Description:
- Added functions to query chunk information:
H5Dget_num_chunks(dset_id, fspace_id, *nchunks)
Gets the number of written chunks that intersect with the given
dataspace. However, in this version, the intersection is not
yet completed. Thus, the number of all written chunks will be
returned.
H5Dget_chunk_info_by_coord(dset_id, *offset, *filter_mask, *addr, *size)
Given a chunk's logical coordinates, returns the chunk's filter,
address, and size.
H5Dget_chunk_info(dset_id, fspace_id, index, *offset, *filter_mask, *addr, *size)
Given a chunk's index, returns the chunk's logical coordinates, filter,
address, and size. The chunk belongs to a set of chunks that have
nonempty intersection with the specified dataspace. However, in
this version, the intersection is not yet completed, and the index
is of all the written chunks.
These functions comply with VOL.
- Fixed some oversights found in the library for the tests in chunk_info.c
to work correctly. The returned value from a callback function was not
checked in H5EA_iterate(), H5FA_iterate(), and H5D__none_idx_iterate().
This oversight caused a callback function to continue iterating even though
it's supposed to stop.
Platforms tested:
Linux/64 (jelly)
Linux/64 (platypus)
Darwin (osx1011test)
* Macro cleanup and obvious warning fixes in parallel code.
* Moved H5FD comm and info dup/free wrapper code to a new H5mpi.c file
and separated it to deal with each MPI type separately.
* commit '7d291d059da6287f769951c4854c566da0b93ce8':
Parallel Fortran tests now use the MPI module instead of including mpif.h.
First stab at gfortran versions. Tested on gfortran 7-9 so far.
Moved the -std=f2008 option to gfortran 9 for now.
Updated gfortran options. * Removed conclude_fc.am, which was causing verbose builds. Need to investigate more, though. * Added gfortran 9 warnings and flags. This will need to be separated into other version blocks based on when the options first appeared. * The standard is now explicitly set at 2008.
* Removed conclude_fc.am, which was causing verbose builds. Need to
investigate more, though.
* Added gfortran 9 warnings and flags. This will need to be separated
into other version blocks based on when the options first appeared.
* The standard is now explicitly set at 2008.
* commit 'dfdc27b04b2e8bcd1985ba90ce6553d8b3805fda':
Fix release notes based on feedback from pull request.
Fix for HDFFV-10800 H5Ocopy failure: The value for the H5F_LIBVER_V18 index in H5O_fill_ver_bounds[], the format version bounds array for fill value message, should be version 3 not 2.
* commit '0742c15fca95965ed4eb87c6d54f41bb07bd4526':
makefile clean-up
move the vol_info to an optional parameter
added H5Pset_vol
added vol connector
added test for vol_connector
removed H5VL constants
removed VOLL NULL connector and replaced it with native VOL
added connector_name test
added missing files
add tests
added more tests
added missing files
Implemented VOL APIs and C constants, no tests
Implemented VOL APIs and C constants, no tests
initial API impl.
disabling tests.
Moved h5cc.in from tools/src/misc to src directory to always create h5cc
whether or not tools are enabled.
Added configuration status of tools and tests to libhdf5.settings.
* H5VLconnector.h for terminal connector things
* H5VLconnector_passthru.h for passthrough connector things
Note that these headers are arranged such that they are included in
hdf5.h so VOL connectors only need to include that. The separation into
multiple headers is mainly for readability.
* commit 'b5ef82a1786605ae86502bc82086047720b7d4ca': (21 commits)
Moving the handling of null prefix into H5_combine_path.
Changing the prefix of external file and VDS from empty string to null for performance improvement.
Improving the condition checking of empty string.
Replacing string operation strdup with assignment for empty string.
Adding back links_env.out which I accidentally removed in my previous commit.
Taking out two unnecessary diff output files.
Taking out unnecessary diff files for output.
Adding some comments.
Some coding style changes.
Adding the standard output files for the external_env.c and vds_env.c tests.
Small correction for my previous commit.
Forgot to add external_common.c and external_common.h.
Updated CMake for the splitting of external.c and vds.c.
Minor fix: removal of unnecessary enum values.
Minor fixes: updating the test vds_env.c according to the set up of vds.c.
Left out this file in previous commit.
HDFFV-10658 - setting and getting properties in API context: 1. switched to use the existing H5F_prefix_open_t for enum type; 2. put the common private function used by external.c and external_env.c into external_common.c
This commit basically has the following changes: 1. restored the datatype, dataspace, and LCPL of the dataset for VOL connector back to the properties. 2. splitted external.c and vds.c because they called HDsetenv in the program, instead using shell scripts to set the environment variables. 3. changed H5CX_get_vds_prefix and H5CX_get_ext_file_prefix to use H5P_peek instead of H5P_get.
HDFFV-10658: I left out this file in my previous commit.
HDFFV-10658: 1. moving HDgetenv to dataset initialization stage to reduce the overhead; 2. restoring the retrieval of three vol properties to H5P_get instead of using API context to prepare for Quincey's upcoming refactoring work.
...
1. switched to use the existing H5F_prefix_open_t for enum type;
2. put the common private function used by external.c and external_env.c into external_common.c
* commit 'd6c2a96ac2f103d90b96d5b39814810e6a31ef99':
Updated the parallel install docs.
Eliminated the need for a separate script variable.
Added a helpful message to the flush script.
Added a shell script so we can run the parallel flush test on OpenMPI.
after deleting attributes in densed storage.
The fix: When deleting attribute nodes from the name index v2 B-tree,
if an attribute is found in the intermediate B-tree nodes, which may be
merged/redistributed in the process, we need to free the dynamically
allocated spaces for the intermediate decoded attribute.
* commit '991996b25122ff010b9af1bbb6cc05be491ce038':
Remove unused CMake files
Update option text
Add reference file to list
HDFFV-10656 remove moved source
HDFFV-10656 Add CHECK_VOL support to CMake
* commit '55e87e5b86aacbb37c57614cd2e869541291caad': (51 commits)
Correct Windows build.
Refactor allocating & releasing pass through wrapper objects.
Fix typo
Remove duplicated comment.
Add a couple of missing prototypes for static routines, along with updating a comment from VOL plugin -> connector.
Switch H5VL_class_value_t from unsigned to int, and add "invalid" ID value.
Correct "make check-vol" regression test target.
Switch H5VL_class_value_t from enum to unsigned integer
Add support for "make check-vol", along with a few minor cleanups, etc.
Refactor infrastructure for setting FAPL information from environment variables during testing, including connecting native, pass-through, and dynamically loaded VOL connectors. Also bring native and pass-through VOL connectors into alignment, removing the "H5VLnative_private.h" header.
Remove example VOL connector, since it's been superceded by the pass-through VOL connector in the src subdirectory.
Add pass-through VOL connector
Update API tracing info for VOL functions and regenerate trace macros for VOL API calls.
Added 'notify' callback for async requests; switched VOL class and info comparison to return comparison value as parameter, so they can return error values; "cancelled" -> "canceled"; switched order of 'wrap_object' and 'free_wrap_ctx' management callbacks.
Add using FAPL from h5_fileaccess() to more tests.
Remove unused test for unimplemented routine.
Switch loc_params to VOL callbacks to pass struct by pointer instead of by value.
Add info_to_str and str_to_info "management" callbacks for serializing and deserializing a connector's info object.
Revert some of the changes to support the original property list value of metadata read attempts.
Add VOL connector info to the flie access property list returned from H5Fget_access_plist(). Also, other misc. cleanups, etc.
...
variables during testing, including connecting native, pass-through, and
dynamically loaded VOL connectors. Also bring native and pass-through
VOL connectors into alignment, removing the "H5VLnative_private.h" header.
Add more sample batch scripts, specifically for sbatch, not for knl
cross compile.
Don't run parallel tests when no parallel test script is configured in
HDF5options.cmake.
* commit 'ee12aff2ca89e92b99590f29096794a84c6bacdc':
Removed a couple of extraneous commas from VOL example.
Cleaned up VOL example.
Updated MANIFEST file.
Fixed VOL example to use a platform-independent format specifier for size_t.
and CMake.
* CMake will no longer install private generated headers and tools
library headers.
* Several empty public header files (which represent internal packages)
were removed. These were only installed by CMake.
* Autotools installs will install H5FDwindows.h.
Fixes HDFFV-10614.
Description:
- Added class DSetAccPropList for the dataset access property list.
- Added wrapper for H5Dget_access_plist to class DataSet
// Gets the access property list of this dataset.
DSetAccPropList getAccessPlist() const;
- Added wrappers for H5Pset_chunk_cache and H5Pget_chunk_cache to class
DSetAccPropList
// Sets the raw data chunk cache parameters.
void setChunkCache(size_t rdcc_nslots, size_t rdcc_nbytes, double rdcc_w0)
// Retrieves the raw data chunk cache parameters.
void getChunkCache(size_t &rdcc_nslots, size_t &rdcc_nbytes, double &rdcc_w0)
- Added two more arguments to H5Location::createDataSet:
const DSetAccPropList& dapl = DSetAccPropList::DEFAULT
const LinkCreatPropList& lcpl = LinkCreatPropList::DEFAULT
- Added one more argument to H5Location::openDataSet:
const DSetAccPropList& dapl = DSetAccPropList::DEFAULT
Platforms tested:
Linux/64 (jelly)
Linux/32 (jam)
Darwin (osx1010test)
* commit '08de02c838c05993fea5febb9c320a679e7f841a':
Changes based on feedback from pull request.
Fix test_misc33() in test/tmisc.c Open the test file read-only so that it can be accessed for testing.
Modifications made based on feedback from pull request.
Fix for HDFFV-10333: 1) Check for valid object header version for a refcount messge 2) Check for invalid fill value size 3) Check for invalid dimension size in a layout message 4) Add --enable-error-stack option to h5stat 5) Add error checks to h5stat.c 6) Add tests to h5stat and h5dump
Fix daily test failure.
