"Bug fix"
Description:
This version is showing very slow HDF5 I/O write speeds (about 40%
of MPIO). It was because the library is writing fill values by
default. There is a no-fill option which set to FALSE. Changed
it to TRUE since MPIO and POSIX are not doing fill values.
This is just a stop gap fix since the better fix would be to
change the library to not write fill-values in this case.
Platforms tested:
modi4 parallel and LANL Q paralllel.
No h5committest test because it is very trivia.
Misc. update:
Purpose:
Code cleanup
Description:
Moved a utility function out so other tests can use the same
function and avoid code duplication. Future utility functions
can also benefit from this move.
Platforms tested:
SunOS 5.7 (arabica)
Linux 2.4 (eirene)
IRIX 6.5.11 (modi4)
HPUX 11.00 (kelgia)
bug fix and cleanup
Description:
Previous change messed things up that fortranlib_test was linked
with multiple MAIN because $(TEST_OBJ) was used as if it was for
fortranlib_test only.
Solution:
Created a separated $(FORTLIBTEST_OBJ) for the fortranlib_test test.
Adjusted other related symbol names.
Also, removed DISTCLEAN=* which held files that should have been
removed in MOSTLYCLEAN.
Platforms tested:
"h5committested" (modi4, sol, verbena).
In the process of running fortran tests in arabica and kelgia.
Misc. update:
cleanup
Description:
Move the removal of libhdf5_fortran.settings from clean to
distclean target.
Platforms tested:
Tested in copper only since the change is trivial.
Misc. update:
Cleanup
Description:
Moved the cleanup of configure generated files to DISTCLEAN as
that should be the place they get removbed.
Platforms tested:
Tested in copper only as change is trivial.
Misc. update:
Cleanup
Description:
Remove h5c++ when distclean is called.
Solution:
Platforms tested:
Tested only in copper since the change is trivial.
Misc. update:
Cleanup
Description:
h5dumpgentest is already included in TEST_PROGS. Removed its old
entry in PROGS.
Platforms tested:
Only in Copper since this is trivial.
Misc. update:
Bug fix and minor code enhancement
Description:
Missing methods to read/write C++ String for an attribute and
a dataset.
Solution:
Added overloaded functions read and write to H5::Attribute and
H5::DataSet.
Also, added another constructor StrType so the need to separately
set the length of the string type can be eliminated. It's minor
but convenient.
Made some minor changes to make error messages more readable.
Platforms:
SunOS 5.7 (arabica)
Linux 2.4 (eirene)
IRIX 6.5.11 (modi4)
HPUX 11.00 (kelgia)
Purpose:
Added sanity check that min number of processes be not greater
than max number of processes.
Platforms tested:
copper only since this is a very simple change.
Misc. update:
Improvement
Description:
CHanged test dataset dimension sizes to (120, 240) so that
it can be tested with nprocs equal to 1, 2, 3, 4, 5, 6, 8, ...
Less restrictive than before.
Platforms tested:
only in CU since this is a pretty trivial change.
Misc. update:
Utility Addition
Description:
Added h5c++ script file to compile HDF5 C++ applications.
Modified so that these are compiled only during a make check-install
action and then only with the h5c++ script.
Platforms tested:
Verbena (Fortran & C++)
Arabica (Fortran & C++)
Modi4 (Fortran & Parallel)
Misc. update:
Bug Fix
Description:
The Dependencies file wasn't being generated correctly.
Solution:
The depned1.in file had old stuff in it. Replaced it with a copy of
the depend1.in file from the C++ directory.
Platforms tested:
Verbena (Only affects GNU platforms).
Misc. update:
Purpose:
Make "New Functions and Tools" section easier to read.
Solution:
Alphabetized functions and tools lists.
Created table with function names on left, full syntax on right.
Platforms tested:
Safari, IE 5
Bug fix.
Description:
The default setting of max transfer size is larger than
the default setting of number of bytes per process, result
in complaints that Transfer size is too big.
Solution:
Cap the min and max transfer size with number of bytes per process.
Always cap min transfer size with the max xfer size.
Platforms tested:
Copper.
Misc. update:
Twist code
Description:
Try to find a happy definition of HSSIZET_MAX, HSSIZET_MIN and HSIZET_MAX
for all platforms.
Platforms tested:
FreeBSD 4.8 (sleipnir)
h5committested
Code cleanup
Description:
h5pset_preserve_f function changed from an INTEGER to a LOGICAL parameter,
change it here too.
Platforms tested:
h5committested
Purpose: Maintenance
Description: Updated test to use new type of the argument for the
h5pset(get)_preserve_f function
Solution:
Platforms tested: arabica nad verbena
Misc. update:
Purpose: Bug fix
Description: One of the arguments to the h5pset(get)_preserve_f and
h5pset(get)_fapl_core_f had fortran INTEGER instead
of LOGICAL type.
Solution: Changed the type
Platforms tested: arabica and verbena
This is a very simple change, and I want to go home :-)
I will be doing more testing anyway on modi4 tomorrow.
Misc. update:
Bug fix
Description:
An earlier checkin changed some of the assumptions about single block
hyperslabs, causing them to fail in odd ways.
Solution:
Fix errors with single block hyperslabs by keying off of count==1 instead
of stride==1.
Platforms tested:
FreeBSD 4.8 (sleipnir) w/parallel
h5committested
added writing more attributes to one test file
Description:
added the writing of 3D attributes to one test file of h5ls
Solution:
Platforms tested:
Windows 2000 (octopus)
Linux 2.4 (rockaway)
SunOS 5.7 (arabica)
IRIX 6.5 (modi4)
Misc. update:
Purpose:
Bug Fix
Description:
The flags for the linker on MacOS X were invalid.
Solution:
If this is a darwin platform (MacOS X), then don't specify the
runtime linker flags.
Platforms tested:
MacOSX (since it only affects it...)
Misc. update:
Code cleanup & performance improvements
Description:
Optimize hyperslabs that are built to detect situations where "regular"
hyperslabs can be recovered from span tree descriptions.
Also, improve "same shape" routine to correctly work with all the different
combinations of selections.
Platforms tested:
FreeBSD 4.8 (sleipnir) w/C++
FreeBSD 4.8 (sleipnir) w/parallel
h5committested
Code cleanup
Description:
Move testing routines into their own module, to avoid linking them into
user's applications needlessly.
Platforms tested:
FreeBSD 4.8 (sleipnir) w/C++
FreeBSD 4.8 (sleipnir) w/parallel
h5committested
API tweak.
Description:
The H5Sget_select_bounds() API call was using hsize_t arrays for retrieving
the 'start' and 'end' coordinates, which is counter to the rest of the dataspace
API.
Solution:
Change the arrays to be hssize_t instead.
Platforms tested:
FreeBSD 4.8 (sleipnir) w/C++
FreeBSD 4.8 (sleipnir) w/parallel
h5committested
Misc. update:
Updated all docs for this change.
Added 1.4 compatibility #ifdef's
Minor corrections to match the man pages. Code provided by Pankaj.
Description:
* do not accept more than 30 file names
* corrected error messages.
Solution:
Platforms tested:
Tested on verbena.
Already tested on several platforms before integration.
Misc. update:
Purpose: Maintenance for 1.6 release
Description: I did more cleanup and restructuring of the file.
Please take a look at the file when you have a chance.
Thank you!
Solution:
Platforms tested:
Misc. update:
Bug fix
Description:
This would sometimes coredump in machines like PSC Alpha cluster
or linux icc if -g is not used.
Solution:
The real problem was because char string arrays were declared
without including space for the null terminator. E.g.,
char s[5][2]={ "s1", "s2", ...}
is not correct. Increased the size and also need to change
the H5S... definitions too.
Platforms tested:
"h5committested"--the h5dumpgentest passed without problem though
it failed in Fortran but that is a current known problem.
Tested in verbena using icc and also in PSC's cluster.
Misc. update:
Can't find this error in the v1.5 know problem list.
