[svn-r28379] HDFFV-9573: Our helper tests bin/release and bin/h5vers are not working properly
bin/release uses the content of MANIFEST file to build the release tarball. But all the files generated by autogen had been removed. Therefore the release would not contain those needed generated files for a proper release.
Solution: Added all those generated files to the tail of MANIFEST and titled it accordingly. (Somehow bin/chkmanifest does not complain against these entries though they are not...
[svn-r28367] Putative fix for SWMR-95: The H5Pset_elink_acc_flags() API call rejects
H5F_ACC_SWMR_READ and H5F_ACC_SWMR_WRITE.
This has been tentatively fixed by updating the API call to accept
the SWMR flags. The only test at this time is a smoke test to
ensure correct flag setting. The SWMR behavior of external links
will be tested later.
Tested on Ubuntu 15.10 (Linux 4.2.0 x86_64) gcc 5.2.1
serial only
[svn-r28362] Merge of r28234 to r28346 from the trunk.
This includes the flex/bison changes to both the autotools
and CMake.
Tested on Ubuntu 15.10 (Linux 4.2.0 x86_64) gcc 5.2.1
serial and parallel (MPICH 3.1.4)
[svn-r28353] Description:
Checkpoint progress on code style & warning cleanups in new dataset code.
Tested on:
MacOSX/64 10.11.1 (amazon) w/serial & parallel
(h5committest not required on this branch)
[svn-r28351] Description:
Checkpoint refactoring work on chunked dataset indexing.
Tested on:
MacOSX/64 10.11.1 (amazon) w/serial & parallel
(h5comittest not required on this branch)
[svn-r28348] Description:
Checkpoint warning and coding style cleanups.
Tested on:
MacOSX/64 10.11.1 (amazon) w/serial & parallel
(h5committest not required on this branch)
[svn-r28344] Description:
Clean up H5LD test code.
Tested on:
MacOSX/64 10.11.1 (amazon) w/serial & parallel
(h5committest not required on this branch)
[svn-r28330] Implementation for version 3 superblock and latest version support.
Tested on quail, kite, osx1010test, platypus, ostrich, moohan, emu, jam.
[svn-r28327] Description:
More progress on revise_chunks review/normalization
Tested on:
MacOSX/64 10.11.1 (amazon) w/serial & parallel
(h5committest not required on this branch)
[svn-r28324] Description:
Code cleanups during review/normalization pass.
Tested on:
MacOSX/64 10.11.1 (amazon) w/serial & parallel
(h5committest not required on this branch)
[svn-r28322] SWMR-26: make check-vfd runs for a long time
The problem is in test/dsets which infinite looped when HDF5_DRIVER is "sec2" because
it had a for loop using a bool
type variable (compress) but the loop tried to increment to a third value. Changed the
type of compress to int allows the loop to run properly.
Secondly, the resas why it did not infinite loop when not in check-vfd is because the
getenv(HDF5_DRIVER) mistakenly assign driver as "nomatch" if it is not defined.
dsets skipp...
[svn-r28309] Changed prefix to swmr2 after creating a snapshot for DLS.
Tested on platypus. There were several issues with the bin/release and bin/h5vers scripts, but tar ball and versions look OK
after manual inspections.
[svn-r28307] Updated the dataset refresh code so that VDS source datasets
are correctly updated.
Tested on Ubuntu 15.10 (Linux 4.2.0 x86_64) w/ gcc 5.2.1
serial + parallel (MPICH 3.1.4)
[svn-r28282] Fix for cache test failure in production mode only.
Cause of failure was stack corruption caused by type mismatch
between expected and actual return value type in the validate
checksum callback.
Tested serial (debug and production) on mercury and jam
parallel (debug) on mercury.
[svn-r28237] Merge of r28228-9 from trunk.
accum test still has issues
Tested on: Ubuntu 15.04 (Linux 3.19 x86_64) gcc 4.9.2
serial autotools and CMake (3.3.2)
[svn-r28218] Merge of r28138, 28196, and 28204 from trunk.
hbool_t changes
Tested on: Ubuntu 15.04 (Linux 3.19 x86_64) gcc 4.9.2
serial w/ Fortran and C++
parallel (MPICH 3.1.4) w/ Fortran
[svn-r28216] Merge of r27969:28132, 28144:28193 from trunk.
Mostly CMake changes, but there's also a few minor library
changes.
Temporarily commented out the accum test. The test builds
but one of the sub-tests fails due to a file issue.
Tested on: Ubuntu 15.04 (Linux 3.19 x86_64) gcc 4.9.2
serial w/autotools
serial w/CMake (3.3.2, w/ fortran and C++)
[svn-r28212] Add code from r28179 back to the branch. This allows h5watch
to open VDS files, though it won't pick up changes to source
files until they are closed. This is a library problem and
will be fixed later.
Tested on: Ubuntu 15.04 (Linux 3.19 x86_64) gcc 4.9.2
[svn-r28179] h5watch can now open VDS files, though it will not detect changes
until the source files are closed.
Tested on: Ubuntu 15.04 (Linux 3.19 x86_64) gcc 4.9.2
serial only
[svn-r28178] Updates to existing VDS + SWMR test code and added reader code.
NOTE: Tests are still manual for now and the reader does not
reflect source dataset changes.
Tested on: Ubuntu 15.04 (Linux 3.19 x86_64) gcc 4.9.2
serial only
[svn-r28164] Added VDS SWMR test generator and writer files. These test files
are executed manually for now.
Tested on: Ubuntu 15.04 (Linux 3.19 x86_64) gcc 4.9.2
serial only
[svn-r28109] Description:
Correct failure when using the family VFD, which requires that the '%'s be
parsed twice.
Tested on:
MacOSX/64 10.10.5 (amazon) w/serial & parallel (and make check-vfd)
(h5committest not required on this branch)
[svn-r28099] Description:
Bring r27958 from trunk to branch.
Tested on:
MacOSX/64 10.10.5 (amazon) w/serial & parallel
(h5committest not required on this branch)
[svn-r28095] Description:
Bring r27884, 27888, 27889, 27890, and 27891 from trunk to branch.
Tested on:
MacOSX/64 10.10.5 (amazon) w/serial & parallel
(h5committest not required on this branch)
[svn-r28094] Description:
Bring r27849 & 27851 from trunk to branch.
Tested on:
MacOSX/64 10.10.5 (amazon) w/serial & parallel
(h5committest not required on this branch)
[svn-r28093] Merge of r27768 and r27774 from trunk
Description:
Complete revamp of package initialization/shutdown mechanism in the library.
Each package now has a single init/term routine.
This new way should avoid packages being re-initialized during library
shutdown and is also be _much_ more proactive about giving feedback for
resource leaks internal to the library.
Introduces a new "module" header file for packages in the library
(e.g src/H5Fmodule.h) which sets up some ne...
[svn-r28092] Description:
Merge the storage message back into the layout message.
Tested on:
MacOSX/64 10.10.5 (amazon) w/serial & parallel
(h5committest not required on this branch)
[svn-r28088] Fixed a mid-block declaration in the dsets test that only exists
when zlib compression is disabled.
Tested on: Ubuntu 15.04 (Linux 3.19 x86_64) gcc 4.9.2
[svn-r28083] Description:
Revise code to factor out check for chunked datasets more clearly.
Tested on:
MacOSX/64 10.10.5 (amazon) w/serial & parallel
(h5committest not required on this branch)
[svn-r28082] Description:
Switch to peek/poke for setting the chunk options.
Tested on:
MacOSX/64 10.10.5 (amazon) w/serial
(h5committest not required on this branch)
[svn-r28068] Removed duplicated Perl check from configure.
Updated the testflushrefresh script to correctly use Perl.
Tested on: Ubuntu 15.04 (Linux 3.19 x86_64) gcc 4.9.2
[svn-r28065] Clean up dcpls in the cork and flushrefresh tests.
Not cleaning them up was causing infinite closing loops with
the incoming init/shutdown changes from the trunk.
Tested on: Ubuntu 15.04 (Linux 3.19 x86_64) gcc 4.9.2
[svn-r28049] Merge of r27997-28047 from the trunk.
Tested on: Ubuntu 15.04 (Linux 3.19 x86_64) gcc 4.9.2
serial w/ Fortran and C++
parallel (MPICH 3.1.4) w/ Fortran
[svn-r28043] Description:
Merge r27818 from trunk to branch.
Tested on:
MacOSX/64 10.10.5 (amazon) w/serial & parallel
(h5committest not required on this branch)
[svn-r28042] Bug fix for failure in testpar/t_cache
Problem appears to have been caused by new code which
dirties a cache entry unexpectedly at file close. This
in turn throws off the write counts maintained by the
test, causing it to fail.
Solution was to manually flush the cache prior to file
close, verify the expected writes, and then de-register
the callback used to obtain write counts from the cache.
After cache shut down, I added additional code to verify
that no test entries were ...