added code warrior project file with szip support
to link with szip, it is necessary the precompiled library
located in
/szip/codewarrior/szip.lib
that is included in our szip distribution
Description:
Solution:
Platforms tested:
Code Warrior
Misc. update:
removed a small path warning on the code warrior project
this project is *without* szip support
Description:
Solution:
Platforms tested:
CW
Misc. update:
h5diff support for long_long
Description:
added h5diff support for long_long;
No automatic test was added because
it seems that the printf format of long_long is slightly different in Linux and IRIX(modi4),
because the diff command complained about a difference in the 2 test files for long_long;
visually, it seems that in IRIX an extra space is added after the print (or maybe it is
just me that is seeing extra spaces :=)
this test is test 5.8, commented in testh5diff.sh
Platforms tested:
Windows 2000 (octopus)
Linux 2.4 (rockaway)
SunOS 5.7 (arabica)
IRIX 6.5 (modi4)
Misc. update:
Bug Fix
Description:
The order the libraries are included in the compilation line is very
important.
Solution:
Need to specify the libraries in the order that they are given. If
we're building the application staticly, then use the static HDF5
library.
Platforms tested:
Modi4
Verbena
Arabica
Misc. update:
bug fix
Description:
the signed / unsigned comparison was not properly handled.
the fixtype function sets the sign for the memory type the same sign as found on disk
and the array_diff function only handled signed cases
Solution:
addded for each INTEGER size type an OR condition with the unsigned version
if the datatypes have different signs , comparison is NOT supported
Platforms tested:
Windows 2000 (octopus)
Linux 2.4 (rockaway)
SunOS 5.7 (arabica)
IRIX 6.5 (modi4)
Misc. update:
Speed up
Description:
Cache the settings of all number types to save configure time.
Reduced configure time by a whopping 40% (11m to 7m :-)
Platforms tested:
Sasn100 only (janus frontend.)
bug fix
Description:
Not all echo commands honors the '-n' option. In those cases,
the output look weired though recognizable.
Solution:
Eliminated the use of -n option. A better fix would be having
configure to set the proper use of the echo command.
Platforms tested:
only on sasn100 (a sun frontend of Janus).
No committest since this is just a simple shell script change.
Bug fix
Description:
(void *)~((size_t)NULL) on the Cray is a different value than
(foo *)~((size_t)NULL) and causes some of the hyperslab algorithms to
fail.
Solution:
Change all the 'void *' forms to 'foo *' forms.
Platforms tested:
Cray SV1
h5committest not needed.
Bug Fix
Description:
Two things:
- static libraries may depend upon the HDF5 library. We should
put them before the HDF5 library on the commandline instead of
afterwards to avoid conflict.
- the CFLAGS weren't being used during linking. This was causing
problems on Arabica where the flags are important to
distinguish between building a 32-bit and 64-bit application.
Solution:
- Added logic so that it catches the static libraries and puts them
before the hdf5 library on the command line. WARNING: This may not
work in all situations...but it should be okay for the good
majority.
- added the CFLAGS macro to the commandline...
Platforms tested:
LInux (Colonelk)
Arabica
Modi4 (Parallel)
(Small changes so doesn't need full h5committesting...)
Misc. update:
Purpose: Support for Fortran Absoft compiler on Linux
Description: Absoft has its own name translation rules
for the names of C functions called from Fortran
Solution: Added "if defined H5_ABSOFT" statement to the linux portion
of the H5f90i.h file to define correct translations.
For now, Fortran users who use ABSOFT fortran compiler have to specify
-DH5_ABSOFT flag for their C compiler.
Platforms tested: Remote laptop running SuSE8.1 with Absoft compiler,
verbena with icc and ifc just to make sure that
our current linux configuration still works.
Since it is a pure Linux change, I didn't test it on
other systems.
Misc. update:
Purpose: Maintenance
Description: all_withf90.zip file contained unnecessary
all.ncb all.opt hdf5.mcp files
Solution: Removed
Platforms tested: Windows 2000 with Intel and MVSC++ C and Fortran
compilers. SZIP fortran test fails in debug mode
while trying to create a file. It is a failure
similar that Bill reported for 1.5.54 failure on Linux
2.4 with pgf90 compiler. I am looking into the problem.
Misc. update:
Purpose: bug fix for Tru64 and test of the latest szip library.
Description: h5diff uses basename function that also is defined
in the system string.h file. That caused compilation to
fail.
Solution: renamed basename to h5diff_basename
Platforms tested: verbena, arabica, arabica-64,
and PSC Compaq machine.
Misc. update:
Purpose: Maintenance and bug fixes and test of latest szip library..
Description: Recently Quincey enabled MEDIUM test for istore. That caused
daily tests to fail on verbena. It looks like that test
creates a pretty big file, and afs write fails. Test passes
when run in /tmp.
Solution: Disable MEDIUM test and filed bug report to investigate the problem
little bit more carefully.
Platforms tested: verbena, arabica, arabica-64 (since I also tested newest szip)
Misc. update:
bug fix and enhancement
Description:
combined the --simple option to the display of NATIVE datatypes
with the -S option , the print of the datatype will always display the full description
making it possible to do an automatic test for catching differences on output
made the case for printing the type of bitfield data, which was not handled
the new test generates a dataset with 10 attributes, each one a different datatype
Platforms tested:
Windows 2000 (octopus)
Linux 2.4 (rockaway)
SunOS 5.7 (arabica)
IRIX 6.5 (modi4)
Misc. update:
Bug fix
Description:
Correct typo which didn't show up during my previous testing in production
mode.
Platforms tested:
FreeBSD 4.8 (sleipnir)
h5committest not necessary.
Bug fix
Description:
The chunk dataspace selection information for raw data I/O are being leaked.
Solution:
Free the chunk information during the cleanup code.
Platforms tested:
Solaris 2.6 (baldric) w/purify
h5committest not needed.
add a new project file for Code Warrior
Description:
the old project was with the old H5Farray.c file, causing a build error
Solution:
removed the file from the project settings
Platforms tested:
CW
Misc. update:
Purpose: Maintenance for the second round of testing
Description: Used bin/h5vers to change version number to 1.5.55
Solution:
Platforms tested: arabica
Misc. update:
Replaced the individual setting of gcc flags with the combined
gnu-flags file.
Platforms tested:
Tested in modi4 with different combinations of gcc. No needed to
run three platforms tests since this is for irix machines only.
Misc. update:
bug fix
Description:
when the list attribute function was called with values of type reference, the format string was not initialized
causing a segmentation fault on printf
Solution:
initialized the string ( with obj no and file no turned on )
Platforms tested:
Windows 2000 (octopus)
Linux 2.4 (rockaway)
SunOS 5.7 (arabica)
IRIX 6.5 (modi4)
Misc. update:
Add printout to identify when the MPIPOSIX driver is being tested instead
of the MPI-IO driver..
Platforms tested:
Tested in copper in parallel mode only since it was just a simple
print statement in the parallel test code.
Misc. update:
Small feature.
Description:
Change the wording for the use of HDF5_PARAPREFIX from Remark to hint.
Also made process 0 be the only one printing the hint rather all
processes duplicating the same hints.
Platforms tested:
Copper, both serial and parallel.
Did not do h5commit test since this is very localized and the above
tests covered both cases already.
Misc. update:
Bug fix
Description:
Fix three (!) bugs in szip filter:
- we were using bytes per pixel instead of bits per pixel
- we were using the size of the slowest changing dimension instead of
the fastest changing dimension to compute the blocks per scanline
parameter
- we swapped two parameters when setting up szip_options block.
Solution:
Addressed bugs above.
Platforms tested:
FreeBSD 4.8 (sleipnir)
h5committest not needed
Code cleanup
Description:
Remove debugging information and match bug-fix changes to szip filter
code.
Platforms tested:
FreeBSD 4.8 (sleipnir)
h5committest not needed.
fixed a typo on the usage message, redone the test files that deal with with
Description:
Solution:
Platforms tested:
Windows 2000 (octopus)
Linux 2.4 (rockaway)
SunOS 5.7 (arabica)
IRIX 6.5 (modi4)
Misc. update:
changes of input/ output format
Description:
on web page /RFC/h5diff
Solution:
Platforms tested:
Windows 2000 (octopus)
Linux 2.4 (rockaway)
SunOS 5.7 (arabica)
IRIX 6.5 (modi4)
Misc. update:
Bug fix
Description:
Raw data I/O on chunked datasets would attempt to allocate data structures
proportional to the number of chunks in the dataset on disk, instead of just
the number of chunks that the I/O operation would interact with, causing
operations on datasets with large #'s of chunks to fail (or become very slow),
even though the actual I/O operation was very modest.
Solution:
This is the "scalability fix" for chunked datasets that I've mentioned
we need to do, althought it's not the complete fix for the issue. Read on
for the details...
Only create data structures for the chunks that the I/O operation will
actually act on, reducing the amount of information allocated in memory,
normally.
I say "normally", because this algorithm has the same problems as the
original algorithm (worse actually, since the data structure for each chunk
is larger now) if _all_ the chunks in a dataset with a lot of chunks are
actually involved in the I/O operation. If that is the case, this code
will fail in a similar way.
To truly fix the problem, we would need to only create data structures for
a particular number of chunks, perform the I/O on just those chunks, then
release the data structures for those chunks and create data structures for
the next set of chunks to access, etc. However, I think this case is pretty
rare right now and we should worry about it after the 1.6.0 release.
Platforms tested:
h5committested
New feature
Description:
Added "fast comparison" code for hsize_t's, since they are used in th
raw data chunking I/O code now.
Platforms tested:
h5committested
Update
Description:
GPFS should only be checked for if the --enable-gpfs flag is used.
Otherwise, it shouldn't.
Platforms tested:
Linux (Configure only fix so only needed to test it to make sure the
configure does the right thing and that the build won't
break. I need to test it on a machine that has GPFS to make
sure it works there...)
Bug Fix
Description:
Lots of things. Mostly, multiple libraries declared on the command
line bothered some compilers. Then the order of the libraries is
important. Then if you are building static, you don't want an
"-lhdf5" flag in your compile line.
Solution:
Various hacks to remove redundant libraries from the compile line and
to keep the order of them and to replace -lhdf5 with the .a file when
building static.
Platforms tested:
Modi4 (Fortran & Parallel)
Verbena (Fortran & C++)
Arabica (Fortran)
Misc. update:
Bug Fix
Description:
The H5HG module had a couple of lines like the following:
HDmemcpy(foo + 1, foo, size);
This would overwrite the first character of "foo" for the total of
"size" bytes.
The H5Pfapl.c is an update of the trace information.
Solution:
Use HDmemmove instead. (Suggested by George Lewandowski at Boeing.
Thanks! :-)
Platforms tested:
Modi4 (Fortran & Parallel)
Verbena (Fortran & C++)
Arabica (Fortran)
Misc. update:
Fix
Description:
The SZLIB stuff was being checked by default. This isn't right as not
everyone will have SZLIB stuff.
Solution:
Made the default "no" instead of "yes"
Platforms tested:
Modi4 (Fortran & Parallel)
Verbena (Fortran & C++)
Arabica (Fortran)
Misc. update:
