Added a test of fill value before any data is written to a dataset.
Rename short_dataset() as dataset_fillvalue() as it reflects better
the tests. Also removed the option of -S since the fill value test
will be tested always.
Platforms tested:
"h5committested"
Misc. update:
Bug fix
Description:
Fix szip FORTRAN tests..
Solution:
Increase the chunk size to be bigger than the szip 'pixels per block'.
Also add in code that checks for a maximum number of errors and doesn't keep
reporting that thousands of errors have occurred.
Platforms tested:
Linux 2.4 (verbena) w/szip & FORTRAN
Not tested w/h5committest since this combination isn't covered by it.
Bug fix
Description:
Add special-case handling to floating-point conversion tests to avoid
problems with denormalized values on Cray T3E & T90 platforms. (Still not
working on Cray SV1, but at least it's closer).
Solution:
Detect denormalized values and don't try to operate on them on the Crays.
Platforms tested:
FreeBSD 4.9 (sleipnir)
Cray T3E (hubble.cray.com)
Cray T90 (gypsy.cray.com)
Bug fix
Description:
Range check the szip 'pixels per block' against the chunk size of the
dataset when attempting to create a new dataset, since the szip library
requires the PPB to be at least the size of the fastest changing dimension
in the chunk.
Platforms tested:
FreeBSD 4.9 (sleipnir)
too minor for h5committest
Code cleanup
Description:
Tweaked down the size of the szip 'pixels per block' for a test, now that
the library is range-checking this value more strictly.
Platforms tested:
FreeBSD 4.9 (sleipnir)
too minor for h5committest
Code cleanup
Description:
Removed redundant testing script from being executed, since it's just
running the program in the "TEST_PROGS" macro.
Platforms tested:
FreeBSD 4.9 (sleipnir)
too minor for h5committest
Code cleanup.
Description:
The H5Eclear() in the VRFY and INFO macros are not needed.
After removing them, there is no need to have a separate
v1.6 Compat version.
Platforms tested:
"h5committested"
Misc. update:
Bug fix
Description:
Fixed bug in hyperslab iteration where certain combinations of flattened
and non-flattened dimensions would cause incorrect locations to be iterated
over.
Platforms tested:
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
h5repack new features
Description:
added support/ tests for contiguous and compact layout processing
Solution:
Platforms tested:
linux
solaris
AIX
Misc. update:
h5repack new features
Description:
added checking routines for the filters that were applied to the output file
added tests for szip filter
Solution:
Platforms tested:
linux
solaris
(IRIX is not available)
:
Misc. update:
Code cleanup, bug fixes
Description:
Wrap up rest of changes necessary for fixing the "short" MPI-I/O read
problem that Robb reported.
Platforms tested:
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Code cleanup
Description:
Switch 'malloc()' to 'HDmalloc()' in a bunch of places.
Platforms tested:
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Bug fixes & code cleanup
Description:
Back out some of the debugging that was inadvertently checked in recently.
Platforms tested:
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Feature.
Description:
Added the short_dataset test (was in v1.6 first.)
Platforms tested:
Tested in eirene (pp) only since these have been tested in v1.6 already.
Misc. update:
Bug fix
Description:
MPIO driver did not detect partial read or write. It assumed if read
completed, all bytes requested are read. Therefore, if a dataset was
not completely written yet, the unread part will not be filled with
default fill value.
Solution:
Added the MPI_Get_elements call to find exactly how many bytes
are read (or written). For read case, if not enough is read,
the rest is padded with zeros. For write case, it is an error
if not able to write them all in first attempt. (It is not
always possible to repeat write for MPIO since file-view may
be changed, collective would hang, etc. So, we do just one attempt.)
Platforms tested:
eirene (pp). Should work for other platforms since these
have been tested in v1.6.
Misc. update:
h5repack new features
Description:
addded the copy routines for hard links
changed the traversal structure to store information about the hard links
added tests
Solution:
Platforms tested:
linux
solaris
IRIX
Misc. update:
Catch up with Unix system.
Description:
H5Tinit.c needs to be consistent with main stream platforms.
Solution:
Use typegen.exe to re-generate H5Tinit.c for windows.
Platforms tested:
windows xp, windows 2000
Misc. update:
Code cleanup for VS.NET
Description:
Clean up several issues that VS.NET was having problems with.
Platforms tested:
Eyeballed against Kent's bug reports.
h5repack new features
Description:
added debugging functions and a hard link generated dataset for the test suite
Solution:
Platforms tested:
linux
solaris 2.7
IRIX
Misc. update:
cvs checkout often included empty directories the contents of which
are no longer needed.
Added -P to cvs command to prune them.
Platforms tested:
no h5committest which does not test snaptest.
Bug fix
Description:
Attributes which were created with scalar dataspaces were reporting their
dataspace as a simple dataspace when queried later.
Solution:
Fix the dataspace handling code when reading in the attribute message from
the file to set the extent type correctly.
Platforms tested:
FreeBSD 4.9 (sleipnir)
too minor to require h5committest
Bug fix
Description:
Fix problems building on Debian Linux when Intel's "ecc" compiler is
not available.
Platforms tested:
Eyeballed - very trivial patch from user.
Bug fix.
Description:
Using a selection offset with hyperslab selections in chunked datasets
was getting into an infinite loop and hanging the application.
Solution:
Apply the selection offset to the hyperslab selection properly.
Platforms tested:
FreeBSD 4.9 (sleipnir) w & w/o parallel
h5committest
cleanup.
Description:
Removed the old command that parses setenv2 and setenv3 which are
replaced by setenvN.
Platforms tested:
No h5committest since it does not test this script.
Pretty sure it works.
Bug fix/code cleanup
Description:
Make H5Pclose_class() decrement the ID's reference count to close a generic
property class ID, which allows the new H5I routines to work correctly.
Platforms tested:
FreeBSD 4.9 (sleipnir)
h5committest
