Seungho Baek [Wed, 16 Apr 2025 05:33:54 +0000 (14:33 +0900)]
[Tizen] Fix spec to use pkgconfig for libtbb
Change-Id: Iee8d31900c54df076a96cad6897725e8475c0596
Signed-off-by: Seungho Baek <sbsh.baek@samsung.com>
Adeel Kazmi [Thu, 19 Sep 2024 15:45:02 +0000 (16:45 +0100)]
Tizen Support
Adeel Kazmi [Thu, 19 Sep 2024 15:42:11 +0000 (16:42 +0100)]
Support for stricter 32-bit ARM architecture
Adeel Kazmi [Thu, 19 Sep 2024 15:23:19 +0000 (16:23 +0100)]
32 Bit Support
Adeel Kazmi [Wed, 18 Sep 2024 15:54:27 +0000 (16:54 +0100)]
Only build the libraries required for DALi
Change-Id: If4b827b70173031a31f2502ed4f52e1387fc2252
pixar-oss [Fri, 23 Feb 2024 23:50:21 +0000 (15:50 -0800)]
Merge release v24.03
sunyab [Thu, 22 Feb 2024 21:44:31 +0000 (13:44 -0800)]
Update CHANGELOG with additional changes for 24.03 release
(Internal change:
2317614)
pixar-oss [Thu, 22 Feb 2024 21:04:12 +0000 (13:04 -0800)]
Adding UTF-8 guide docs to doxygen and Sphinx docs
(Internal change:
2317598)
pixar-oss [Wed, 21 Feb 2024 17:22:33 +0000 (09:22 -0800)]
Merge pull request #2919 from dgovil/boost_python_clang15_fixes
Disable BOOST_UNORDERED_HAVE_PIECEWISE_CONSTRUCT for MacOS builds
(Internal change:
2317210)
pixar-oss [Fri, 16 Feb 2024 17:02:37 +0000 (09:02 -0800)]
Updates for 24.03 release notes
(Internal change:
2316703)
pixar-oss [Wed, 14 Feb 2024 18:18:45 +0000 (10:18 -0800)]
Adding Terms of Use link to openusd.org page footer
(Internal change:
2315572)
pixar-oss [Thu, 8 Feb 2024 18:22:18 +0000 (10:22 -0800)]
Fixing Javascript issue for sphinx doc pages
(Internal change:
2315166)
blevin [Tue, 6 Feb 2024 19:50:44 +0000 (11:50 -0800)]
Fix hdPrman builds for RenderMan 24.0
pull over line from
2312852 to 24.0
(Internal change:
2315002)
sunyab [Mon, 5 Feb 2024 19:40:05 +0000 (11:40 -0800)]
Add support for Python 3.11 to usd-core PyPI package
The manylinux docker image used for Linux builds was bumped
to the earliest revision that includes Python 3.11.
(Internal change:
2314835)
sunyab [Mon, 5 Feb 2024 18:09:16 +0000 (10:09 -0800)]
Update USD version for pending 24.03 release
(Internal change:
2314801)
sunyab [Sat, 3 Feb 2024 04:04:57 +0000 (20:04 -0800)]
pcp: Revert specializes culling optimization
Change
2312006 implemented prim index culling for specializes
nodes. This noticeably reduced memory usage on several test
production shots, but increased memory usage in at least one
other shot. More crucially, it caused runtime performance
regressions in authoring workflows that added new opinions
to prims in sites that were now being culled due to this
optimization. In this case, Pcp needs to fully recompute the
entire prim index to pick up the new opinions.
We lean towards better performance over memory usage, so
I'm reverting this change for now until we can figure out
how to address the authoring issue.
(Internal change:
2314652)
pixar-oss [Sat, 3 Feb 2024 04:04:54 +0000 (20:04 -0800)]
Fixing doxygen markdown+HREF formatting issue
(Internal change:
2314643)
pixar-oss [Sat, 3 Feb 2024 04:04:46 +0000 (20:04 -0800)]
Updating doc stage.Traverse() comments regarding PrimRange
Fixes #222
(Internal change:
2314591)
unhyperbolic [Sat, 3 Feb 2024 04:04:30 +0000 (20:04 -0800)]
UsdGeomTetMesh faces computation: changing the map so that the key does not carry information that should be part of the value.
(Internal change:
2314459)
blevin [Sat, 3 Feb 2024 04:04:24 +0000 (20:04 -0800)]
LPE AOV's now use the RenderVar name rather than the LPE
expression as the channel name. This both gives expected
outputs in EXR channels and fixes name collisions when the
same LPE is used with multiple summary statistics, such
as variance and mse.
Strip ri:displayChannel namespace from params dictionary
keys, so e.g. ri:displayChannel:statistics -> statistics,
as expected by RenderMan API.
(Internal change:
2314430)
pellerington [Sat, 3 Feb 2024 04:04:17 +0000 (20:04 -0800)]
Add PxrMesh adapter to help mesh lights work on older versions of USD
(Internal change:
2314386)
blevin [Sat, 3 Feb 2024 04:04:03 +0000 (20:04 -0800)]
Fix crashes when UsdRenderProduct's orderedVars points at
a target prim that does not exist.
(Internal change:
2314341)
tallytalwar [Sat, 3 Feb 2024 04:04:00 +0000 (20:04 -0800)]
Small typo fix in value clips docs
(Internal change:
2314333)
blevin [Sat, 3 Feb 2024 04:03:53 +0000 (20:03 -0800)]
Fix up comment & diagnostic message for actual env var name
(Internal change:
2314302)
davidgyu [Sat, 3 Feb 2024 04:03:48 +0000 (20:03 -0800)]
[hgiMetal] Runtime test for fragment discard return
A previous fix made sure to return from the fragment shader main
function after executing discard_fragment() in order to avoid
undesired side effects from buffer writes, e.g. when collecting
translucent fragments for Order Independent Translucent (OIT)
rendering.
This introduces a runtime check to disable this early return on
systems where it can cause other problems with performance or
functionality.
This follows the pattern used directly in the lines above
to detect and implement a similar sort of workaround for
AMD Vega GPU devices. Ideally, we'd check for something
like MTLGPUFamilyApple9 feature support or inspect the device
architecture name, but we also want to maintain build
compatibility with earlier versions of Xcode and the macOS SDK.
(Internal change:
2314233)
unhyperbolic [Sat, 3 Feb 2024 04:03:10 +0000 (20:03 -0800)]
Add note about mismatch with Riley API
(Internal change:
2314232)
clach [Sat, 3 Feb 2024 04:03:04 +0000 (20:03 -0800)]
[hgiVulkan] Add implementation for HgiVulkanBlitCmds::FillBuffer.
(Internal change:
2314208)
klucknav [Sat, 3 Feb 2024 04:02:54 +0000 (20:02 -0800)]
Add imaging support for UsdGeomTetMesh
(Internal change:
2314180)
(Internal change:
2314671)
jloy [Sat, 3 Feb 2024 04:02:46 +0000 (20:02 -0800)]
hdar: fix symbol visibility typo
(Internal change:
2314172)
diyajoy [Sat, 3 Feb 2024 04:02:23 +0000 (20:02 -0800)]
Update UsdGeomSubset::GetUnassignedIndices to work correctly for
elementType edge.
(Internal change:
2314156)
pixar-oss [Sat, 3 Feb 2024 04:01:34 +0000 (20:01 -0800)]
Merge pull request #2905 from marcodiiga/dev_findentry_nullptr_check
Added nullptr check for _FindEntry (like _InsertEntry and related)
(Internal change:
2314131)
blevin [Sat, 3 Feb 2024 04:00:39 +0000 (20:00 -0800)]
HdPrman: Add TfToken support for Riley params, among other
types, by re-using type case dispatch provided by
SetParamFromVtValue
(Internal change:
2314105)
sunyab [Sat, 3 Feb 2024 04:00:32 +0000 (20:00 -0800)]
pcp: Cache prim index traversals for variant selections
Pcp now caches information computed used when searching
for variant selections during prim indexing. This includes
the translated path at each node where we look for
authored variant selections, as well as whether that
node actually contained any authored variant selections.
This greatly improves performance when dealing with prim
indexes that require recomputing variant selections many
times. For example, in a synthetic test case involving
a prim with 1000 references, adding 500 variant sets to
that prim previously took ~25.5 seconds; with this change,
it takes ~6.8 seconds. In a production test case with
many inherit and variant arcs, the time spent computing
variants dropped from ~30 seconds to ~4.5 seconds.
The caching mechanism involves two new objects:
- PcpNodeRef_PrivateSubtreeConstIterator, an iterator
for traversing over a specified subtree within a
prim index.
- Pcp_TraversalCache, an object that uses the above
iterator to traverse a subtree and caches information
at each node in the traversal.
Together, these objects allow us to transform the search
for variant selections from a recursive traversal to a
simple iteration, avoid expensive path translations, and
skip over nodes where we've already determined no
authored variant selections exist.
(Internal change:
2314082)
(Internal change:
2314085)
(Internal change:
2314122)
sunyab [Sat, 3 Feb 2024 04:00:23 +0000 (20:00 -0800)]
pcp: Ensure implied opinions are considered for ancestral variants
This change fixes a long-standing composition bug where
variant selections authored in implied inherit locations
would be ignored during recursive prim indexing for
ancestral opinions. This was because ancestral variants
were evaluated immediately, before implied inherit arcs
that might have contributed opinions were added to the graph.
Pcp now defers the evaluation of ancestral variants as late
as possible to ensure that all possible sources of opinions are
available when searching for variant selections. Conceptually,
this is the same thing Pcp does for "direct" variant selections.
However, this is trickier to implement because it requires Pcp
to evaluate variants on a node at a different level of namespace
than the node's current site.
This approach means that variants will never be evaluated in
recursive prim indexing calls. This allows us to remove some
complicated code that was needed to deal with traversing a
prim index across those recursive calls and replace it with a
straightforward strength-order traversal of a single prim index.
The TrickySpookyVariantSelectionInClass museum case contains
examples of where this occurs. This test began passing several
years ago, leading us to believe the bug had been fixed by some
other change. It turns out that we were just getting lucky
(or unlucky) and the test was set up in a way that just
happened to get the right answer, but for the wrong reason.
This test had been disabled in a previous change, but with
this fix we can reenable it.
(Internal change:
2314080)
unhyperbolic [Sat, 3 Feb 2024 04:00:09 +0000 (20:00 -0800)]
UsdImagingStageSceneIndex::SetTime: adding forceDirtyingTimeDeps flag
(Internal change:
2314005)
Pixneb [Sat, 3 Feb 2024 03:59:14 +0000 (19:59 -0800)]
Ts: fix buggy derivative cases, and improve test coverage.
Derivatives are supposed to reflect what happens in value evaluation. There
were several cases where derivatives disagreed with the curve shape given by
value evaluation.
Most of these cases involved linear and/or held knots.
Some example cases - not an exhaustive list:
- Derviatives at all linear knots were incorrectly evaluated from one of the
spline's extrapolation slopes.
- Ineffective tangent values on linear and held knots were incorrectly being
considered at neighboring knots.
- Ineffective extrapolation-facing tangent values on Bezier knots were
incorrectly being returned verbatim, even when held extrapolation means that
these slopes are zero. This was the only case involving Bezier knots.
This change completely rewrites Ts_Eval, and partially rewrites some of its
helpers. Some test baselines that codified the incorrect behavior have been
updated.
Add testTsDerivatives. I believe this covers all possible derivative cases.
Derivatives are well-defined for value types that are, in the language of
TsTraits, extrapolatable. This is a bit of a misnomer. It is true that
"extrapolatable" means "we will perform extrapolation". But it also means "we
can determine a slope".
(Internal change:
2313969)
pixar-oss [Sat, 3 Feb 2024 03:59:08 +0000 (19:59 -0800)]
Merge pull request #2918 from pmolodo/pr/testTsSimplify-fix
ts: fix for TsSimplify + testTsSimplify
(Internal change:
2313961)
pixar-oss [Sat, 3 Feb 2024 03:58:54 +0000 (19:58 -0800)]
Merge pull request #2873 from dgovil/build_usd/ignore_homebrew
Add the ability to request CMake to ignore paths
(Internal change:
2313862)
diyajoy [Sat, 3 Feb 2024 03:58:47 +0000 (19:58 -0800)]
Add support for elementTypes "tetrahedron" and "face" to
UsdGeomSubset for use with UsdGeomTetMesh.
(Internal change:
2313837)
jloy [Sat, 3 Feb 2024 03:58:42 +0000 (19:58 -0800)]
vt: allow NRVO in static VtValue::Cast
The in-place VtValue::Cast method returns a reference to *this. In
return position, this requires the compiler to copy construct a
temporary from the reference.
(Internal change:
2313813)
blevin [Sat, 3 Feb 2024 03:56:18 +0000 (19:56 -0800)]
usdrecord now accepts a -sesssionLayer argument.
Remove the recently added --vars argument from usdrecord,
since it is currently unclear whether we need it, and
--sessionLayer provides a more general superset of
capabilities.
(Internal change:
2313796)
rajabala [Sat, 3 Feb 2024 03:55:31 +0000 (19:55 -0800)]
[hd] Improve write-up.
(Internal change:
2313776)
rajabala [Sat, 3 Feb 2024 03:54:43 +0000 (19:54 -0800)]
[hdPrman] Use shutter curve on the camera when available.
- Update HdPrmanCamera to query and cache shutter curve attributes.
- Rework HdPrman_CameraContext to use the camera's shutter curve when available. It no longer needs to cache any of this state nor expose API to set it.
- Add unit test for various shutter interval and curve scenarios. See movingSphere.usda for notes on having motion blur show up correctly when using usdview.
- Update existing motion blur tests to use the fallback shutter curve for batch rendering so the baselines can remain unchanged.
Note that usdImaging wasn't updated to accomplish this. The reason this works is because the 1.0 and 2.0 getters (UsdImagingDelegate::GetCameraParamValue and UsdImagingDataSourceCameraPrim) fallback to querying the attribute on the USD prim. The shutter curve attributes are auto-applied by the UsdRi PxrCameraAPI schema on any UsdGeomCamera prim and thus always available when using hdPrman with USD.
It would however be useful to add an UsdImagingAPISchemaAdapter for PxrCameraAPI to allow its attributes to be inspectable in the Hydra Scene Browser.
(Internal change:
2313615)
matthewcpp [Sat, 3 Feb 2024 03:54:29 +0000 (19:54 -0800)]
The change handles changes that were the result of the asset localization that relate to processing multiple references to the same dependencies.
A cache has been added to asset localization delegates in order to prevent the user processing function from being triggered multiple times for the same layer / asset path combination.
UsdUtilsComputeAllDependencies has been modified to ensure that the same resolved apths do no appear multiple times in the same output array.
(Internal change:
2313564)
pixar-oss [Sat, 3 Feb 2024 03:53:59 +0000 (19:53 -0800)]
Merge pull request #2859 from nvmkuruc/arutf
Specify invalid UTF-8 characters in URI scheme registration errors
(Internal change:
2313563)
pixar-oss [Sat, 3 Feb 2024 03:39:34 +0000 (19:39 -0800)]
Merge pull request #2852 from erikaharrison-adsk/feature-hgi-vulkan-general-stability
Autodesk: General HgiVulkan stability
(Internal change:
2313559)
o2b3d [Sat, 3 Feb 2024 03:25:59 +0000 (19:25 -0800)]
Convert latin-1 encoded files to utf-8
(Internal change:
2313544)
klucknav [Sat, 3 Feb 2024 03:25:46 +0000 (19:25 -0800)]
[hdPrman] Use the recently added 'disableDepthOfField' RenderSettings option in HdPrman.
(Internal change:
2313517)
tallytalwar [Sat, 3 Feb 2024 03:25:14 +0000 (19:25 -0800)]
Update sdr to use SdfValueTypeNames::Int# values
- Shaders using sdr which have support for int2, int3, int4 can now
leverage SdfValueTypeNames->Int2, SdfValueTypeNames->Int3,
SdfValueTypeNames->Int4 appropriately, without having to specify
sdrUsdDefinitionType.
- Updates to following tests:
- testUsdShadeShaderDef
- testOslParser
(Internal change:
2313496)
pixar-oss [Sat, 3 Feb 2024 03:25:05 +0000 (19:25 -0800)]
Allow playback from SdfChangeList.
Currently SdfChangeList entries are ambiguous when it comes to
reproducing the edits they indicate. For example, an entry can
have both didAddNonInertPrim and didRemoveNonInertPrim but it
can't tell you if it was an add followed by a remove or a remove
followed by an add or even add/remove/add, etc.
With this change entries that would be ambiguous when combined are
kept separate. There is no change to the API.
(Internal change:
2313470)
sunyab [Sat, 3 Feb 2024 03:24:57 +0000 (19:24 -0800)]
pcp: Add/update museum tests for upcoming ancestral variant fix
The new test cases capture the current behavior in a variety of
corner-ish cases that caused internal tests to fail during
development of the fix mentioned above. Checking these in
first helps verify that these behaviors do not change after
that fix lands.
(Internal change:
2313409)
rajabala [Sat, 3 Feb 2024 03:24:54 +0000 (19:24 -0800)]
[hdsi] Fix typo.
(Internal change:
2313366)
blevin [Sat, 3 Feb 2024 03:24:50 +0000 (19:24 -0800)]
usd: Python wrap for Usd.UsdFileFormat.Tokens
(Internal change:
2313325)
pixar-oss [Sat, 3 Feb 2024 03:24:38 +0000 (19:24 -0800)]
Merge pull request #2906 from nvmkuruc/utf8iteratortype
Change value type of `TfUtf8CodePointIterator` to `TfUtf8CodePoint`
(Internal change:
2313272)
blevin [Sat, 3 Feb 2024 03:24:31 +0000 (19:24 -0800)]
Add a --vars argument to usdrecord, allowing expression variables
to be specified on the command line. They will be set on the session
layer before recording.
(Internal change:
2313259)
pixar-oss [Sat, 3 Feb 2024 03:23:46 +0000 (19:23 -0800)]
Adding support for orientationsf in Prman
See #2529
(Internal change:
2313218)
Dhruv Govil [Thu, 25 Jan 2024 20:51:48 +0000 (12:51 -0800)]
Disable BOOST_UNORDERED_HAVE_PIECEWISE_CONSTRUCT for MacOS builds
Paul Molodowitch [Thu, 25 Jan 2024 06:30:21 +0000 (22:30 -0800)]
ts: fix for TsSimplify + testTsSimplify
On windows debug builds, was getting a "Debug Assertion Failed" on
simplify.cpp, line 379. This is because this we were using iterators
that have become invalid because the underlying data was altered.
Matt Kuruc [Mon, 11 Dec 2023 05:05:01 +0000 (21:05 -0800)]
Specify invalid UTF-8 characters in URI scheme registration errors
pixar-oss [Mon, 22 Jan 2024 23:05:11 +0000 (15:05 -0800)]
Merge pull request #2907 from erikaharrison-adsk/adsk/bugfix/HeaderIncludeError
Autodesk: Include <optional> to fix external build error
(Internal change:
2313062)
diyajoy [Mon, 22 Jan 2024 22:37:27 +0000 (14:37 -0800)]
Update DoubleConversion version to v3.3.0 in tf.
Fixes #2218
(Internal change:
2313015)
cvj [Mon, 22 Jan 2024 22:37:17 +0000 (14:37 -0800)]
Make HdImageShaderSchema constants return a schema instead of a container data source.
(Internal change:
2312854)
klucknav [Mon, 22 Jan 2024 22:37:08 +0000 (14:37 -0800)]
Add disableDepthOfField bool to RenderSettings
This is adding a 'disableDepthOfField' bool to the UsdRenderSettingsBase
Schema, as well as the HdRenderProductSchema. A follow up change will
add the ability for HdPrman to use this option.
(Internal change:
2312852)
gitamohr [Mon, 22 Jan 2024 22:37:03 +0000 (14:37 -0800)]
tf: Fix buggy bugs not properly decrementing python exception state reference
counts, and not properly saving/restoring exception state in some corner cases.
(Internal change:
2312838)
rajabala [Mon, 22 Jan 2024 22:36:58 +0000 (14:36 -0800)]
[usd] Fix comment typo.
(Internal change:
2312773)
cvj [Mon, 22 Jan 2024 22:36:54 +0000 (14:36 -0800)]
Make HdSceneIndexAdapterSceneDelegate handle image shaders correctly.
(Internal change:
2312761)
cvj [Mon, 22 Jan 2024 22:36:48 +0000 (14:36 -0800)]
ImageShader filePath is a string, not an asset path.
(Internal change:
2312644)
unhyperbolic [Mon, 22 Jan 2024 22:36:43 +0000 (14:36 -0800)]
Hydra schemas: making material-related schema Getters return schemas instead of just a container data source.
(Internal change:
2312440)
pixar-oss [Mon, 22 Jan 2024 22:36:40 +0000 (14:36 -0800)]
Merge pull request #2889 from nvmkuruc/untangleutf8iterator
Untangle `TfUtf8CodePointIterator` from view class
(Internal change:
2312405)
rajabala [Mon, 22 Jan 2024 22:36:35 +0000 (14:36 -0800)]
[hd, hdui] Add a button on the Hydra Scene Browser to dump scene index contents to file.
(Internal change:
2312284)
matthewcpp [Mon, 22 Jan 2024 22:36:30 +0000 (14:36 -0800)]
This change adds an optional UsdUtilsProcessingFunc parameter to UsdUtilsComputeAllDependencies. This allows for users to inject custom logic into asset discovery. It is important to note that in this case, modifications to asset paths and dependencies are not wrriten back to the underlying layers. The motivation for this change was to allow for filtering of certain layers in order to match specific behavior of previous versions of this function.
(Internal change:
2312250)
jmk [Mon, 22 Jan 2024 22:36:26 +0000 (14:36 -0800)]
Adjust formatting of SdfPathExpression documentation to annotate
syntactic elements as inline code.
Also corrected the formatting of an example expression which resulted
in odd output in the generated HTML. This required the manual use of
<code> tags to get around some unexpected Doxygen behavior.
(Internal change:
2312036)
sunyab [Mon, 22 Jan 2024 22:36:20 +0000 (14:36 -0800)]
pcp: Implement culling for specializes nodes
Change
2311004 caused a regression in memory usage because
nodes due to ancestral opinions beneath specializes nodes
were not being culled as they had been before.
For example, if prim /A specialized /B/C/D, Pcp would
recursively compute the prim index for /B/C/D. Prior to the
above change, Pcp would cull nodes in that prim index before
attaching it to the prim index for /A. After that change,
Pcp would attach the unculled prim index first, then attempt
to cull nodes later. However, this caused Pcp to hit a
clause in the culling code that explicitly ignored specializes
nodes, meaning the nodes for /B/C/D would be skipped over
entirely.
Culling for specializes nodes was originally unimplemented
because of the complicated duplication of node subtrees that
was part of how the specializes arc was implemented. Adding
this fixes the regression and actually reduces memory usage
even further than the original numbers.
(Internal change:
2312006)
pixar-oss [Mon, 22 Jan 2024 22:36:13 +0000 (14:36 -0800)]
doc that previewtex uses transparent black border color
Fixes #2770
(Internal change:
2311916)
whar [Mon, 22 Jan 2024 22:36:01 +0000 (14:36 -0800)]
Adding asset[] support to usdImaging Hydra type dispatcher
(Internal change:
2311901)
pixar-oss [Mon, 22 Jan 2024 22:32:08 +0000 (14:32 -0800)]
PcpMuseum test output/baseline cleanup: removing all remaining "_pcp"
suffixes from output file and baselines. This required renaming existing
baseline files using the "_pcp" suffix to text files without the suffix.
In some cases, the suffixless baseline didn't exist already so a simple
perforce rename was all that was required.
In other cases the suffixless file did exist but was not being used
anymore. These files were updated to match the "_pcp" baseline (if they
didn't already) and the "_pcp" baseline deleted.
(Internal change:
2311696)
pixar-oss [Mon, 22 Jan 2024 22:31:23 +0000 (14:31 -0800)]
Merge pull request #2851 from erikaharrison-adsk/feature-hgi-vulkan-unit-test
Autodesk: Feature hgi vulkan unit test
(Internal change:
2311664)
(Internal change:
2311695)
Matt Kuruc [Fri, 29 Dec 2023 23:13:40 +0000 (15:13 -0800)]
Change value type of `TfUtf8CodePointIterator` to `TfUtf8CodePoint`
Matt Kuruc [Tue, 26 Dec 2023 16:00:48 +0000 (08:00 -0800)]
Untangle `TfUtf8CodePointIterator` from view class
blevin [Fri, 12 Jan 2024 18:41:20 +0000 (10:41 -0800)]
Register UsdImaging data source support for SdfPathExpression.
(Internal change:
2311586)
gitamohr [Fri, 12 Jan 2024 18:41:15 +0000 (10:41 -0800)]
sdf: A bugfix in the SdfPathExpression parser: only allow the pattern chars
'-', '!', and ']' once an initial '[' has been seen. Previously a case like
/foo//-/foo/bar// which should parse as a set-difference ('/foo// -
/foo/bar//') would parse as a single path pattern with a '-' element. Also
beef up some python wrapping and add a more exhuastive Python API test.
(Internal change:
2311465)
pixar-oss [Fri, 12 Jan 2024 18:40:49 +0000 (10:40 -0800)]
Updating the testPcpMuseum tests so that the clean up step, where
absolute paths to the test run directory are made relative in the test
output, happens directly in the testPcpCompositionResults script.
This requires baseline updates to a few tests as the the original
cleanup step was sometimes deleting more output text than
necessary.
(Internal change:
2311459)
pixar-oss [Fri, 12 Jan 2024 18:40:45 +0000 (10:40 -0800)]
Merge pull request #2877 from dgovil/FixUsdShadeOutputValidation
UsdOutput::IsOutput should check attr validity
(Internal change:
2311435)
comand [Fri, 12 Jan 2024 18:40:38 +0000 (10:40 -0800)]
Remove uses of the MENV30 define.
(Internal change:
2311372)
pixar-oss [Fri, 12 Jan 2024 18:40:24 +0000 (10:40 -0800)]
Merge pull request #2885 from sirpalee/pr/compliance-checker-producing-attr
Using get value producing attrs when checking texture assets in NormalMapTextureChecker
(Internal change:
2311331)
pixar-oss [Fri, 12 Jan 2024 18:40:19 +0000 (10:40 -0800)]
Merge pull request #2888 from nvmkuruc/updatedocs
Update documentation for UTF-8 utilities
(Internal change:
2311183)
pixar-oss [Fri, 12 Jan 2024 18:40:15 +0000 (10:40 -0800)]
Merge pull request #2890 from tylerm-nv/usdviewq-utf8-search
UTF8 search support in usdviewq
(Internal change:
2311157)
ucbearkat [Fri, 12 Jan 2024 18:38:40 +0000 (10:38 -0800)]
Refactoring HdSt_CodeGen to take metaData as constructor arg.
This allows us to delay the construction of the codeGen object since the hash can be computed
from the constructor args. Prior to this change the codeGen object was created on every
invocation of HdStExtCompGpuComputationResource::_Resolve just so the metadata member
can be retrieved and filled in by HdSt_ResourceBinder.
Although HdSt_CodeGen::CompileComputeProgram was being called only for the first instance of
each program merely constructing the codeGen object on every invocation was doubling the time
taken in the _Resolve function when the use case involved 1000s of extComputations, as in
an example production crowds case.
With this change we can delay the construction of the codeGen object and make sure it is created
only for the first instance of each compute program. For the example case this change provides
approx 50% speedup in HdStExtCompGpuComputation::Execute and consequently >10% speedup
in frame rendering time.
(Internal change:
2311005)
sunyab [Fri, 12 Jan 2024 18:38:35 +0000 (10:38 -0800)]
pcp: Defer node culling passes to the end of indexing
When recursively computing prim indexes to incorporate
ancestral opinions due to composition arcs, Pcp would scan
and mark cullable nodes in those indexes before adding them
to the "final" requested prim index. These culling passes
assumed that all ancestral composition arcs had already been
processed.
An upcoming fix will defer the handling of ancestral variant
arcs, breaking this assumption. In anticipation of this, this
change defers the culling pass until the final prim index has
been fully computed. This has the side benefit of removing
some semi-complicated logic that would "uncull" nodes in
certain cases.
This change causes the TrickySpookyVariantSelectionInClass
test case to fail, but the behavior that test was intended to
verify was already broken to begin with; the test just happened
to be set up in such a way to mask that failure. That behavior
will be fixed by the upcoming change, so we just disable the
test for now.
Besides that, I don't expect any practically noticeable changes
in behavior or performance. Pcp did not rely on node culling
for any optimizations during indexing, and we should wind up
with the same nodes being culled and removed after indexing
is completed. Landing this change first will help verify that
and allow us to track unexpected regressions if they occur.
(Internal change:
2311004)
pixar-oss [Fri, 12 Jan 2024 18:38:30 +0000 (10:38 -0800)]
UsdNamespaceEditor now finds all relationship targets and attribute
connections for properties on the stage the stage that refer to the
paths changed by a namespace edit and attempts to modify them to
refer to the updated path (or remove the target altogether in the case
of namespace deletion).
The test here currently only handles the basic case where all target and
connection paths exist in specs in the root layer stack. The
UsdNamespaceEditor does handle targets and connections that are defined
across composition arcs, but the test cases for this will be added in
an upcoming change.
(Internal change:
2310948)
pixar-oss [Fri, 12 Jan 2024 18:38:25 +0000 (10:38 -0800)]
README fix -- avoid recommending /opt/local for install
(Internal change:
2310922)
clach [Fri, 12 Jan 2024 18:37:54 +0000 (10:37 -0800)]
[usdImagingGL] Add test setup for testUsdImagingGLDomeLightTaskController to
the open source build.
(Internal change:
2310873)
Pixneb [Fri, 12 Jan 2024 18:36:54 +0000 (10:36 -0800)]
Initial version of Ts library.
This library implements animation splines.
It builds, but isn't used by anything yet. It's a first milestone
that will allow us to begin co-developing with external partners and collecting
public feedback.
Much work remains. See the README for a rough list of tasks.
(Internal change:
2310847)
(Internal change:
2310875)
(Internal change:
2310876)
(Internal change:
2311059)
(Internal change:
2311623)
rajabala [Fri, 12 Jan 2024 18:34:29 +0000 (10:34 -0800)]
[hd, usdImaging] Rework material binding schemata and resolution.
hd:
- Remove 'bindingStrength' from HdMaterialSchema. This is an authoring affordance that doesn't need to be transported to a render delegate.
- Update HD_API_VERSION.
usdImaging:
- Add schemata to represent direct and collection-based material bindings.
- Update materialBindingAPIAdapter to build and invalidate data sources based on the schemata.
- Move direct binding flattening from hd to usdImaging.
- Add a scene index to compute the resolved material bindings. Currently, it just wraps the flattened direct binding for a given purpose; this will be updated to factor in collection material bindings in a future change.
(Internal change:
2310706)
vverovvero [Fri, 12 Jan 2024 18:34:24 +0000 (10:34 -0800)]
Add documentation to HdContainerDataSourceEditor::Finish()
(Internal change:
2310692)
diyajoy [Fri, 12 Jan 2024 18:34:19 +0000 (10:34 -0800)]
Add support for elementType "edge" to UsdGeomSubset.
(Internal change:
2310632)
gitamohr [Fri, 12 Jan 2024 18:34:12 +0000 (10:34 -0800)]
tf: Wrap lines to fit in 80 cols following change
2310294.
(Internal change:
2310626)
Marco Alesiani [Thu, 11 Jan 2024 20:48:32 +0000 (21:48 +0100)]
Added nullptr check for _FindEntry (like _InsertEntry and related)
kapoorv [Wed, 10 Jan 2024 16:08:30 +0000 (11:08 -0500)]
cleaned up code around
1. VkPipelineVertexInputStateCreateInfo (following latest fixes proposed by Pixar)
2. Instance Layer Extensions (following latest fixes proposed by Pixar)
3. UV flipping fix in shader selection for Vulkan (following latest fixes proposed by Adobe)
4. Undoing temp code aound flipping color correction UVs flip
(cherry picked from commit
db0a46bede5b9504bb6954dc8ac7b1b5d6e5f364)
Pal Mezei [Wed, 10 Jan 2024 00:33:38 +0000 (11:33 +1100)]
Using UsdShade.Input.IsInput to verify that the value producing attr is an input value.