This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "MPICH primary repository". The annotated tag, v3.2b2 has been created at ced902bb0853c2b89c4c4e5e2d92d49ec0064bfb (tag) tagging c895bd9cc8bdb1701ac149d91742727ecd3cbca7 (commit) replaces v3.2b1 tagged by Huiwei Lu on Fri Apr 24 10:18:22 2015 -0500 - Log ----------------------------------------------------------------- tagging 'v3.2b2' Antonio J. Pena (5): Fix unsafe datatype release in several RMA tests Fix win_shared_create* tests Revert "Fix unsafe datatype release in several RMA tests" Moved a request assert to an earlier location Disable MPICH2 tests by default in old nightlies Antonio Pena Monferrer (1): Fix several RMA + async failures Charles J Archer (8): Update polling loop for OFI cq_read EAGAIN return OFI semantics: close AV after closing endpoint Simplification of OFI initialization flow using allocinfo OFI tsearch removed from API in favor of FI_PEEK+trecvmsg OFI Netmod: Add CVAR enhancements for OFI provider selection OFI: Update for removed FI_CANCEL flag OFI: Fix multiple providers failing probe-unexp Add ordering check to mprobe test case Halim Amer (3): Applied the PPoPP patch Revert "Applied the PPoPP patch" Hot fix for commit abb56764 Huiwei Lu (3): Fixes build failure with per-object thread-cs Changed MPICH version to 3.2b2 Modify CHANGES for 3.2b2 release Junchao Zhang (6): Change file extension from .F90 to .f90 in mpi_f08 Delete module files generated in F08 config test Revise F08 binding build system Define MPI_MESSAGE_NULL directly with an integer Enhance F08 buildiface to auto-set compile time constants Enhance F08 buildiface to auto-set int kinds of c_Ain/Count/Offset Ken Raffenetti (16): move error checking inside guards in info_get.c track OFI library (libfabric) in EXTERNAL_LIBS gforker: implement PMI_Abort handler gforker: fixup for [7bb8bd59] testsuite: Makefile cleanup portals4: fix incorrect usage of max_eqs limit portals4: remove unused header file portals4: add missing newline to error message portals4: fixup for [7d1fd292] portals4: increase max_entries limit portals4: improvements to VC close portals4: overhaul MPICH control message handling portals4: revert [722d85a4] and [d459c025] portals4: tuning portals4: fix anysource_matched mxm: fix anysource_matched Min Si (3): Fixed the Fortran common symbol issue on Mac. A workaround for FreeBSD pthread mallc/free bug. Revert "A workaround for FreeBSD pthread mallc/free bug." Norio Yamaguchi (11): Delete unused code in netmod-IB Code cleanup in netmod-IB Fix fail to build with new libc Fix allocation size and release size in netmod-IB Fix memory leak in netmod-IB Fix strict-aliasing rules break in netmod-IB Fix sending size of non-contiguous data in IB Fix invalid access before connection established Fix some build warnings in netmod-IB Fix operation of long messages in netmod-IB Add copyright in netmod-IB Paul Coffman (7): ROMIO One-sided Collective IO Aggregation base code ROMIO Collective IO One-sided aggregation algorithm base code ROMIO Collective IO One-sided aggregation algorithm GPFS enablement ROMIO Collective IO One-sided aggregation algorithm non-contiguous ROMIO Collective IO One-sided aggregation algorithm avoid ROMIO Collective IO One-sided aggregation algorithm support read-modify-write PAMI_Rput_typed / PAMI_Rget_typed utilization for derived types Pavan Balaji (14): Reduce the number of processes in the commcreatep test. Bump up the bcast2/bcast3 timeouts a bit more. Remove unnecessary backslashes in man pages. Get rid of bcast3.c Cosmetic changes to the bcast2 test. Simplify the bcast test. Update .gitignore. Free communicators when not using them. Increase bcast_full time limit. Run putfence1 and getfence1 only with two processes. Replace xfail ticket number. Cleanup threaded progress. Initial version of the intelligent thread yielding. Fix arbitrary poll count before yielding. Rob Latham (7): update ROMIO header checks for modern autoconf use autoconf routines for AIO struct checks check for aio-lite request/presence romio standalone: implement missing check fix "is present" case for MPI_Count check Revert "When decrementing a ref-count to 0, do not bother with atomics." move window creation out of common code Sameh Sharkawi (4): PAMID: CUDA AWARE support in collectives didn't handle local copy PAMID: Dynamically loading libcuda instead of static linking to libmpi.so PAMID: Fix CUDA-aware NULL buffer assert PAMID: MPI_Allreduce/MPI_Reduce coredump w/ DOUBLE_INT datatype Sangmin Seo (1): Fix wrong alias names. Su Huang (1): PAMID: Unexpected mpi_gather message print out when MP_PRINTENV=yes Wesley Bland (4): Remove unused version of MPID_Comm_agree Remove MPID_Comm_failed_bitarray Remove fixed size variable usage Extend timeouts for putfence1/getfence1 Xin Zhao (5): Move 'stream_offset' out of RMA packet struct. Set size of IMMED data in RMA packets to 8 bytes. Delete assert that no longer makes sense. Bug-fix: correct the wrong judgement in RMA function. Modify comments about atomicity in GACC/FOP handlers. ----------------------------------------------------------------------- hooks/post-receive -- MPICH primary repository