums.h"
#include <backend/DriverEnums.h>
#include "utils/compiler.h"
#include <utils/compiler.h>
namespace filament::backend {

View File

@@ -17,15 +17,13 @@
#ifndef TNT_FILAMENT_BACKEND_PRIVATE_SAMPLERGROUP_H
#define TNT_FILAMENT_BACKEND_PRIVATE_SAMPLERGROUP_H
#include "backend/DriverApiForward.h"
#include <utils/compiler.h>
#include <utils/FixedCapacityVector.h>
#include <backend/DriverApiForward.h>
#include <backend/DriverEnums.h>
#include <backend/Handle.h>
#include <backend/SamplerDescriptor.h>
#include <utils/FixedCapacityVector.h>
#include <utils/ostream.h>
#include <stddef.h>
namespace filament::backend {

Some files were not shown because too many files have changed in this diff Show More