Maintenance for windows
Description:
Remove project file filenamedll.dsp because its source file filename.c was removed
Solution:
Platforms tested:
MSVS 6.0 on windows XP
Misc. update:
Configuration feature
Description:
Serial test output is now stored in log files and printed when all tests
in a directory complete, or when a test fails. This should make test output
more readable and useful.
Also made changes to clean up ii_files directories that are created by some
C++ compilers/linkers.
Also fixed a few minor Makefile bugs.
Solution:
When serial tests run, their output is saved in *.log or *.logsh. While
running, tests only print when they begin and when they complete; their
more specific output (from the log file) is printed if the test fails or
when all tests have completed.
Comments welcome.
Platforms tested:
mir, modi4 (parallel and serial), copper, shanti
Bug fix
Description:
Rewrite code for mounting files to clean up layers of kludges and implement
a much cleaner and more maintainable design.
Platforms tested:
FreeBSD 4.11 (sleipnir)
Linux 2.4
Correct a typo for function declaration.
It causes problems for windows DLLs.
Description:
In file hdf5\hl\fortran\src\H5LTf90proto.h
Line 183 'n' might be missed in the very beginning
"h5ltget_dataset_ndims_c"
should change to
"nh5ltget_dataset_ndims_c"
Solution:
change it.
Platforms tested:
too minor to test at other platform,
Fang reported this, so I assume it should work on windows.
Misc. update:
improve documentation by comments
changed the return value of H5DSget_num_scales for a case when there are no scales attached from FAIL (-1) to 0
Description:
Solution:
Platforms tested:
linux
solaris
Misc. update:
cleanup.
Description:
The install directory (--prefix) has been changed to $PWD/hdf5 long time
ago. It is no need to reset to installdir any more.
Platforms tested:
Change is trivial and will be tested in Daily tests.
Misc. update:
support thread-safe on windows
Description:
pthread on windows won't return a numeric value from pthread_safe.
An error message applying on linux won't apply on windows,
Print some other message.
Solution:
Platforms tested:
windows xp, too minor to test at other platforms.
Misc. update:
Description: When converting floating-point numbers to integers and the values of
floating-point number are greater than the maximal value of integer, Cray X1 generates
floating exception.
Solution: Added a test in configure to detect Cray X1's exception. Set a flag to
indicate the machine that can handle overflow converting all floating-point to all
integer types. This flag should be set for all machines, except for Cray X1 where
floating exception is generated when the floating-point value is greater than the
maximal integer value.
Platforms tested: Cray X1 and h5committest.
Bug fix
Description:
Update Makefiles now that docs are no longer in main hdf5 tree.
Solution:
Changed root-level Makefile.am not to recurse into docs tree.
Removed docs Makefiles from configure.in.
Removed config/commence-doc.am, since it is no longer used.
Platforms tested:
mir (change to Makefiles only)
Description:
All HDF5 user documentation has been moved to a separate hdf5doc/
repository, managed under Subversion.
With this 'cvs commit', all files are stripped from hdf5/doc/.
THIS CHANGE IS APPLIED ONLY TO THE HDF5 DEVELOPMENT BRANCH,
post Release 1.6.x; it is not applied to the release branches.
Bug fixes
Description:
A number of minor changes to Makefiles. Some files will now be cleaned
properly, some comments are more informative, etc.
Platforms tested:
heping, mir, modi4
upgrade.
Description:
Release used to copy the updated release_docs/HISTORY.txt and
release_docs/RELEASE.txt to the doc/html/ADGuide. The doc/ is
removed from the source tree. So, remove the command that
does the copying.
Note that once the separated doc/ location is known, release needs
to copy the two files into them and committed them via sub-version
commands.
Platforms tested:
No test because it is a trivial change and it is not easy to change
without triggering the full release process.
Misc. update:
Bug fix
Description:
Correct memory leak when a dataset is attempted to be opened, but turns
out to be a group or named datatype.
Also, clean up code that was leading to the leak and zero out empty
group entries to help prevent similar errors in the future.
Platforms tested:
FreebSD 4.11 (sleipnir)
Too minor to require h5committest
Bug fix
Description:
Further progress on fixing file mounting to work properly.
Platforms tested:
FreeBSD 4.11 (sleipnir)
Linux 2.4
Too minor to require h5committest
bug fix.
Description:
The deploy command would exit whether it succeeded or not.
Solution:
Fixed it so that it exits only when deploy has encountered failures.
Platforms tested:
Trivial change. Will watch the daily test results.
Misc. update:
Description: On VMS assertion failed due to a typo ("/" instead of "%")
Solution: Fixed a typo
Platforms tested: VMS (code is not exercised on other platforms)
Misc. update:
Purpose:
Make collective IO tests work with mpiposix
Description:
Although internally HDF5 is using independent IO for mpiposix driver,
it is still a good test to see whether HDF5 APIs can allow collective IO calls to pass
into mpiposix driver.
Previous tests ignore this test; now add it.
Solution:
Platforms tested:
AIX (copper), too minor to need to test with three platforms.
Misc. update:
Bug fix
Description:
The testh5diff.sh script fails on mcr because of random srun messages
thrown into the output
Solution:
Added some code to filter out the srun messages before the output is compared.
Platforms tested:
LLNL mcr
Misc. update:
Updating phase 2 work of collective IO
Description:
The current collective IO tests cannot test for number of processors bigger than
some values, this change will lift those restriction. However, the test may be slower.
Solution:
Platforms tested:
linux 2.4, AIX 5.1, Linux 2.4 IA64 and IRIX 6.5
(I haven't tested big number of processors with the restriction of the machine)
Misc. update:
Updating phase 2 work of collective IO
Description:
When the size is bigger than some value for MPI derived data type, mpich 1.2.5 and lower
version won't work any more, so in the configuration file, we will check the version number
and switch from collective IO to independent IO by turning off complex derived datatype.
Solution:
Platforms tested:
linux 2.4, AIX 5.1, Linux 2.4 IA64 and IRIX 6.5
Misc. update:
Updating phase 2 work of collective IO
Description:
When the size is bigger than some value for MPI derived data type, mpich 1.2.5 and lower
version won't work any more, so in the configuration file, we will check the version number
and switch from collective IO to independent IO by turning off complex derived datatype.
Solution:
Platforms tested:
linux 2.4, AIX 5.1, Linux 2.4 IA64 and IRIX 6.5
Misc. update:
Bug fix
Description:
An earlier bug fix to plotting when the minimum number of processors was not 1
caused the -procs argument to be ignored.
Solution:
Added a check to see what the -procs argument was set to and plot accordingly.
Platforms tested:
colonelk
Misc. update:
bug fix
Description:
rwdsetexample and attrexample both modify the same file created by
dsetexample. Need to serialize all three of them.
Platforms tested:
Did 50 times of make check in fortran/examples. It used to fail
at least once. Now it does not fail.
Description:
The use of FileCreatPropList::DEFAULT sometimes caused failure
in the reference counting area. This occurs to all the default
property lists, which also include FileAccPropList::DEFAULT,
DSetCreatPropList::DEFAULT, and DSetMemXferPropList::DEFAULT.
H5P_DEFAULT was used to create these default prop lists and
because its value is 0, the id of these prop lists are 0, which
is rejected by the H5I functions during the reference counting.
Solution:
The main action to fix the above problem was to use
H5P_FILE_CREATE
H5P_FILE_ACCESS
H5P_DATASET_CREATE
H5P_DATASET_XFER
to define the default property lists accordingly. Yet, when this
fix was applied, some bug in reference counting was revealed.
It appeared that some ids were not incremented but were passed in
for decrementing. The following actions were then taken to fix and
improve the current use of reference counting H5I functions.
* added private func IdComponent::p_valid_id to verify that the
id is a valid id and can be passed into an H5I function
* used p_valid_id to validate an id before calling an H5I functions
in the reference-counting member functions incRefCount,
decRefCount, and getCounter
* changed to use member function incRefCount, decRefCount, and
getCounter instead of the C APIs H5Iinc_ref, H5Idec_ref, and
H5Iget_ref throughout IdComponent.
In addition, overloadings were added for incRefCount, decRefCount,
and getCounter to take an id different than the id of the current
instance; they can be convenient during debugging.
Platforms tested:
Linux 2.4 (heping)
SunOS 5.8 64-bit (sol)
AIX 5.1 (copper)
Bug fix
Description:
Correct error where the dataset sieve buffer was inadvertantly disabled,
resulting in poor raw data I/O performance for certain I/O patterns.
Solution:
Hook dataset sieve buffer size setting back up.
Platforms tested:
FreeBSD 4.11 (sleipnir)
Solaris 2.9 (shanti)
Description: 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.
Solution: Made the library compare both the correct checksum and incorrect
checksum generated from v1.6.2 or before. This makes the library be
backward compatible but not forward compatible.
Platforms tested: h5committest and fuss.
Misc. update: Documented the forward incompatibility problem in the
Known Problem section in RELEASE.txt
Description: -xopnemp=stabs linking flag is needed on Solaris 2.9
in order to build Fortran library.
Solution: Added the flag to LDFLAGS if system version is 5.9
I don't really like the fix since this flag is now always added
for 5.9 systems even if Fortran is not enabled. If someone has
a better solution, please let me know :-)
Platforms tested: shanti, sol
Misc. update:
Description: Removed -DH5_ABSOFT flag for C compiler when Absoft f95 compiler
was used; this flag is not needed anymore since we automatically construct
names of C functions called from Fortran.
Solution:
Platforms tested: pommier
Misc. update:
update project file
Description:
new source codes added (H5BP.c H5BPcache.c H5BPdbg.c and H5BPtest
.c)
VS: [details about the changes, algorithm, etc...]
Platforms tested:
MSVS 6.0 on windows XP
Misc. update:
Update project file.
Description:
New source codes added (H5BP.c H5BPcache.c H5BPdbg.c and H5BPtest.c)
Solution:
Platforms tested:
MSVS 6.0 on windows XP
Misc. update:
new test hor h5repack, to syncronize tests between unix and windows
it requires a new file added
Description:
Solution:
Platforms tested:
linux
Misc. update:
Bug fix
Description:
The ".chksh" file for a test script was being created in the "source"
location rather than the build location. This can cause problems when
multiple builds are running because "slower" machines will see the ".cshsh"
file from faster machines and will not run the test script as they should.
Solution:
Use 'basename' command to strip off the path of the script and create
the ".chksh" file in the build location.
Platforms tested:
FreeBSD 4.11 (sleipnir)
h5committest
Catch up with linux.
Description:
dll b+tree project file b_plus_treedll.dsp was added due to
b+tree.c added on July 6
Solution:
Platforms tested:
MSVS 6.0 on windows XP
Misc. update:
