HDF5 version 1.9.81 currently under development ================================================================================ INTRODUCTION This document describes the differences between HDF5-1.9.0 and HDF5 1.9.x snapshot, and contains information on the platforms tested and known problems in HDF5-1.9.x. For more details check the HISTORY*.txt files in the HDF5 source. Links to HDF5 1.9.x source code can be found on The HDF Group's development FTP server at the following location: ftp://ftp.hdfgroup.uiuc.edu/pub/outgoing/hdf5/snapshots User documentation for the snapshot can be accessed directly at this location: http://www.hdfgroup.uiuc.edu/HDF5/doc_dev_snapshot/H5_dev/ For more information, see the HDF5 home page: http://www.hdfgroup.org/HDF5/ If you have any questions or comments, please send them to the HDF Help Desk: help@hdfgroup.org CONTENTS - New Features - Support for new platforms and languages - Bug Fixes since HDF5-1.8.0 - Platforms Tested - Known Problems New Features ============ Configuration: ------------- - Configure now generates Makefiles that build in "silent make mode" by default in which compile and link lines are significantly simplified for clarity. To override this and view actual compile and link lines during building, the --disable-silent-rules flag can be used at configure time, or the 'make' command can be followed by V=1, to indicate a "verbose" make. (MAM - 2011/4/14). - Added mpicc and mpif90 as the default C and Fortran compilers for Linux systems when --enable-parallel is specified but no $CC or $FC is defined. (AKC - 2011/2/7) - Added a new configure option, "--enable-unsupported", which can be used to stop configure from preventing the use of unsupported configure option combinations, such as c++ in parallel or parallel HDF5 with threadsafe. Use at your own risk, as it may result in a library that won't compile or run as expected! (MAM - 2010/11/17 - Bug 2061) - PHDF5 changed to use "mpiexec", instead of mpirun, as the default MPI applications startup command as defined in the MPI-2 definition, section 4.1. (AKC - 2010/6/11 - Bug 1921) - Configure now adds appropriate defines for supporting large (64-bit) files on all systems, where supported, by default, instead of only linux. This largefile support is controllable with the --enable-largefile configure option. This is replacing the linux-specific --enable-linux-lfs option, which has been removed from configure. (MAM - 2010/05/05 - 1772/1434) - Upgraded versions of autotools used to generate configuration suite. We now use Automake 1.11.1, Autoconf 2.65, and Libtool 2.2.6b. MAM 2010/04/15. - Added the xlc-* and mpcc_r-* BASENAME patterns to be recognized as IBM compilers so that the ibm compiler options can be added properly. This allows non-system-default compiler command names (e.g. xlc-m.n.k.l) be recognized. AKC 2009/11/26. - Configuration suite now uses Automake 1.11 and Autoconf 2.64. MAM 2009/08/11. - Changed default Gnu fortran compiler from g95 to gfortran since gfortran is more likely installed with gcc now. -AKC 2009/07/19- - Added libtool version numbers to generated c++, fortran, and hl libraries. MAM 2009/04/19. - Regenerated Makefile.ins using Automake 1.10.2. MAM 2009/04/19. - Added a Make target of check-all-install to test the correctness of installing via the prefix= or $DESTDIR options. AKC - 2009/04/14 - Configuration suite now uses Libtool 2.2.6a. MAM 2008/10/24 - Configuration suite now uses Autoconf 2.61, Automake 1.10.1. MAM 2008/05/05. - The new configure option "--disable-sharedlib-rpath" disables embedding the '-Wl,-rpath' information into executables when shared libraries are produced, and instead solely relies on the information in LD_LIBRARY_PATH. (MAM - 2008/05/15) Library: -------- - Added ability to cache files opened through external links. Added new public functions H5Pset_elink_file_cache_size(), H5Pget_elink_file_cache_size(), and H5Fclear_elink_file_cache(). (NAF - 2011/02/17) - Removed all old code for Metraowerks compilers, bracketed by __MWERKS__). Metraowerks compiler is long gone. (AKC - 2010/11/17) - Added support for threadsafety on windows using the windows threads library. Use the HDF5_ENABLE_THREADSAFE option in CMake while on a windows platform to enable this functionality. This is supported on Windows Vista and newer Windows operating systems. (MAM - 2010/09/10) - When a mandatory filter failed to write data chunks, the dataset couldn't close (bug 1260). The fix releases all resources and closes the dataset but returns a failure. (SLU - 2010/9/8) - H5Tset_order and H5Tget_order now support all data types. A new byte order H5T_ORDER_MIXED has been added specifically for compound datatype and its derived type. Please see bug #1934. (SLU - 2010/8/23) - Improved performance of the chunk cache by avoiding unnecessary b-tree lookups of chunks already in cache. (NAF - 2010/06/15) - Greatly improved performance of extending a dataset with early allocation. (NAF - 2010/03/24 - 1637) - Added support for filtering densely stored groups. Many of the API functions related to filters have been extended to support dense groups as well as datasets. Pipeline messages can now be stored in a group's object header. (NAF/QAK - 2009/10/8) - The embedded library information is displayed by H5check_version() if a version mismatch is detected. Also changed H5check_version() to suppress the warning message totally if $HDF5_DISABLE_VERSION_CHECK is 2 or higher. (Old behavior treated 3 or higher the same as 1, that is print a warning and allows the program to continue. (AKC - 2009/9/28) - If a user does not care for the extra library information insert in the executables, he may turn it off by --disable-embedded-libinfo during configure. (AKC - 2009/9/15) - Corrected problem where library would re-write the superblock in a file opened for R/W access, even when no changes were made to the file. (QAK - 2009/08/20, Bz#1473) - Separated "factory" free list class from block free lists. These free lists are dynamically created and manage blocks of a fixed size. H5set_free_list_limits() will use the same settings specified for block free lists for factory free lists. (NAF - 2009/04/08) - Added support for dense attributes to H5Ocopy. (XCao/NAF - 2009/01/29) - Added H5Pset_elink_cb and H5Pget_elink_cb functions to support a user-defined callback function for external link traversal. (NAF - 2009/01/08) - Added H5Pset_elink_acc_flags and H5Pget_elink_acc_flags functions to allow the user to specify the file access flags used to open the target file of an external link. (NAF - 2009/01/08) - Added H5Pset_chunk_cache() and H5Pget_chunk_cache() functions to allow individual rdcc configuration for each dataset. Added H5Dget_access_plist() function to retrieve a dataset access property list from a dataset. (NAF - 2008/11/12) - Added H5Iis_valid() function to check if an id is valid without producing an error message. (NAF - 2008/11/5) - Added two new public routines: H5Pget_elink_fapl() and H5Pset_elink_fapl(). (see bug #1247) (VC - 2008/10/13) - Improved free space tracking in file to be faster. (QAK - 2008/10/06) - Added 'mounted' field to H5G_info_t struct. (QAK - 2008/07/15) Parallel Library: ----------------- - Modified parallel tests to run with arbitrary number of processes. The modified tests are testphdf5 (parallel dataset access), t_chunk_alloc (chunk allocation), and t_posix_compliant (posix compliance). The rest of the parallel tests already use in the code the number of processes available in the communicator. (CMC - 2009/04/28) Fortran Library: ---------------- - Added for the C APIs the Fortran wrappers: h5dget_access_plist_f h5iis_valid_f h5pset_chunk_cache_f h5pget_chunk_cache_f (MSB - 2009/04/17) C++ Library: ------------ - New member functions + Overloaded CommonFG::getObjnameByIdx to take char* for name + Overloaded CommonFG::getObjTypeByIdx to return type name as a char*. (BMR - 2010/05/02) + DataSet::getInMemDataSize() to simplify getting the dataset's data size in memory. (BMR - 2009/07/26) - These member functions were added as wrapper for H5Rdereference to replace the incorrect IdComponent::dereference(). void H5Object::dereference(H5File& h5file, void* ref) void H5Object::dereference(H5Object& obj, void* ref) In addition, these constructors were added to create the associated objects by way of dereference: Attribute(H5Object& obj, void* ref); Attribute(H5File& file, void* ref); DataSet(H5Object& obj, void* ref); DataSet(H5File& file, void* ref); DataType(H5Object& obj, void* ref); DataType(H5File& file, void* ref); Group(H5Object& obj, void* ref); Group(H5File& obj, void* ref); (BMR - 2008/08/10) Tools: ------ - h5diff: Added new level for -v (verbose) option. The new levels are 1 and 2. So -v1 and -v2 can be specified to view more information about attributes differences. Bug#2121 (JKM 2011/3/23) - h5dump: Added new option --enable-error-stack. This option will display error stack information in the output stream. This is useful when the "h5dump: Unable to print data" message is output. (ADB - 2011/02/24) - h5diff: Add a new flag --exclude-path. Specified path to an object will be excluded from comparing the two files or two groups. If group is specified all the member objects will be excluded. (JKM - 2010/09/16). - h5ls: Add new flag --no-dangling-links. (refer to --help for details) (JKM - 2010/06/15) - h5ls: Add new flag --follow-symlinks. (refer to --help for details) (JKM - 2010/05/25) - h5diff: Add new flag --no-dangling-links. (refer to --help for details) (JKM - 2010/02/10) - h5diff: Add new flag --follow-symlinks. (refer to --help for details) (JKM - 2010/01/25) - h5diff: fix for displaying garbage value on LE machine for BE data. (JKM - 2009/11/20) - h5dump: subsetting now allows default for count. Also trailing ; in short form can be omitted after last specified value. (ADB - 2009/09/04) - h5dump/h5ls: now can display data in region references using new -R, --region flag. (ADB - 2009/09/04) - h5diff: new flag, -c, --compare, list objects that are not comparable. (PVN - 2009/4/10 - 1368) - h5diff new flag, -N, --nan, avoids NaNs detection. (PVN - 2009/4/10) - h5dump correctly specifies XML dtd / schema urls (ADB - 2009/4/3 - 1519) - h5repack now handles group creation order. (PVN - 2009/4/2 - 1402) - h5dump: added a printing of the compression ratio of uncompressed and compressed sizes for cases where compression filters are present. (PVN - 2008/05/01) - h5dump: added an option to allow a user defined formatting string for printf regarding floating point numbers. (PVN - 2008/05/06) - h5dump: support for external links, display the object that the external link points to. (PVN - 2008/05/12) - h5repack: add a userblock to an HDF5 file during the repack. (PVN - 2008/08/26) - h5repack: add 2 options that call H5Pset_alignment in the repacked file. (PVN - 2008/08/29) - h5ls: added capability to traverse through external links when the -r (recursive) flag is given. (NAF - 2008/09/16) - h5ls: added -E option to enable traversal of external links. h5ls will not traverse external links without this flag being set. (NAF - 2008/10/06) - h5diff: added support for long double (PVN - 2008/10/28) - h5dump: binary output defaults to NATIVE with -b optionally accepting the form of binary output (NATIVE, FILE, BE, LE). (PVN - 2008/10/30) - h5diff: return 1 for file differences when both file graphs differ by any object. Error return code was changed to 2 from -1. (PVN - 2008/10/30) - h5import: TEXTFPE (scientific format) was deprecated. Use TEXTFP instead (PVN - 2008/10/30) - h5repack: When user doesn't specify a chunk size, h5repack now defines a default chunk size as the same size of the size of the hyperslab used to read the chunks. The size of the hyperslabs are defined as the size of each dimension or a predefined constant, whatever is smaller. This assures that the chunk read fits in the chunk cache. (PVN - 2008/11/21) - h5diff: h5diff treats two INFINITY values different. Fixed by checking (value==expect) before call ABS(...) at h5diff_array.c This will make that (INF==INF) is true (INF is treated as an number instead of NaN) (PC -- 2009/07/28) - h5diff: add option "--use-system-epsilon" to print difference if (|a-b| > EPSILON) Change default to use strict equality (PC -- 2009/09/12) High-Level APIs: ------ - Table: In version 3.0 of Table, "NROWS" (used to store number of records) was deprecated (PVN - 2008/11/24) Documentation ------------- Support for new platforms, languages and compilers. ======================================= - Intel V11.1 uses now -O3 optimization in production mode (EIP - 2010/10/08) - PathScale compilers are recognized and can build the HDF5 library properly. AKC - 2009/7/28 - Bug Fixes since HDF5-1.8.0 release ================================== Library ------- - The datatype handler created with H5Tencode/decode used to have the reference count 0 (zero). I have fixed it. It is 1 (one) now. (SLU - 2011/2/18) - Fixed a bug that caused big endian machines to generate corrupt files when using the scale-offset filter with floating point data or fill values. Note that such datasets will no longer be readable by any machine after this patch. (NAF - 2010/02/02 - Bug 2131) - Retrieving a link's name by index in the case where the link is external and the file that the link refers to doesn't exist will now fail gracefully rather than cause a segmentation fault. (MAM - 2010/11/17) - Modified library to always cache symbol table information. Libraries version 1.6.3 have a bug which causes them to require this information for some operations. (NAF - 2010/09/21 - 1864) - Fixed a bug that could occur when getting information for a new-style group that was previously opened through a file handle that was later closed. (NAF - 2010/09/15) - Added define check in H5public.h if stdint.h is supported by the C++ compiler. This define is only available on Windows with VS2010 and using CMake to build the library. (ADB - 2010/09/13 - Bug 1938) - H5Eset_current_stack now also closes the error stack to be set as the default. This is to avoid a potential problem (Bug 1799). (SLU - 2010/9/7) - Fixed the bug in the filter's public CAN_APPLY function. The return value should be htri_t not herr_t (Bug #1239). (SLU - 2010/8/5) - Fixed a bug in the direct I/O driver that could render files with certain kinds of unaligned data unreadable or corrupt them. (NAF - 2010/07/28) - valgrind reported an error of copying data to itself when a new attribute is written (Bug #1956). I fixed it by taking out the memcpy step in the attribute code. (SLU - 2010/07/28) - Fixed a bug that could cause file corruption when using non-default sizes of addresses and/or lengths. This bug could also cause uncorrupted files with this property to be unreadable. This bug was introduced in 1.8.5. (NAF - 2010/07/16 - 1951) - Fixed a file corruption bug that could happen when shrinking a compressed dataset. (NAF - 2010/05/20) - Fixed some memory leaks in VL datatype conversion when strings are used as fill values. (MAM - 2010/05/12 - BZ# 1826) - Fixed a bug when copying objects with NULL references with the H5O_COPY_EXPAND_REFERENCE_FLAG flag set. (NAF - 2010/04/08 - 1815) - Files can now be concurrently opened more than once using the core file driver, as long as the backing store is used. (NAF - 2010/03/09) - Added support for H5O_COPY_EXPAND_EXT_LINK_FLAG to H5Ocopy. External links will now be expanded if this flag is set. (NAF - 2010/03/05 - 1733) - Fixed a bug where the library, when traversing an external link, would reopen the source file if nothing else worked. (NAF - 2010/03/05) - Fixed an intermittent bug in the b-tree code which could be triggered by expanding and shrinking chunked datasets in certain ways. (NAF - 2010/02/16) - H5Tdetect_class said a VL string is a string type. But when it's in a compound type, it said it's a VL type (Bug #1584). I fixed it to be consistent. It always return string type. (SLU - 2009/12/10) - Fixed a bug where writing and deleting many global heap objects (i.e. variable length data) would render the file unreadable. Previously created files exhibiting this problem should now be readable. (NAF - 2009/10/27 - 1483) - Fixed incorrect return value for H5Pget_preserve. (AKC - 2009/10/08 - 1628) - Fixed an assertion failure that occurred when H5Ocopy was called on a dataset using a vlen inside a compound. (NAF - 2009/10/02 - 1597) - Fixed incorrect return value for H5Pget_filter_by_id1/2 in H5Ppublic.h. (NAF - 2009/09/25 - 1620) - Fixed a bug where properties weren't being compared with the registered compare callback. (NAF - 2009/09/25 - 1555) - Fixed a bug where H5Pget_fitler_by_id would succeed when called for a filter that wasn't present. (NAF - 2009/06/25 - 1250) - Fixed an issue with committed compound datatypes containing a vlen. Also fixed memory leaks involving committed datatypes. (NAF - 2009/06/10 - 1593) - Added versioning to H5Z_class_t struct to allow compatibility with 1.6 API. (NAF - 2009/04/20 - 1533) - Fixed a problem with using data transforms with non-native types in the file. (NAF - 2009/04/20 - 1548) - Added direct.h include file to windows section of H5private.h to fix _getcwd() warning. (ADB - 2009/04/14 - 1536) - Fixed a bug that prevented external links from working after calling H5close(). (NAF - 2009/04/10 - 1539) - Modified library to write cached symbol table information to the superblock, to allow library versions 1.3.0 to 1.6.3 to read files created by this version. (NAF - 2009/04/08 - 1423) - Changed skip lists to use a deterministic algorithm. The library should now never call rand() or srand(). (NAF - 2009/04/08 - 503) - Fixed a bug where H5Lcopy and H5Lmove wouldn't create intermediate groups when that property was set. (NAF - 2009/04/07 - 1526) - Fixed a bug that caused files with a user block to grow by the size of the user block every time they were opened. (NAF - 2009/03/26 - 1499) - Fixed a rare problem that could occur with files using the old (pre 1.4) array datatype. (NAF - 2009/03/23) - Modified library to be able to open files with corrupt root group symbol table messages, and correct these errors if they are found. Such files can only be successfully opened with write access. (NAF - 2009/03/23 - 1189) - Removed the long_long #define and replaced all instances with "long long". This caused problems with third party products. All currently supported compliers support the type. (ADB - 2009/03/05) - Fixed various bugs that could prevent the fill value from being written in certain rare cases. (NAF - 2009/02/26 - 1469) - Fixed a bug that prevented more than one dataset chunk from being cached at a time. (NAF - 2009/02/12 - 1015) - Fixed an assertion failure caused by opening an attribute multiple times through multiple file handles. (NAF - 2009/02/12 - 1420) - Fixed a problem that could prevent the user from adding attributes (or any object header message) in some circumstances. (NAF - 2009/02/12 - 1427) - Fixed a bug that could cause problems when an attribute was added to a committed datatype using the committed datatype's datatype. (NAF - 2009/02/12) - Fixed a bug that could cause problems when copying an object with a shared message in its own object header. (NAF - 2009/01/29) - Changed H5Tset_order to properly reject H5T_ORDER_NONE for most datatypes. (NAF - 2009/01/27 - 1443) - Fixed a bug where H5Tpack wouldn't remove trailing space from an otherwise packed compound type. (NAF - 2009/01/14) - Fixed up some old v2 btree assertions that get run in debug mode that were previously failing on compilation, and removed some of the more heavily outdated and non-rewritable ones. (MAM - 2008/12/15) - Fixed a bug that could cause problems when "automatically" unmounting multiple files. (NAF - 2008/11/17) - H5Ovisit and H5Ovisit_by_name will now properly terminate when the callback function returns a positive value on the starting object. (NAF - 2008/11/03) - Fixed an error where a null message could be created that was larger than could be written to the file. (NAF - 2008/10/23) - Corrected error with family/split/multi VFD not updating driver info when "latest" version of the file format used. (QAK - 2008/10/14) - Corrected alignment+threshold errors to work correctly when metadata aggregation is enabled. (QAK - 2008/10/06) - Changed H5Fget_obj_count and H5Fget_obj_ids to ignore objects registered by the library for internal library use. (NAF - 2008/10/06) - Fixed potential memory leak during compound conversion. (NAF - 2008/10/06) - Changed the return value of H5Fget_obj_count from INT to SSIZE_T. Also changed the return value of H5Fget_obj_ids from HERR_T to SSIZE_T and the type of the parameter MAX_OBJS from INT to SIZE_T. (SLU - 2008/09/26) - Fixed an issue that could cause data to be improperly overwritten during compound type conversion. (NAF - 2008/09/19) - Fixed pointer alignment violations that could occur during vlen conversion. (NAF - 2008/09/16) - Fixed problem where library could cause a segmentation fault when an invalid location ID was given to H5Giterate(). (QAK - 2008/08/19) - Fixed improper shutdown when objects have reference count > 1. The library now tracks reference count due to the application separately from that due to internal library routines. (NAF - 2008/08/19) - Fixed assertion failure caused by incorrect array datatype version. (NAF - 2008/08/08) - Fixed an issue where mount point traversal would fail when using multiple handles for the child. (NAF - 2008/08/07) - Fixed an issue where mount points were inaccessible when using multiple file handles for the parent. The mount table is now in the shared file structure (the parent pointer is still in the top structure). (NAF - 2008/08/07) - when an attribute was opened twice and data was written with one of the handles, the file didn't have the data. It happened because each handle had its own object structure, and the empty one overwrote the data with fill value. This is fixed by making some attribute information like the data be shared in the attribute structure. SLU - 2008/07/22 - Fixed issue where a group could have a file mounted on it twice. (QAK - 2008/07/15) - Fixed a Windows-specific issue in the ohdr test which was causing users in some timezones to get false errors. This a deficiency in the Windows mktime() function, and has been handled properly. SJW - 2008/06/19 - Fixed the problem with the searching of target file for H5Lcreate_external(). The searching pattern will depend on whether the target file's pathname is an absolute or a relative path. Please see the description in the RM for H5Lcreate_external(). (VC - 2008/04/08) - Fixed possible file corruption bug when encoding datatype descriptions for compound datatypes whose size was between 256 & 511 bytes and the file was opened with the "use the latest format" property enabled (with H5Pset_libver_bounds). (QAK - 2008/03/13) - Fixed bug in H5Aget_num_attrs() routine to handle invalid location ID correctly. (QAK - 2008/03/11) - H5Dset_extent: when shrinking dimensions, some chunks were not deleted. (PVN - 2009/01/8) - Added code to maintain a min_clean_fraction in the metadata cache when in serial mode. (MAM - 2009/01/9) Configuration ------------- - Removed config/ibm-aix6.x. All IBM-AIX settings are in one file, ibm-aix. (AKC - 2011/4/14) - Shared C libraries are no longer disabled on Mac when Fortran is enabled. Shared Fortran libraries are still not supported on Mac, so configure will disable them by default, but this is overridable with the new --enable-unsupported configure option. The configure summary has been updated to reflect the fact that the shared-ness of the C++/Fortran wrapper libraries may not align with the C library. (MAM - 04/11/2011 - HDFFV-4353). - Removed recognition of the parallel compilers of LAM(hcc) and ChMPIon(cmpicc) since we have no access to these two MPI implementations and cannot verify their correctness. (AKC - 2010/7/14 - Bug 1921) - Removed the following config files, as we no longer support them: config/dec-osf*, config/hpux11.00, config/irix5.x, config/powerpc-ibm-aix4.x config/rs6000-ibm-aix5.x config/unicos* MAM - 2009/10/08 - Modified configure and make process to properly preserve user's CFLAGS (and company) environment variables. Build will now properly use automake's AM_CFLAGS for any compiler flags set by the configure process. Configure will no longer modify CFLAGS directly, nor will setting CFLAGS during make completely replace what configure has set up. MAM - 2009/10/08 - Support for TFLOPS, config/intel-osf1, is removed since the TFLOPS machine has long retired. AKC - 2009/10/06. - Added $(EXEEXT) extension to H5detect when it's executed in the src/Makfile to generate H5Tinit.c so it works correctly on platforms that require the full extension when running executables. MAM - 2009/10/01 - BZ #1613 - Configure will now set FC and CXX to "no" when fortran and c++ are not being compiled, respectively, so configure will not run some of the compiler tests for these languages when they are not being used. MAM - 2009/10/01 - The PathScale compiler (v3.2) was mistaken as gcc v4.2.0 but it fails to recognize some gcc options. Fixed. (see bug 1301). AKC - 2009/7/28 - - The --enable-static-exec flag will now properly place the -static flag on the link line of all installed executables. This will force the executable to link with static libraries over shared libraries, provided the static libraries are available. MAM - 2009/08/31 - BZ #1583 - The --includedir=DIR configuration option now works as intended, and can be used to specify the location to install C header files. The default location remains unchanged, residing at ${prefix}/include. MAM - 2009/03/10 - BZ #1381 - Configure no longer removes the '-g' flag from CFLAGS when in production mode if it has been explicitly set in the CFLAGS environment variable prior to configuration. MAM - 2009/03/09 - BZ #1401. - Fixed error with 'make check install' failing due to h5dump needing other tools built first. MAM - 2008/10/24. - When using shared szip, it is no longer necessary to specify the path to the shared szip libraries in LD_LIBRARY_PATH. MAM - 2008/10/24. - The file libhdf5_fortran.settings is not installed since its content is included in libhdf5.settings now. AKC - 2008/10/21 - "make DESTDIR=xxx install" failed to install some tools and files (e.g., h5cc and fortran modules). Fixed. AKC - 2008/10/8. Performance ------------- - perf_serial test added to Windows projects and check batch file. (ADB - 2009/06/11) Tools ----- - Updated h5dump test case script to prevent entire test failure upon source directory is read-only. Bug# HDFFV-4342 (JKM 2011/4/12) - Updated to unify option name to '--enable-error-stack' for printing HDF5 error stack messages for HDF5 tools. h5ls and h5dump for now. For h5ls, this replaces "-e/--errors" option, which is deprecated. Bug#2182 (JKM 2011/3/30) - Fix h5diff for --use-system-epsilon option: the calculation changed from ( |a - b| / b ) to ( |a - b| ). This was decided for better performance. Bug#2184 (JKM 2011/3/24) - Fixed output for H5T_REFERENCE in h5dump. According to the BNF document the output of a H5T_REFERENCE should be followed by the type; ::= H5T_REFERENCE { } ::= H5T_STD_REF_OBJECT | H5T_STD_REF_DSETREG Previously this was only displayed if the -R option was used. Bug#1725 (ADB 2011/3/28) - Fix h5diff issues for #1: h5diff compared attributes correctly only when two objects have the same number of attributes and attribute names are identical, #2: didn't display useful information about attribute difference. Bug#2121 (JKM 2011/3/17) - Fixed memory leak for h5diff when accessing symbolic links with --follow-symlink option. Bug#2214 (JKM 2011/3/18) - Fixed memory leak for h5diff when access variable length string data. Bug#2216 (JKM 2011/3/18) - Fixed and improved help page for -a option of h5ls. Bug#1904 (JKM 2011/3/11) - Fixed h5dump not to include attribute values in the output file when h5dump "-y -o output_file" options were used. The problem was introduced in HDF5 1.8.6 by showing data pointed by region references. (XCAO 2011/3/9) - Fixed h5copy to be able to copy any object into the same HDF5 file. Previously h5copy displayed error message when target file is same as source file. (XCAO 2011/3/8) - Fixed h5dump for skipping some values for long array type dataset on Windows. This issue only occurred on Windows due to the different return behavior from _vsnprintf() funtion. Bug#2161 (JKM 2011/3/3) - Fixed h5dump for skipping array indices every certain number when the array type dataset is relatively big. The certain number varies according to the size of array. Bug#2092 (JKM 2011/2/15). - Fixed h5diff for the segfault when compares compound datasets with combination of fixed length string types and vlen string types in certain orders. bug#2089 (JKM 2010/12/28) - Improve h5diff performance. 1) use HDmemcmp() before comparing each elements. 2) replace expensive H5Tequals() calls 3) retrieve datatype information at dataset level not each element level for compound datasets - Fixed h5ls to display nested compound type with curly bracket when -S (--simple) option is used with -l (--label), so it shows which member (in curly bracket) belong to which nested compound type and make the output make sense. bug#1979 (JKM 2010/11/09) - Fixed h5diff to handle variable-length strings in a compound dataset correctly. (also variable-length string array in a compound dataset) Garbage values were displayed when h5diff compared multiple variable-length strings in a compound type dataset. Bug#1989 (JKM 2010/10/28) - Fixed h5copy to fail gracefully when copying object to non-exist group without -p option. Bug#2040 (JKM 2010/10/18) - Fixed to compare member objects and groups recursively when two files or groups are specified to be compared. Bug#1975 (JKM 2010/9/16) - Make h5repack be able to convert a layout to COMPACT for small size dataset as default. bug#1896 (JKM 2010/09/15) - Change h5ls not to manipulate special characters in object name or attribute name for smart display. bug#1784 (JKM 2010/06/28) - Fixed h5ls to return exit code 1 (error) when non-existent file is specified. bug#1793. (JKM 2010/04/27) - h5copy failed to copy dangling link when the link is specified directly. bug#1817. (JKM 2010/04/22) - h5repack lost attributes from a dataset of reference type. bug#1726. (JKM 2010/3/25) - h5repack sets NULL for object reference value for group or named datatype. bug#1814. (JKM 2010/03/19) - h5diff: fixed incorrect behavior (hang) in parallel mode when specify invalid options (ex: -v and -q) (JKM 2010/02/17) - h5dump/h5ls display buffer resize fixed in tools library. (ADB - 2009/07/21 - 1520) - Fixed many problems that could occur when using h5repack with named datatypes. (NAF - 2009/4/20 - 1516/1466) - h5dump, h5diff, h5repack were not reading (by hyperslabs) datasets that have a datatype datum size greater than H5TOOLS_BUFSIZE, a constant defined as 1024Kb, such as array types with large dimensions (PVN - 2009/4/1 - 1501) - h5import: By selecting a compression type, a big endian byte order was being selected (PVN - 2009/3/11 - 1462) - zip_perf.c had missing argument on one of the open() calls. Fixed. (AKC - 2008/12/9) - h5dump now checks for uniqueness of committed datatypes. (NAF - 2008/10/15) - Fixed unnecessary indentation of committed datatypes in h5dump. (NAF - 2008/10/15) - Fixed bugs in h5stat:segmemtation fault when printing groups and print warning message when traversal of objects is unsuccessful. (see bug #1253) (VC- 2008/10/13) - Fixed bug in h5ls that prevented relative group listings (like "h5ls foo.h5/bar") from working correctly (QAK - 2008/06/03) - Fixed bug in h5diff that prevented datasets & attributes with variable-length string elements from comparing correctly. (QAK - 2008/02/28) - h5import bug on Windows w/binary datasets. fread in windows needs a binary file to be open with 'rb' instead of 'r' otherwise it terminates execution if an end of file character is found on the input file. Besides that the binary file generated needs to be open with 'wb' , otherwise an end of line character is read twice. (PVN - 2008/02/19) - Fixed bug in h5dump that caused binary output to be made only for the first dataset, when several datasets were requested. (PVN - 2008/04/07) - h5dump: when doing binary output (-b), the stdout printing of attributes was done incorrectly. Removed printing of attributes when doing binary output. PVN - 2008/06/05 High-Level APIs: ------ - Fixed a bug in H5DSattach_scale, H5DSis_attached and H5DSdetach_scale caused by using H5Tget_native_type function to determine the native type for reading REFERENCE_LIST attribute. The bug was exposed on Mac PPC. (EIP - 2010/05/22 -1851) - Fixed a bug in the H5DSdetach_scale function when 0 bytes were allocated after the last reference to a dim. scale was removed from the list of references in a VL element of the DIMENSION_LIST attribute; modified the function to comply with the Spec: DIMENSION_LIST attribute is deleted now when no dimension scales left attached. (EIP - 2010/05/14 -1822) - Fixed a bug where the H5TB API would forget the order of fields when added out of offset order. (NAF - 2009/10/27 - 1582) - H5DSis_attached failed to account for different platform types. Added a get native type call. (ADB - 2009/9/29 - 1562) - Dimension scales: The scale index return value in H5DSiterate_scales was not always incremented. (PVN - 2009/4/8 - 1538) Fortran High-Level APIs: ------ - Lite: The h5ltget_dataset_info_f function (gets information about a dataset) was not correctly returning the dimension array. (PVN - 2009/3/23) - Lite: the h5ltread_dataset_string_f and h5ltget_attribute_string_f functions had memory problems with the g95 fortran compiler. (PVN � 5/13/2009) 1522 Documentation ------------- F90 APIs -------- C++ APIs -------- - The constructor PropList::PropList(id) was fixed to act properly according to the nature of 'id'. When 'id' is a property class id, a new property list will be created. When 'id' id a property list id, a copy of the property list will be made. (BMR - 2010/5/9) - The parameters 'size' and 'bufsize' in CommonFG::getLinkval and CommonFG::getComment, respectively, now have default values for user's convenience. (BMR - 2009/10/23) - NULL pointer accessing was fixed, bugzilla 1061. (BMR - 2009/10/05) - read/write methods of DataSet and Attribute classes were fixed to handle string correctly. (BMR - 2009/07/26) - Fixed bug that caused segfaults in Attribute::read. (BMR - 2008/04/20) - Fixed bug in PropList::getClassName to use portable HDfree instead of free. (BMR - 2008/04/20) - Fixed a design bug which allowed an Attribute object to create/modify attributes (bugzilla #1068). The API class hierarchy was revised to address the problem. Classes AbstractDS and Attribute are moved out of H5Object. Class Attribute now multiply inherits from IdComponent and AbstractDs and class DataSet from H5Object and AbstractDs. In addition, the data member IdComponent::id was moved into subclasses: Attribute, DataSet, DataSpace, DataType, H5File, Group, and PropList. (BMR - 2008/08/10) - IdComponent::dereference was incorrect and replaced as described in "New Features" section. (BMR - 2008/08/10) Platforms Tested ================ AIX 6.1 xlc 11.1.0.3 (NCSA BP) xlC 11.1.0.3 xlf 13.1.0.3 mpcc_r 11.1.0.3 mpxlf_r 13.1.0.3 FreeBSD 6.2-STABLE i386 gcc 3.4.6 [FreeBSD] 20060305 (duty) g++ 3.4.6 [FreeBSD] 20060305 gcc 4.2.1 20080123 g++ 4.2.1 20080123 gfortran 4.2.1 20070620 FreeBSD 6.2-STABLE amd64 gcc 3.4.6 [FreeBSD] 20060305 (liberty) g++ 3.4.6 [FreeBSD] 20060305 gcc 4.2.1 20080123 g++ 4.2.1 20080123 gfortran 4.2.1 20080123 Linux 2.6.9 (RHEL4) Intel 10.0 compilers (abe.ncsa.uiuc.edu) Linux 2.4.21-47 gcc 3.2.3 20030502 (osage) Linux 2.6.9-42.0.10 gcc 3.4.6 20060404 (kagiso) PGI 7.0-7 (pgcc, pgf90, pgCC) Intel 9.1 (icc, ifort, icpc) Linux 2.6.16.27 x86_64 AMD gcc 4.1.0 (SuSE Linux), g++ 4.1.0, g95 (GCC 4.0.3) (smirom) PGI 6.2-5 (pgcc, pgf90, pgCC) Intel 9.1 (icc, iort, icpc) Linux 2.6.5-7.252.1-rtgfx #1 SMP ia64 Intel(R) C++ Version 9.0 (cobalt) Intel(R) Fortran Itanium(R) Version 9.0 SGI MPI SunOS 5.10 Sun C 5.9 SunOS_sparc Patch 124867-11 2009/04/30 (linew) Sun Fortran 95 8.3 SunOS_sparc Patch 127000-11 2009/10/06 Sun C++ 5.9 SunOS_sparc Patch 124863-16 2009/09/15 IA-64 Linux 2.4.21.SuSE_292.til1 ia64 (NCSA tg-login) gcc 3.2.2 Intel(R) C++ Version 8.1 Intel(R) Fortran Compiler Version 8.1 mpich-gm-1.2.5..10-intel-r2 Windows XP Visual Studio 2008 w/ Intel Fortran 11.1 Visual Studio 2010 Cygwin(1.7.7 native gcc(4.3.4) compiler and gfortran) Windows XP x64 Visual Studio 2008 w/ Intel Fortran 11.1 Visual Studio 2010 Cygwin(1.7.7 native gcc(4.3.4) compiler and gfortran) Windows Vista Visual Studio 2008 w/ Intel Fortran 11.1 Windows Vista x64 Visual Studio 2008 w/ Intel Fortran 11.1 Windows 7 Visual Studio 2010 Windows 7 x64 Visual Studio 2010 MAC OS 10.5 (Intel) gcc i686-apple-darwin8-gcc-4.0.1 (GCC) 4.0.1 G95 (GCC 4.0.3 (g95 0.91!) Nov 21 2006) GNU Fortran (GCC) 4.3.0 20070810 Alpha Open VMS 7.3 Supported Configuration Features Summary ======================================== In the tables below y = tested and supported n = not supported or not tested in this release x = not working in this release ( ) = footnote appears below second table = testing incomplete on this feature or platform W or C indicates workstation or cluster, respectively. Platform C F90 F90 C++ zlib SZIP parallel parallel SunOS5.10 64-bit n y n y y y SunOS5.10 32-bit n y n y y y AIX 6.1 32- and 64-bit n y n y y y Windows XP n y(3) n(3) y y y Windows XP x64 n y(3) n(3) y y y Windows Vista n y(3) n(3) y y y Windows Vista x64 n y(3) n(3) y y y Mac OS X 10.5 Intel n y n y y y FreeBSD 6.2 32-bit n n n y y y FreeBSD 6.2 64-bit RedHat EL4 2.6.9 i686 GNU W y(2) y(4) y(2) y y y RedHat EL4 2.6.9 i686 Intel W n y n y y n RedHat EL4 2.6.9 i686 PGI W n y n y y n SuSe Linux 2.6.16 x86_64 GNU (5) W y(2) y n y y y SuSe Linux 2.6.16 x86_64 Int (5) W n y n y n n SuSe Linux 2.6.16 x86_64 PGI (5) W n y n y n n RHL9 Linux 2.4 Xeon Lustre Intel C n y n y y n RHEL3 Linux 2.4 Xeon Intel W n y n n y n RHEL4 Linux 2.6 Xeon Lustre Int C n y n y y n SuSE Linux 2.4 ia64 Intel C y(1) y y y y y SuSe Linux 2.6.5 SGI Altix ia64 Intel C n y n y n y Alpha OpenVMS 7.3.2 n y n y n n Platform Shared Shared Shared static- Thread- C libs F90 libs C++ libs exec safe SunOS 5.10 32-bit y y y x y SunOS 5.10 64-bit y y y x y AIX 6.1 32- and 64-bit n n n x y Windows XP y y(3) y y n Windows XP x64 y y(3) y y n Windows Vista y y(3) y y y Windows Vista x64 y y(3) y y y Mac OS X 10.5 Intel y y y x n FreeBSD 6.2 32-bit y y y x n FreeBSD 6.2 64-bit y y y x n RHEL4 2.6.9 i686 GNU W y y(4) y x y RHEL4 2.6.9 i686 Intel W y y y x n RHEL4 2.6.9 i686 PGI W y y y x n SuSE Linux 2.6.16 x86_64 GNU (5) W y y y x y SuSE Linux 2.6.16 x86_64 Intel(5) W y y y x n SuSE Linux 2.6.16 x86_64 PGI(5) W y y y x n RHL9 Linux 2.4 Xeon Lustre Intel C y y y x n RHEL3 Linux 2.4 Xeon Intel W y n n x n RHEL4 Linux 2.6 Xeon Lustre Intel C y y y x n SuSE Linux 2.4 ia64 Intel C y y y x n SuSe Linux 2.6.5 SGI Altix ia64 Intel C n n n x n Notes: (1) Using mpich 1.2.6. (2) Using mpich2 1.0.6. (3) Using Visual Studio 2008 (Cygwin shared libraries are not supported) (4) With PGI and Absoft compilers. (5) AMD Opteron x86_64 Compiler versions for each platform are listed in the preceding "Platforms Tested" table. Known Problems ============== * --with-mpe configure option does not work with Mpich2. AKC - 2011/03/10) * While working on the 1.8.6 release of HDF5, a bug was discovered that can occur when reading from a dataset in parallel shortly after it has been written to collectively. The issue was exposed by a new test in the parallel HDF5 test suite, but had existed before that. We believe the problem lies with certain MPI implementations and/or filesystems. We have provided a pure MPI test program, as well as a standalone HDF5 program, that can be used to determine if this is an issue on your system. They should be run across multiple nodes with a varying number of processes. These programs can be found at: http://www.hdfgroup.org/ftp/HDF5/examples/known_problems/ * Parallel mode in AIX will fail some of the testcheck_version.sh tests where it treats "exit(134) the same as if process 0 had received an abort signal. This is fixed and will be available in the next release. AKC - 2009/11/3 * Some tests in tools/h5repack may fail in AIX systems when -q32 mode is used. The error is due to insufficient memory requested. Request a large amount of runtime memory by setting the following environment variable for more memory. LDR_CNTRL=MAXDATA=0x20000000@DSA AKC - 2009/10/31 * The PathScale MPI implementation, accessing a Panasas file system, would cause H5Fcreate() with H5F_ACC_EXCL to fail even when the file is not existing. This is due to the MPI_File_open() call failing if the amode has the MPI_MODE_EXCL bit set. (See bug 1468 for details.) AKC - 2009/8/11 * Parallel tests failed with 16 processes with data inconsistency at testphdf5 / dataset_readAll. Parallel tests also failed with 32 and 64 processes with collective abort of all ranks at t_posix_compliant / allwrite_allread_blocks with MPI IO. (CMC - 2009/04/28) * For SNL, spirit/liberty/thunderbird: The serial tests pass but parallel tests failed with MPI-IO file locking message. AKC - 2007/6/25. * On Intel 64 Linux cluster (RH 4, Linux 2.6.9) with Intel 10.0 compilers use -mp -O1 compilation flags to build the libraries. Higher level of optimization causes failures in several HDF5 library tests. * For HPUX 11.23 many tools tests failed for 64-bit version when linked to the shared libraries (tested for 1.8.0-beta2) * For SNL, Red Storm: only paralle HDF5 is supported. The serial tests pass and the parallel tests also pass with lots of non-fatal error messages. * on SUN 5.10 C++ test fails in the "Testing Shared Datatypes with Attributes" test * configuring with --enable-debug=all produces compiler errors on most platforms. Users who want to run HDF5 in debug mode should use --enable-debug rather than --enable-debug=all to enable debugging information on most modules. * On Mac OS 10.4, test/dt_arith.c has some errors in conversion from long double to (unsigned) long long and from (unsigned)long long to long double. * On Altix SGI with Intel 9.0 testmeta.c would not compile with -O3 optimization flag. * On VAX, Scaleoffset filter isn't supported. The filter cannot be applied to HDF5 data generated on VAX. Scaleoffset filter only supports IEEE standard for floating-point data. * On Cray X1, a lone colon on the command line of h5dump --xml (as in the testh5dumpxml.sh script) is misinterpereted by the operating system and causes an error. * On mpich 1.2.5 and 1.2.6, we found that if more than two processes contribute no IO and the application asks to do IO with collective, we found that when using 4 processors, a simple collective write will be hung sometimes. This can be verified with t_mpi test under testpar. * The dataset created or rewritten with the v1.6.3 library or after can't be read with the v1.6.2 library or before when Fletcher32 EDC(filter) is enabled. There was a bug in the calculating code of the Fletcher32 checksum in the library before v1.6.3. The checksum value wasn't consistent between big-endian and little-endian systems. This bug was fixed in Release 1.6.3. However, after fixing the bug, the checksum value is no longer the same as before on little-endian system. The library release after 1.6.4 can still read the dataset created or rewritten with the library of v1.6.2 or before. SLU - 2005/6/30 * For the version 6(6.02 and 6.04) of Portland Group compiler on AMD Opteron processor, there's a bug in the compiler for optimization(-O2). The library failed in several tests but all related to multi driver. The problem has been reported to the vendor. * On IBM AIX systems, parallel HDF5 mode will fail some tests with error messages like "INFO: 0031-XXX ...". This is from the command poe. Set the environment variable MP_INFOLEVEL to 0 to minimize the messages and run the tests again. The tests may fail with messages like "The socket name is already in use". HDF5 does not use sockets (except for stream-VFD). This is due to problems of the poe command trying to set up the debug socket. Check if there are many old /tmp/s.pedb.* staying around. These are sockets used by the poe command and left behind due to failed commands. Ask your system administrator to clean them out. Lastly, request IBM to provide a mean to run poe without the debug socket. * The C++ library's tests fails when compiling with PGI C++ compiler. The workaround until the problem is correctly handled is to use the flag "--instantiate=local" prior to the configure and build steps, as: setenv CXX "pgCC --instantiate=local" for pgCC 5.02 and higher * The stream-vfd test uses ip port 10007 for testing. If another application is already using that port address, the test will hang indefinitely and has to be terminated by the kill command. To try the test again, change the port address in test/stream_test.c to one not being used in the host. * The --enable-static-exec configure flag will only statically link libraries if the static version of that library is present. If only the shared version of a library exists (i.e., most system libraries on Solaris, AIX, and Mac, for example, only have shared versions), the flag should still result in a successful compilation, but note that the installed executables will not be fully static. Thus, the only guarantee on these systems is that the executable is statically linked with just the HDF5 library. * With the gcc 2.95.2 compiler, HDF 5 uses the `-ansi' flag during compilation. The ANSI version of the compiler complains about not being able to handle the `long long' datatype with the warning: warning: ANSI C does not support `long long' This warning is innocuous and can be safely ignored. * Certain platforms give false negatives when testing h5ls: - Cray J90 and Cray T90IEEE give errors during testing when displaying some floating-point values. These are benign differences due to the different precision in the values displayed and h5ls appears to be dumping floating-point numbers correctly. * Not all platforms behave correctly with szip's shared libraries. Szip is disabled in these cases, and a message is relayed at configure time. Static libraries should be working on all systems that support szip, and should be used when shared libraries are unavailable. There is also a configure error on Altix machines that incorrectly reports when a version of szip without an encoder is being used. * On some platforms that use Intel and Absoft compilers to build HDF5 fortran library, compilation may fail for fortranlib_test.f90, fflush1.f90 and fflush2.f90 complaining about exit subroutine. Comment out the line IF (total_error .ne. 0) CALL exit (total_error) * Information about building with PGI and Intel compilers is available in INSTALL file sections 5.7 and 5.8 * On at least one system, (SDSC DataStar), the scheduler (in this case LoadLeveler) sends job status updates to standard error when you run any executable that was compiled with the parallel compilers. This causes problems when running "make check" on parallel builds, as many of the tool tests function by saving the output from test runs, and comparing it to an exemplar. The best solution is to reconfigure the target system so it no longer inserts the extra text. However, this may not be practical. In such cases, one solution is to "setenv HDF5_Make_Ignore yes" prior to the configure and build. This will cause "make check" to continue after detecting errors in the tool tests. However, in the case of SDSC DataStar, it also leaves you with some 150 "failed" tests to examine by hand. A second solution is to write a script to run serial tests and filter out the text added by the scheduler. A sample script used on SDSC DataStar is given below, but you will probably have to customize it for your installation. Observe that the basic idea is to insert the script as the first item on the command line which executes the the test. The script then executes the test and filters out the offending text before passing it on. #!/bin/csh set STDOUT_FILE=~/bin/serial_filter.stdout set STDERR_FILE=~/bin/serial_filter.stderr rm -f $STDOUT_FILE $STDERR_FILE ($* > $STDOUT_FILE) >& $STDERR_FILE set RETURN_VALUE=$status cat $STDOUT_FILE tail +3 $STDERR_FILE exit $RETURN_VALUE You get the HDF make files and test scipts to execute your filter script by setting the environment variable "RUNSERIAL" to the full path of the script prior to running configure for parallel builds. Remember to "unsetenv RUNSERIAL" before running configure for a serial build. Note that the RUNSERIAL environment variable exists so that we can can prefix serial runs as necessary on the target system. On DataStar, no prefix is necessary. However on an MPICH system, the prefix might have to be set to something like "/usr/local/mpi/bin/mpirun -np 1" to get the serial tests to run at all. In such cases, you will have to include the regular prefix in your filter script. * H5Ocopy() does not copy reg_ref attributes correctly when shared-message is turn on. The value of the reference in the destination attriubte is wrong. This H5Ocopy problem will affect h5copy tool