[arrayfire] annotated tag upstream/3.3.2+dfsg1 created (now a143e5e)

[Ghislain Vaillant]

This is an automated email from the git hooks/post-receive script.

ghisvail-guest pushed a change to annotated tag upstream/3.3.2+dfsg1
in repository arrayfire.

        at  a143e5e   (tag)
   tagging  b6183497598566cf51faf3be5f643fd7a38f6338 (commit)
  replaces  upstream/3.3.1+dfsg1
 tagged by  Ghislain Antony Vaillant
        on  Tue Apr 26 18:33:42 2016 +0100

- Log -----------------------------------------------------------------
Upstream version 3.3.2+dfsg1
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
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=dpsc
-----END PGP SIGNATURE-----

Ghislain Antony Vaillant (2):
      Prevent inclusion of <sys/sysctl.h> on gnu hurd.
      Merge tag 'v3.3.2' into upstream/latest

Pavan Yalamanchili (17):
      BUGFIX: Handle errors being returned by the C binary functions
      Handle errors for CUDA and OpenCL specific functions
      BUGFIX: Do not take control of external OpenCL device and contexts
      Incrementing the version to 3.3.2
      Merge pull request #1368 from ghisvail/fix/gnu-hurd-support
      Merge pull request #1367 from shehzan10/cmake-fixes
      BUILD: fix to disable using CPUID on demand
      Merge pull request #1380 from shehzan10/cpu-async-gcc
      Merge branch 'fixes_32' of pavanky/arrayfire into hotfix-3.3.1
      TEST: Adding rgb <--> gray conversion tests
      BUGFIX: Fixing rgb to gray conversion
      BUILD: Add options to build deb and rpm packages on demand
      DOCS: Updating documentation for replace.
      Merge pull request #1373 from shehzan10/sort
      Merge pull request #1389 from pavanky/rpm_deb_builds
      Converting std::pair to std::tuple to get for sorting in CPU backend
      Use memAlloc and memFree instead of std::vector internally

Shehzan Mohammed (44):
      Merge pull request #1354 from pavanky/fixes-3.3.2
      More detailed message about CUDA computes in CMake
      Disable JPEG_GREYSCALE flag in imageio if not defined in freeimage header file
      If CUDA_CUDA_LIBRARY is not found, check for stub and throw error
      Updated boost.compute to latest develop commit
      Use stable sort for thrust and boost compute
      Allow complex types as value type in sort_by_key
      Improvements to sort
      Revert "Updated boost.compute to latest develop commit"
      Boost.Compute sort/sort_by_key are stable in v0.5. So revert to that
      Call modDims when setDataDims is called
      Fix temp Array T and moddims in CPU batched sort
      Call thrust/compute::sort_by_key instead of kernel::sort_by_key wrapper
      PERF Add batched sort to sort_index for CPU and CUDA
      Disable CPU Async if GCC Version is less than 4.8.4
      Fixes for compiling CPU/OpenCL on Windows with static MKL and freeimage
      Add NOMINMAX for windows for tests when build out of source
      Check for MKL options after find blas
      Fix ordering of data from sort
      Merge pull request #1385 from shehzan10/hotfix-3.3.2
      Merge pull request #1387 from pavanky/rgb2gray_fix
      Fix reordering of data for sort_index in cpu and cuda
      Added sort_by_key batching to CPU and CUDA
      Added tests for sort_index and sort_by_key for higher dimensions
      Add sort index, sort by key batching to OpenCL
      Combine sort_index and sort_by_key kernels in CPU
      Combine sort_index and sort_by_key kernels in CUDA
      Combine sort_index and sort_by_key kernels in OpenCL
      Fix sort calls from harris and orb in CUDA
      Fix sort calls from harris and orb in OpenCL
      Clean up sort tests
      Fixed sort_by_key kernel for OpenCL
      Instantiate sort_by_key kernels in separately
      Instantiate sort_by_key kernels in separately in opencl
      Instantiate sort_by_key kernels in separately in cpu
      Sort by key cuda - create pair memory using memalloc, reasonable heuristic
      Remove sort 0 dim restriction note from documentation
      Add multi dimension support to median, tests
      Fixes to Using on OSX page
      OSX Installer: better error handling from brew failures
      Updated release notes for v3.3.2
      Merge pull request #1391 from pavanky/compile_fix
      Merge pull request #1393 from shehzan10/hotfix-3.3.2
      Merge pull request #1394 from arrayfire/hotfix-3.3.2

-----------------------------------------------------------------------

No new revisions were added by this update.

-- 
Alioth's /usr/local/bin/git-commit-notice on /srv/git.debian.org/git/debian-science/packages/arrayfire.git