Changelog for hdf5-openmpi4-1.10.7-bp153.3.61.x86_64.rpm :

* Fri Jan 29 2021 Egbert Eich - Fix update_so_version.sh, write so versions to file being included.
* Fri Nov 06 2020 Ana Guerrero Lopez - Update to version 1.10.7
* Add metadata cache optimization to reduce skip list usage.
* Add BEST_EFFORT value to HDF5_USE_FILE_LOCKING environment variable.
* Add H5Pset/get_file_locking() API calls.
* Add Mirror VFD. Use TCP/IP sockets to perform write-only (W/O) file I/O on a remote machine. Must be used in conjunction with the Splitter VFD.
* Add Splitter VFD. Maintain separate R/W and W/O channels for \"concurrent\" file writes to two files using a single HDF5 file handle.
* Fixed an assertion failure in the parallel library when collectively filling chunks.
* Fortran/C++ libs: Add wrappers for H5Pset/get_file_locking() API calls. h5pget_file_locking_f() h5pset_file_locking_f()
* Added new Fortran parameters: H5F_LIBVER_ERROR_F H5F_LIBVER_NBOUNDS_F H5F_LIBVER_V18_F H5F_LIBVER_V110_F
* Add new Fortran API: h5pget_libver_bounds_f.
* h5repack added options to control how external links are handled.
* The tools library was updated by standardizing the error stack process.
* The H5DSis_scale function was updated to return \"not a dimension scale\" (0) instead of failing (-1), when CLASS or DIMENSION_SCALE attributes are not written according to Dimension Scales Specification.
* Bug Fixes: + Fix bug and simplify collective metadata write operation when some ranks have no entries to contribute. This fixes parallel regression test failures with IBM SpectrumScale MPI on the Summit system at ORNL. + Fixed use-of-uninitialized-value error. Appropriate initialization of local structs was added to remove the use-of-uninitialized-value errors reported by MemorySanitizer. + Creation of dataset with optional filter. A fix is applied to allow the creation of a dataset in such situation, as specified in the user documentation. + Explicitly declared dlopen to use RTLD_LOCAL. + H5Sset_extent_none() sets the dataspace class to H5S_NO_CLASS which causes asserts/errors when passed to other dataspace API calls. + Fixed the segmentation fault when reading attributes with multiple threads The problem was fixed by setting the file pointer to the current opened file pointer when the attribute was accessed. Similar patch up was done before when reading dataset with variable length string datatype. + Don\'t allocate an empty (0-dimensioned) chunked dataset\'s chunk index, until the dataset\'s dimensions are increased. + Fortan: Corrected INTERFACE INTENT(IN) to INTENT(OUT) for buf_size in h5fget_file_image_f. + h5diff fixed a command line parsing error. + h5diff added a command line option to ignore attributes. + h5diff added another level to the verbose argument to print filenames. + h5repack was fixed to repack the reference attributes properly.
* For details check: https://support.hdfgroup.org/ftp/HDF5/releases/hdf5-1.10/hdf5-1.10.7/src/hdf5-1.10.7-RELEASE.txt- version 1.10.6
* Added S3 and HDFS Virtual File Drivers (VFDs) to HDF5. Instructions to enable them when configuring HDF5 on Linux and Mac may be found at. https://portal.hdfgroup.org/display/HDF5/Virtual+File+Drivers+-+S3+and+HDFS.
* Added new wrappers for H5Pset/get_create_intermediate_group() LinkCreatPropList::setCreateIntermediateGroup() LinkCreatPropList::getCreateIntermediateGroup()
* h5repack was fixed to repack datasets with external storage to other types of storage.
* Bug Fixes: + Improved performance when creating a large number of small datasets by retrieving default property values from the API context instead of doing skip list searches. More work is required to achieve parity with HDF5 1.8. + Fixed user-created data access properties not existing in the property list returned by H5Dget_access_plist. Thanks to Steven Varga for submitting a reproducer and a patch. + Inappropriate linking with deprecated MPI C++ libraries. HDF5 does not define
*_SKIP_MPICXX in the public headers, so applications can inadvertently wind up linking to the deprecated MPI C++ wrappers. MPICH_SKIP_MPICXX and OMPI_SKIP_MPICXX have both been defined in H5public.h so this should no longer be an issue. HDF5 makes no use of the deprecated MPI C++ wrappers. + fcntl(2)-based file locking incorrectly passed the lock argument struct instead of a pointer to the struct, causing errors on systems where flock(2) is not available. This bug affects HDF5 1.10.0 through 1.10.5. fcntl(2)-based file locking now correctly passes the struct pointer. + Fixed a bug caused by a bad tag value when condensing object header messages. + Fixed an issue when creating a file with non-default file space info together with library high bound setting to H5F_LIBVER_V18. This was fixed by setting and checking the proper version in the file space info message based on the library low and high bounds when creating and opening the HDF5 file. + Fixed an issue where copying a version 1.8 dataset between files using H5Ocopy fails due to an incompatible fill version. + Fixed a bug that would cause an error or cause fill values to be incorrectly read from a chunked dataset using the \"single chunk\" index if + Fixed a bug that could cause an error or cause fill values to be incorrectly read from a dataset that was written to using H5Dwrite_chunk if the dataset was not closed after writing. + Fixed memory leak in scale offset filter.
* For details check: https://support.hdfgroup.org/ftp/HDF5/releases/hdf5-1.10/hdf5-1.10.6/src/hdf5-1.10.6-RELEASE.txt- Security bugs fixed:
* CVE-2018-13870: heap-based buffer over-read in the function H5O_link_decode in H5Olink.c (bsc#1101493)
* CVE-2018-13869: memcpy parameter overlap in the function H5O_link_decode in H5Olink.c (bsc#1101495)
* CVE-2018-17438: A SIGFPE signal is raised in the function H5D__select_io() of H5Dselect.c in the HDF HDF5 through 1.10.3 library during an attempted parse of a crafted HDF file, because of incorrect protection against division (bsc#1109570)
* CVE-2018-17435: A heap-based buffer over-read in H5O_attr_decode() in H5Oattr.c in the HDF HDF5 through 1.10.3 library allows attackers to cause a denial of service via a crafted HDF5 file. (bsc#1109567)- Refresh patches
* Thu Jul 23 2020 Egbert Eich - Add build support for gcc10 to HPC build (bsc#1174439).
* Mon Jul 20 2020 Egbert Eich - Add missing openmpi4 flavors.- Enable build of non-HPC flavors for all non-HPC builds.
* Wed Jun 24 2020 Alin M Elena - add openmpi4 packages to build and disable build for leap
* Fri Apr 17 2020 Egbert Eich - Fix .so number in baselibs.conf for libhdf5_fortran libs (boo#1169793).
* Thu Apr 02 2020 Egbert Eich - Fix library link flags on pkg-config file for HPC builds (boo#1134298).
* Mon Jan 20 2020 Stefan Brüns - Remove bogus undefines of suffix and mpi_flavor. suffix is the name of an RPM built-in (like expand or echo), and since RPM 4.15 trying to alter a built-in is an error. Fixes built of serial flavor on Tumbleweed. See https://rpm.org/user_doc/macros.html
* Tue Dec 10 2019 Stefan Brüns - Add compat provides for renamed hdf5-openmpi{,-devel} package, see https://en.opensuse.org/openSUSE:Package_dependencies- Fix openmpi1 package naming for SLE/Leap <= 15.1
* Thu Nov 28 2019 Egbert Eich - Add HPC support for gcc8 and gcc9 (jsc#SLE-7766 & jsc#SLE-8604).
* Thu Nov 28 2019 Egbert Eich - Disable openmpi1 builds for SLE/Leap > 15.1.- Enable openmpi3 builds for Leap and SLE > 15.1 (jsc#SLE-7773).
* Thu Oct 24 2019 Nicolas Morey-Chaisemartin - openmpi has been renamed into openmpi1
* Tue Sep 10 2019 Ana Guerrero Lopez - Use -ffat-lto-objects to provide a proper static library.
* Tue Sep 03 2019 Ana Guerrero Lopez - Re-list hdf5-mpi.patch in the spec file for the sake of a avoiding a script rejecting the package automatically.
* Fri Aug 23 2019 Ana Guerrero Lopez - Update to version 1.10.5: (jsc#SLE-8501)
* https://support.hdfgroup.org/ftp/HDF5/releases/hdf5-1.10/hdf5-1.10.5/src/hdf5-1.10.5-RELEASE.txt- Security bugs fixed:
* CVE-2018-17233: A SIGFPE signal is raised in the function H5D__create_chunk_file_map_hyper. (bsc#1109166)
* CVE-2018-17434: Memory leak in the H5O__chunk_deserialize() function in H5Ocache.c (bsc#1109167)
* CVE-2018-17437: A SIGFPE signal is raised in the function H5D__chunk_set_info_real. (bsc#1109168)- Bump fortran library soname, sonum_F from 100 to 102.- Adjust library installation path, use %hpc_prefix/lib64 in x86_64 and %hpc_libdir in all other cases- Patches refreshed:
* hdf5-LD_LIBRARY_PATH.patch
* hdf5-1.8.11-abort_unknown_host_config.patch
* PPC64LE-Fix-long-double-handling.patch
* hdf5-Remove-timestamps-from-binaries.patch
* Disable-phdf5-tests.patch- Patch disabled, upstream have changed the file that was being patched, while it could be ported, it\'s unknown if this patch is still needed.
* hdf5-mpi.patch- Patch removed, merged upstream
* 0001-Fix-return-value-in-test_libver_bounds_open.patch
* Thu Jul 04 2019 Egbert Eich - Since build machine size issues do not seem to go way, make the %check stage optional: enable with \'--with check\'.
* Thu Apr 25 2019 Christian Goll - set higher constraints for succesfull mpich tests (boo#133222)
* Mon Apr 01 2019 normandAATTlinux.vnet.ibm.com- Add _constraints to avoid transient build error ppc64le
* Sat Mar 09 2019 Stefan Brüns - Update URL: and Source: to use https://.
* Fri Mar 01 2019 Egbert Eich - Like on PPC ignore transient \'make check\' failures on aarch64.
* Fri Mar 01 2019 Egbert Eich - Remove ringdisabled magic as hdf5 is no longer a ring package. This should resolve a number of problems in the review channels for Factory with packages requiring hdf5 libs.
* Wed Feb 13 2019 Egbert Eich - Consolidate use of openmpi1, openmpi2, openmpi3 for non-HPC builds.
* Fri Jan 25 2019 eichAATTsuse.com- Remove dependency to the hdf5-devel-data file for HPC: Currently, there are no macro files for HPC - these files will have to be handled differently.- Re-disable all non-HPC builds for SLE.
* Mon Jan 14 2019 eichAATTsuse.com- Reenable \'serial\' build for SLE: this build creates hdf5-devel-data.
* Thu Jan 03 2019 eichAATTsuse.com- Do not build non-HPC on SLE: This helps to save some build cycles.
* Wed Dec 19 2018 Todd R - Use upstream-recommended %{_rpmconfigdir}/macros.d directory for the rpm macros.
* Wed Dec 12 2018 Todd R - Also build for regular openmpi2 and openmpi3
* Thu Nov 22 2018 Stefan Brüns - Build HDF5 C++ bindings also for the MPI variants, there is no reason not to do it.- Enable sz2 support also for MPI variants.
* Mon Nov 19 2018 eichAATTsuse.com- Replace git-archive generated by downloaded tar-ball.
* Sat Nov 17 2018 eichAATTsuse.com- Update to 1.10.4: https://support.hdfgroup.org/ftp/HDF5/releases/hdf5-1.10/hdf5-1.10.4/src/hdf5-1.10.4-RELEASE.txt
* Update Disable-phdf5-tests.patch
* Replace and update hdf5-buildcompare.patch by hdf5-Remove-timestamps-from-binaries.patch
* Fri Nov 16 2018 eichAATTsuse.com- Remove lib
*_cpp and lib
*_hl_cpp dependencies for MPI for the time being.
* Sun Nov 11 2018 eichAATTsuse.com- Create main master package, fix dependencies.
* Fri Sep 21 2018 Stefan Brüns - Update to 1.10.3: https://support.hdfgroup.org/ftp/HDF5/releases/hdf5-1.10/hdf5-1.10.3/src/hdf5-1.10.3-RELEASE.txt- Update to 1.10.2: https://support.hdfgroup.org/ftp/HDF5/releases/hdf5-1.10/hdf5-1.10.2/src/hdf5-1.10.2-RELEASE.txt- Bump soversions (101 -> 103) for libhdf5 and libhdf5_cpp- Drop Fix-warnings-for-missing-returns.patch, fixed upstream- Add 0001-Fix-return-value-in-test_libver_bounds_open.patch- Rebase patches (tab / space conversion):
* PPC64LE-Fix-long-double-handling.patch
* hdf5-buildcompare.patch
* hdf5-mpi.patch- Rebase Disable-phdf5-tests.patch- Security bugs fixed:
* CVE-2017-17505: NULL pointer dereference in the function H5O_pline_decode allowing for DoS via crafted file (bsc#1072087)
* CVE-2017-17506: Out of bounds read in the function H5Opline_pline_decode (bsc#1072090)
* CVE-2017-17508: divide-by-zero vulnerability in function H5T_set_loc (bsc#1072108)
* CVE-2017-17509: out of bounds write vulnerability in function H5G__ent_decode_vec (bsc#1072111)
* CVE-2018-11202: NULL pointer dereference in H5S_hyper_make_spans in H5Shyper.c allows a remote denial of service attack (bsc#1093641)
* CVE-2018-11203: division by zero in H5D__btree_decode_key in H5Dbtree.c It could allow a remote denial of service attack (bsc#1093649)
* CVE-2018-11204: A NULL pointer dereference was discovered in H5O__chunk_deserialize in H5Ocache.c (bsc#1093655)
* CVE-2018-11206: A out of bounds read was discovered in H5O_fill_new_decode and H5O_fill_old_decode in H5Ofill.c (bsc#1093657)
* CVE-2018-11207: division by zero was discovered in H5D__chunk_init in H5Dchunk.c (bsc#1093653)
* Tue Jun 19 2018 eichAATTsuse.com- Handle license file using %license where applicable.- Package example package also for HPC build and create a master file for it.
* Fri Apr 27 2018 eichAATTsuse.com- Fix HPC library master packages dependency: make it require the correct flavor (bsc#1091237).
* Sat Apr 07 2018 eichAATTsuse.com- Only build one examples package for all flavors, do not include dependencies as these would be flavor specific (bsc#1088547).
* Sat Mar 31 2018 stefan.bruensAATTrwth-aachen.de- Fix some typos, one trivial, two mispelled macros
* Sat Feb 10 2018 jengelhAATTinai.de- Trim description for size and drop most of the marketing blurb. Fix RPM groups.
* Sat Feb 10 2018 eichAATTsuse.com- Make module files package arch dependent: it contains arch-dependent paths (boo#1080442).
* Sat Feb 10 2018 eichAATTsuse.com- Prepend PKG_CONFIG_PATH in modules file (boo#1080426).
* Fri Feb 09 2018 eichAATTsuse.com- Add support for openmpi2 for HPC (FATE#325089).
* Fri Feb 09 2018 cgollAATTsuse.com- Fix summary in module files (bnc#1080259)
* Thu Feb 08 2018 eichAATTsuse.com- Disable %check stage for mpich builds on s390(x) (bsc#1080022).
* Fri Jan 12 2018 eichAATTsuse.com- Disable the openmpi3 flavor in some products.
* Fri Jan 05 2018 eichAATTsuse.com- Switch from gcc6 to gcc7 as additional compiler flavor for HPC on SLES.- Add support for mpich and openmpi3 for HPC.
* Thu Nov 09 2017 normandAATTlinux.vnet.ibm.com- temporarily disable make check for PowerPC boo#1058563
* Tue Nov 07 2017 eichAATTsuse.com- Update to 1.10.1: https://support.hdfgroup.org/ftp/HDF5/releases/hdf5-1.10/hdf5-1.10.1/src/hdf5-1.10.1-RELEASE.txt- Fix-warnings-for-missing-returns.patch: replace: hdf5-non_void_return.patch- PPC64LE-Fix-long-double-handling.patch: replace: hdf5-ldouble-ppc64le.patch- Disable-phdf5-tests.patch: replace: hdf5_disable_testphdf5.patch- Disabling validation tests for the MPI variant for now as they seem to cause issues in OBS.
* Thu Oct 26 2017 eichAATTsuse.com- Fix name: baselib.conf -> baselibs.conf
* Sun Oct 22 2017 eichAATTsuse.com- Remove openssl-devel and krb5-devel from the BuildRequires.
* Tue Oct 17 2017 eichAATTsuse.com- Add magic to limit the number of flavors built in the OBS ring to non-HPC and mvapich2 builds.- Fix %libname() for non-HPC builds.- Fix some summaries and a Requires: of devel-static.
* Fri Sep 08 2017 eichAATTsuse.com- Convert package to multibuild.- Add HPC build using environment modules (FATE#321717, FATE#321710).
* Tue Aug 29 2017 hpjAATTurpla.net- fix compiler flag substitution
* Mon Jul 03 2017 badshah400AATTgmail.com- Add libsz2-devel to list of Requires for -devel packages, this is required for compilation against hdf5 when szlib support is enabled.- Minor specfile cleanups using spec-cleaner.
* Fri Jun 16 2017 junghansAATTvotca.org- Enabled szlib support through libaec\'s szlib API
* Wed Oct 05 2016 toddrme2178AATTgmail.com- Fix baselibs.conf so numbers. Fixes bnc#1002569
* Tue Sep 06 2016 normandAATTlinux.vnet.ibm.com- new hdf5_disable_testphdf5.patch to disable only testphdf5 for all arches and not all mpi tests for ppc64le tempo patch as long as boo#997252 not solved
* Mon Sep 05 2016 normandAATTlinux.vnet.ibm.com- temporarily disable mpi test for ppc64le
* Tue Jul 26 2016 adrianAATTsuse.de- update to 1.10.0-patch1 https://www.hdfgroup.org/ftp/HDF5/releases/hdf5-1.10/hdf5-1.10.0-patch1/src/hdf5-1.10.0-patch1-RELEASE.txt
* Mon May 23 2016 toddrme2178AATTgmail.com- update to 1.8.17, changes documented here: https://www.hdfgroup.org/ftp/HDF5/releases/hdf5-1.8.17/src/hdf5-1.8.17-RELEASE.txt
* Fri Apr 15 2016 normandAATTlinux.vnet.ibm.com- add hdf5-mpi.patch same as fedora commit id 78ddcff82b10fd80eafb7b2dc5e4ce2bb9ecbe3c should avoid transient hung of t_pflush1 test
* Fri Mar 18 2016 toddrme2178AATTgmail.com- Split libraries into C, C++, and Fortran versions to comply with openSUSE library packaging policies. Fixes BNC#971602- Make Fortran builds required.- Manually set so numbers. The autodetected ones weren\'t matching the real ones.
* Mon Mar 07 2016 olafAATTaepfle.de- Remove timestamps from binaries hdf5-buildcompare.patch
* Thu Jan 28 2016 sweet_f_aAATTgmx.de- update to 1.8.16, changes documented here: https://www.hdfgroup.org/ftp/HDF5/releases/hdf5-1.8.16/src/hdf5-1.8.16-RELEASE.txt- rebase hdf5-ldouble-ppc64le.patch- remove hdf5-cpp-compilation-errors.patch- prepare independent so versions and add TODO
* Fri Aug 21 2015 toddrme2178AATTgmail.com- Currently the regular, openmpi, and mpivachi2 versions of the library packages all provide the exact same libraries. So any package built against the -devel package will end up with an \"unresolvable\" build error because there are multiple versions of the library it can use. To avoid this, I have added generic provides for the library package so packages can depend on the version they really need, avoiding these build issues.
* Thu Aug 13 2015 badshah400AATTgmail.com- Add hdf5-cpp-compilation-errors.patch: Reverts a change from v1.8.15 to H5api_adpt.h that caused compilation errors for C++ HDF5 applications; patch taken from upstream svn.
* Mon Jul 27 2015 toddrme2178AATTgmail.com- Update subpackage groups- Make subpackage summaries and descriptions more consistent- Provide rpm macros for version identification- Spec file cleanups
* Mon Jul 27 2015 toddrme2178AATTgmail.com- Use explicit library requirements to avoid ambiguities. Without this, packages that depend on hdf5 won\'t know which library to use.
* Thu Jul 23 2015 toddrme2178AATTgmail.com- Fix baselibs.conf
* Wed Jul 22 2015 toddrme2178AATTgmail.com- update to 1.8.15
* New Features + Library - Added Memory Allocation Functions that Use the Library\'s Allocator - H5Pset_istore_k and H5Pset_sym_k added validation of \"ik\" to not exceed the max v1 btree entries (2 bytes) to these two routines. - Added Functions to Control the Value of H5PL_no_plugin_g without Using an Environment Variable + Parallel Library - Attached an attribute destroy callback to MPI_COMM_SELF that shuts down the HDF5 library when MPI_COMM_SELF is destroyed, in other words, on MPI_Finalize. This should fix several issues that users see when they forget to close HDF5 objects before + Fortran API - Added Global Variables. These new global variables are equivalent to the C definitions without the \'_F\': o H5G_UDLINK_F o H5G_SAME_LOC_F o H5O_TYPE_UNKNOWN_F o H5O_TYPE_GROUP_F o H5O_TYPE_DATASET_F o H5O_NAMED_DATATYPE_F o H5O_TYPE_NTYPES_F + C++ API - New Wrappers for C Functions H5P[s/g]et_libver_bounds - New Wrappers to Get the Object Header\'s Version - New DataType Constructor + Support for New Platforms, Languages, and Compilers - Support for Linux 3.10.0-123.20.1.el7 added (LK - 2015/04/01)
* Bug Fixes since HDF5-1.8.14 + Library - Addition of Error Tracing Functionality to Several C API Calls - H5Rdereference Now Checks for HADDR_UNDEF or Uninitialized References - Fixed Bugs in H5Sextent_copy + Parallel Library - Fixed a Potential Memory Error - Parallel Test Problem Fixed - MPE Support + Tools - h5repack crashed on enumerated 8-bit type. - h52gif crashed non-8bit images. - perform/benchpar.c retired. - Source perform/ directory moved to tools/perform. + Fortran API - Fortran Fails with --enable-fortran2003 and Intel 15.x Compilers - h5tenum_insert_f Does Not Work with Default 8 Byte Integers (xlf compiler) - Some Fortran APIs Never Returned the Error State - Fixed h5pget_class_f + C++ API - Combined Two H5File::getObjCount Overloaded Methods The following two methods o ssize_t getObjCount(unsigned types) const; o ssize_t getObjCount() const; were combined into one: o ssize_t getObjCount(unsigned types = H5F_OBJ_ALL) const; - Many Warnings Were Removed - Functionality Deprecation The following two constructors of classes AbstractDs, IdComponent, H5Location, and H5Object are no longer appropriate after the data member \"id\" had been moved from IdComponent to the sub-classes in previous releases. o (const hid_t h5_id); o (const & original); The copy constructors were no-op and removed in 1.8.15. The other constructors will be removed from 1.10 release, and then from 1.8.17 if their removal does not cause any problems. + High-Level APIs: - Suppress Warnings from Flex/Bison-generated Code - Changed hdf5_hl.h to Include the HDF5 Main Library \"hdf5.h\" - H5PTcreate_fl Does Not Convert to Memory Datatype - Fix for H5LT Attribute Functions + Fortran High-Level APIs: - Internal Library Fix for Missing Argument Declaration- Build in thread-safe mode to allow parallel reads
* These are only unsupported to avoid confusing casual users. The fotran and C++ libraries don\'t support thread-safe mode, while the C library does. So you have to explicitly allow building the thread-safe C library and the non-thread-safe C++ and fortran libraries in order to make sure only people who have done their research do so. Since users are going to be accessing this through other programs, this doesn\'t matter.- Build fortran 2003 libraries- Rebase hdf5-non_void_return.patch- Remove irrelevant hdf5-configure-long-double.patch
* Mon May 11 2015 dmuellerAATTsuse.com- update hdf5-non_void_return.patch for GCC 5.x
* Tue Jan 13 2015 pthAATTsuse.de- Tag baselibs.conf as source.
* Fri Jan 09 2015 sweet_f_aAATTgmx.de- fix typos in ChangeLog
* Wed Nov 19 2014 pgajdosAATTsuse.com- updated to 1.8.14:
* bugfix release, see RELEASE.txt