OSL 1.11.14.0

Release 1.11.14 -- 10 May 2021 (compared to 1.11.13)

• oslc/OSLCompiler: Check for errors when when writing oso (for example,
  disk volume full). #1360
• Fix crash generating closure function calls with LLVM 11. #1361
• Build: LLVM 12 compatibility. #1351
• Build: Support for CMake 3.20 #1354
• Build: Be robust to certain OpenEXR 2.x config fies. #1356
• Build: Fixes to deal with newer OIIO builds that are C++14.

OSL 1.11.13.0

Release 1.11.13 -- 1 Apr 2021 (compared to 1.11.12)

• Updated vector2.h with a mod(vector2,vector2) function. #1312
• Fix unnecesary error messages to stderr when encountering broken point
  clouds. #1333

OSL 1.11.12.0

Release 1.11.12 -- 1 Mar 2021 (compared to 1.11.11)

• Build: Changes for smooth compile against upcoming Imath 3.0. #1332 #1341
• Build: We weren't properly hiding non-public symbols. #1337
• CI: Cleanup and get rid of the last Travis vestiges. #1334
• CI: Speed up CI runs by using GHA cache + ccache. #1335
• CI: Separate CI process into several individual "steps." #1338

OSL 1.11.11.0

Release 1.11.11 -- 1 Feb 2021 (compared to 1.11.10)

• Build: Make the CMake build system more friendly to being a subproject.
  #1304
• Build: Improve dependency on Imath/OpenEXR include files.
• CI: Fixes to Mac CI. #1315
• Docs: Update CLAs, Charter, GOVERNANCE, particularly after acceptance of
  new CLAs and moving the repo to the ASWF GitHub account. #1308 #1240 #1314
• Internals: Switch from deprecated call to OIIO::parallel_image to the new
  version. #1317

OSL 1.11.10.0

Release 1.11.10 -- 1 Dec 2020 (compared to 1.11.9)

• Bug fix: optimizer interaction with output param that also is userdata.
  #1295
• Ensure that LoadMemoryCompiledShader returns false upon parse failure.
  #1302
• Work to ensure that OIIO will build correctly against the upcoming
  Imath 3.0 and OpenEXR 3.0. #1299

OSL 1.11.9.0

Release 1.11.9 -- 1 Nov 2020 (compared to 1.11.8)

• Build properly against Cuda 11 and OptiX 7.1. #1232
• Windows compile fixes. #1263 #1285
• PugiXML build fixes on some systems. #1262
• Fix asymptomatic potential runtime optimizer bug where certain
  multi-component values were not correctly recognized as nonzero. #1266
• Fix false positive with "debug_uninit" where certain while loops might be
  flagged as using uninitialized values when in fact they were not. #1252
• Slightly optimize by not copying output params to layers that will not be
  used. #1253
• Internals: use the final keyword in certain internal classes where
  applicable. #1260
• Cuda/OptiX back end: Add __CUDADEVRT_INTERNAL__ define to bitcode
  generation, needed to avoid duplicate cudaMalloc symbols with CUDA9+ #1271
• Fix build break against recent OIIO master change where m_mutex field was
  removed from ImageInput. #1281

OSL 1.11.8.0

Release 1.11.8 -- 1 Oct 2020 (compared to 1.11.7)

• Fix broken derivatives of the optional "alpha" return of texture calls
  when the normal color channels return don't have their derivatives used
  but the alpha does. #1258
• Building vs Python: instead of defaulting to searching for python 2.7
  specifically and needing to set PYTHON_VERSION if you want (or have) a
  different one, default to whichever version is found. If multiple versions
  of Python are on the system, you can still use the PYTHON_VERSION cmake
  variable to disambiguate which one you want. #1249
• For the experimental OptiX support, explicitly set the OptiX pipeline
  stack size (fixes some bugs). #1254
• ShadingSystem statistics output (which is printed with testshade --help,
  by the way) now includes information about the SIMD capabilities specified
  at OSL build time, those available at runtime, and the versions of LLVM,
  OIIO, and Imath used. This should help in debugging and issue reporting
  by making it easy to know certain build time choices just by running
  testshade. #1258

OSL 1.11.7.3

Release 1.11 -- 1 Sept 2020 (compared to 1.10)

ASWF adoption changes:

• GOVERNANCE.md and ASWF/Technical-Charter.md document the project
  governance and how the Technical Steering Committee works. An archive of
  meeting minutes can be found in ASWF/meetings. #1137 (1.11.5)
• All the copyright notices now read "Copyright Contributors to the Open
  Shading Language Project," all files now have proper SPDX identifiers
  instead of long boilerplate notices, and 3rd party included code now has
  its licenses documented in the LICENSE-THIRD-PARTY.md file. #1145 (1.11.5)
• The official mail list has switched to https://lists.aswf.io/g/osl-dev
  #1167

Dependency and standards changes:

• LLVM 7.0-10.0: Support for LLVM 4, 5, and 6 have been dropped. Support
  for LLVM 8, 9, and 10 have been added. Note that using LLVM 10 requires
  building for C++14 or later, because LLVM 10's APIs no longer support
  C++11. #981 #1046 #1058 #1128 #1152 #1162 #1206
• OpenImageIO 2.0-2.2: Support for OIIO 1.8 has been dropped; a minimum of
  OIIO 2.0 is needed to build OSL. (Additionally, a minimum of OIIO 2.1 is
  strongly recommended for anyone using the SIMD batch shading mode.) #1038
  (1.11.0)
• CMake minimum is now 3.12. #1072 (1.11.1)

OSL Language and oslc compiler:

• New syntax to reference color and point/vector/normal components as
  named struct components, such as C.r, C.g, C.b, or P.x, P.y, P.z.
  #1049 (1.11.0)
• oslc compilation speed-ups with faster retrieval of source lines when
  pasted into oso output. #938 (1.11.0)
• Writing to function parameters not marked as output was only
  recently properly recognized as an error (it was documented as illegal
  all along). Now we demote to a warning, treating it as a full error was
  too disruptive. #944 (1.10.3/1.11.0)
• Improve oso output efficiency. #938 (1.10.3/1.11.0)
• Fix bug related to counting the proper number of values used for
  array default value initialization. #948 (1.10.3/1.11.0)
• Slight modification to the formatting of floating point values in the oso
  to ensure full precision preservation for float values. #949 (1.10.3/1.11.0)
• Fix crash when encountering empty-expression for (;;). #951 (1.10.3/1.11.0)
• Fix bug in code generation of certain while loops that directly
  tested a variable as a zero/nonzero test, such as:

      i = 3;
      while (i)
          --i;
  

  whereas the following had worked (they should have been identical):

      i = 3;
      while (i != 0)
          --i;
  

  #947 (1.10.3/1.11.0)
• Improve warnings about ill-advised use of the comma operator.
  #978 (1.10.4/1.11.0)
• Fix an assertion/crash when passing initialization-lists as parameters to
  a function, where the function argument expected was as array. #983
  (1.10.4/1.11.0)
• Fix an assertion/crash for certain type constructors of structs, where the
  struct name was not declared. (This is an incorrect shader, but of course
  should have issued an error, not crashed.) #988 (1.10.4/1.11.0)
• Improve oslc type error detection for variable declarations with init
  expressions. Note that this may flag some warnings and errors that went
  undetected before, involving initialization assignments of incompatible
  types. #991, #993 (1.10.5/1.11.0)
• Allow shader parameters with empty brace initializers, such as:

      shader foo ( float myarray[] = {} )
  

  Un-lengthed arrays, if they aren't connected or given a concrete instance
  value, still default to length 1 (we don't allow 0-length arrays, that's a
  whole other can of worms), but this is now treated as equivalent to
  = { default_value } (0 for int or float, "" for string). #967 (1.11.0)
• More correct handling of escaped string literals (such as "\tfoo\n")
  in string and also in string metadata. #1073 (1.11.1)
• New command line argument oslc --embed-source embeds the preprocessed
  source code in the .oso file itself (helpful for certain debugging
  situations). #1081 (1.11.3)
• oslc can generate dependency files, honoring the -M, -MM, -MD, -MMD, -MF,
  and -MT flags with similar meanings to gcc and clang. #1085 #1107 (1.11.3)
  Example:

      $ oslc -MMD test.osl
      Compiled test.osl -> test.oso
      $ cat test.d
      test.oso: test.osl \
          dist/macosx/share/OSL/shaders/color2.h \
          dist/macosx/share/OSL/shaders/stdosl.h
  

• New shader-level metadata: [[ int range_checking=0 ]] disables the
  automatic generation of range checking code for the shader. #1112 (1.11.4)
• Support has been added for pointcloud_get() to be able to retrieve
  per-point string data from point clouds. #1157 (1.11.5)
• Shader output parameters that are marked [[ int lockgeom=1 ]] will
  automatically bind to any identically-named userdata, just like non-output
  parameters have always done. #1200 (1.11.6)

OSL Standard library:

• Extend linearstep() and smooth_linearstep() to work with color, point,
  vector, and normal types (previously restricted to float). #994
  (1.10.5/1.11.0)
• Extend transformc() to understand translating between linear and sRGB
  color spaces. #1013 (1.11.0)
• cbrt() cube root was added to the standard library. #1164 #1166 (1.11.6)

API changes, new options, new ShadingSystem features (for renderer writers):

• ShadingSystem::convert_value() will now allow promotion of int or
  float to float[4]. #940 (1.11.0)
• Removed the varieties of RendererServices::texture(), texture3d(), and
  environment() that lack the errormessage parameter and had been marked
  as deprecated since OSL 1.8. #945 (1.11.0)
• Deprecate the auto-allocation of contexts and per-thread-info. Certain
  calls that took a Context* pointer that were allowed to be NULL now
  by convention require the renderer to pass a valid context. #958 (1.11.0)
• ShadingSystem: thread-safe/reentrant shader specification via new
  varieties of Parameter(), Shader(), ConnectShaders(), and
  ShaderGroupEnd() that take an explicit ShaderGroup reference and are
  therefore "stateless" and thread-safe. If you exclusively use these new
  methods for shader specification, it's possible for multiple threads to
  specify shader groups simultaneously. #984, #985, #986, #1000 (1.11.0)
  #1067 (1.11.1)
• New ShadingSystem options:
  • "opt_warnings" enables warnings about things that couldn't be
    optimized and may be performance issues. #1010 (1.11.0)
  • "gpu_opt_error" enables full error status of the subset of those
    warnings that are also hard no-go's when executing on GPUs. #1010
    (1.11.0)
  • "lazyerror", if set to 0, will unconditionally (and non-lazily) run
    any shader layers that still contain error() calls after
    optimization, so that error messages can't be elided by optimizations
    causing the layer to be eliminated or never run. When "opt_warnings"
    is enabled, warnings will be issued about shaders that couldn't
    optimize away all of their error calls. Note that the default value is
    1, giving the old behavior or lazily evaluating shader layers that
    contain error calls. #1191 (1.11.5)
• RendererServices::get_texture_handle() has changed slightly to require
  a ShadingConntext* parameter, and get_texture_info() has changed to
  have a parameter allowing the caller to provide a texture_thread_info
  handle and a context, as well as to provide a pointer to a ustring where
  error messages should be placed. #1033 (1.11.0)
• The LLVM JIT optimization level (controlled by the llvm_optimize shading
  system attribute) now defaults to 1, not 0. That should be the best
  performance while still having reasonable JIT times. #1114 (1.11.4)
• We no longer automatically build the MaterialX shaders (you need to set
  build option OSL_BUILD_MATERIALX to ON). This will eventually be deprecated
  entirely. #1136 (1.11.5)
• Helper varieties of ShadingSystem::Parameter() and ReParameter() that
  handle the common cases of a single float, int, or string. #1195 #1196
  (1.11.6)
• ShaingSystem::optimize_group() now takes an optional parameter that
  allows the caller to only do the optimize step without the JIT step.
  #1226 (1.11.71)

Continued work on experimental SIMD batched shading mode:

• Continued work on SIMD-friendly math and noise: #1108 (1.11.4)
• Support in LLVM_Util for various wide (SIMD) data types. #1175 (1.11.6)

Continued work on experimental OptiX rendering:

• Build option USE_OPTIX=1 enable experimental OptiX support.
• testshade and testrender now take --optix flags to run tests on OptiX.
  You can also force either to run in OptiX mode with environment variable
  TESTSHADE_OPTIX=1. #996 #1001 #1003
• Various fixes to how strings are represented on the Cuda side. #973 (1.11.0)
• More complete support of all the noise varieties. #980 (1.11.0)
• Got spline() function working. #974 #1011 (1.11.0)
• Work towards getting texture calls working. #974 (1.11.0)
• printf works for multiple values. #1007 (1.11.0)
• Work on color-related functions. #1012 (1.11.0)
• Support for native OSL closures. #1028 #1029 (1.11.0)
• Work on matrix ops. #1054 (1.11.1)
• Allow init ops. #1059 (1.11.1)
• Fixes to string closure params. #1061 (1.11.1)
• Add support for OptiX 7 #1111 #1203 (1.11.6)

Performance improvements:

• Constant fold array accesses even if they are out of bounds. #1035 (1.11.0)
• Speed up the concat() OSL function, the new implementation is very
  careful to avoid extra copies and unnecessary allocations. #1104,#1105
  (1.11.3)
• Shader performance improved slightly by better inlining during the JIT.
  #1114 (1.11.4)
• Range checking of matrix component access is skipped for certain cases
  where it was unnecessary. #1174 (1.11.6)
• LLVM JIT+optimize t...

OSL 1.11.7.2-RC2

Release candidate for 1.11. The branch is frozen for release. If no critical bugs are identified, we anticipate a full release on Sept 1, 2020.

OSL 1.11.7.1-RC1

Release candidate for 1.11 release. This branch is now frozen for all but critical bugs, with an expected 1.11 release on 1 Sept 2020.