summaryrefslogtreecommitdiffstats
path: root/pablo
diff options
context:
space:
mode:
authorRobb Matzke <matzke@llnl.gov>1998-12-14 16:19:21 (GMT)
committerRobb Matzke <matzke@llnl.gov>1998-12-14 16:19:21 (GMT)
commit28e684b0af6f181b2a1acc662d6ed2cc67ca97fb (patch)
treea5d6954568d8cca7a313dbb92c623a1cf9ec88eb /pablo
parent16b789409def33adf5a64ba0adaeaf6204a20dad (diff)
downloadhdf5-28e684b0af6f181b2a1acc662d6ed2cc67ca97fb.zip
hdf5-28e684b0af6f181b2a1acc662d6ed2cc67ca97fb.tar.gz
hdf5-28e684b0af6f181b2a1acc662d6ed2cc67ca97fb.tar.bz2
[svn-r968] Changes since 19981207
---------------------- ./MANIFEST Added tcompound2.{dmp,h5} and tdset2.{dmp,h5} used by the h5toh4 testing. ./configure.in ./configure [REGENERATED] ./src/H5config.h.in [REGENERATED] Added check for <sys/timeb.h> for Win32. Determines sizeof(char) just so we have a complete list of type sizes defined as preprocessor symbols. ./src/H5detect.c ./src/H5T.c ./src/H5Tpkg.h Detects alignment constraints. In order for this to work the operating system must not correct unaligned data (for instance, on the DEC Alpha one might need to say `uac p sigbus' before running H5detect). ./src/H5private.h Includes <signal.h> because it's needed by H5detect.c to detect alignment constraints. ./src/H5Tconv.c Added extra checks to all hardware conversion functions so they align data when necessary before the conversion. This slows down the conversions somewhat but they're still much faster than the software conversions. ./test/dtypes.c By setting a constant at the top of the source you can test conversions where the data is aligned on various byte boundaries. By setting a constant at the top of the source you can simulate architectures that have alignment constraints on architectures that don't. Changed typo SIZEOF_LDOUBLE to SIZEOF_LONG_DOUBLE in a few places.
Diffstat (limited to 'pablo')
0 files changed, 0 insertions, 0 deletions