----------------------
./src/H5.c
Fixed a signed vs. unsigned comparison.
./src/H5D.c
Setting a fill value of all zeros will cause the fill value to
be written to the dataset instead of relying on the low-level
file driver initializing unwritten areas with zero.
./src/H5D.c
./src/H5F.c
./src/H5Fprivate.h
./src/H5G.c
./src/H5Gpkg.h
./src/H5Gprivate.h
./src/H5O.c
./src/H5T.c
More file mounting stuff.
./src/H5I.c
Fixed a bug where trying to close an invalid object id caused
a core dump. For instance, H5Gclose(-1).
./MANIFEST
./test/Makefile.in
./test/mount.c [NEW]
Mounting tests.
./src/H5R.c
Fixed a couple (herr_t)NULL casts.
----------------------
./src/H5D.c
Fill values are working for contiguous datasets now except
there are two things that need more support from the data
space layer, specifically the ability to form a selection from
the difference of two selections. They are (1) extending an
external contiguous dataset, (2) optimization by delaying the
fill until after the first H5Dwrite().
Renamed H5D_allocate() to H5D_init_storage() since allocation
is only part of the story. Added a data space argument so it
doesn't have to query the space from the object header -- the
space is always available in the caller anyway.
Removed `#ifdef HAVE_PARALLEL' from a few places where it
wasn't necessary. We don't need it around code that doesn't
compile anything from mpi.h or mpio.h.
./src/H5Fistore.c
Uncommented H5F_istore_alloc() for non-parallel and moved the
`#ifdef HAVE_PARALLEL' just around Kim's barrier.
./src/H5Fmpio.c
Wrapped a couple long lines.
Got rid of two signed vs. unsigned comparison warnings.
./MANIFEST
./test/Makefile.in
./test/fillval.c [NEW]
Added tests for fill values. The contiguous dataset extend
test is disabled until H5S_SELECT_DIFF is implemented.
./tools/Makefile.in
Fixed a bug where `make test' didn't build the executables
first. This should cause the snapshots to start up again.
./Makefile.in
Changed to build in `test' directory before `tools'
directory. We want the library tests to pass before we even
start considering the tools. You can still build and/or test
the tools independent of the library tests passing.
----------------------
./MANIFEST
Changed freebsd2.2.1 to freebsd2.2.7 to match change in file name.
Added a few missing files. Are all of these supposed to be
distributed?
./tools/testfiles/h5dumptst.c
+ ./tools/testfiles/tall-1.ddl
+ ./tools/testfiles/tall-2.ddl
+ ./tools/testfiles/tall-3.ddl
./tools/testfiles/tall.ddl
./tools/testfiles/tall.h5
+ ./tools/testfiles/tattr-1.ddl
+ ./tools/testfiles/tattr-2.ddl
+ ./tools/testfiles/tattr-3.ddl
+ ./tools/testfiles/tattr-4.ddl
./tools/testfiles/tattr.ddl
./tools/testfiles/tattr.h5
+ ./tools/testfiles/tdset-1.ddl
+ ./tools/testfiles/tdset-2.ddl
+ ./tools/testfiles/tdset-3.ddl
+ ./tools/testfiles/tdset-4.ddl
./tools/testfiles/tdset.ddl
./tools/testfiles/tdset.h5
+ ./tools/testfiles/tgroup-1.ddl
+ ./tools/testfiles/tgroup-2.ddl
+ ./tools/testfiles/tgroup-3.ddl
./tools/testfiles/tgroup.ddl
./tools/testfiles/tgroup.h5
+ ./tools/testfiles/tslink-1.ddl
+ ./tools/testfiles/tslink-2.ddl
./tools/testfiles/tslink.ddl
./tools/testfiles/tslink.h5
./src/H5B.c
./src/H5Bprivate.h
./src/H5D.c
./src/H5Dprivate.h
./src/H5Farray.c
./src/H5Fistore.c
./src/H5Fprivate.h
./src/H5Gstab.c
./src/H5P.c
./src/H5Ppublic.h
./src/H5Sall.c
./src/H5Shyper.c
./src/H5Spoint.c
./src/H5Sprivate.h
./test/istore.c
Implemented split ratios as documented in previous
e-mails. Frank, the new API functions are H5Pget_btree_ratios()
and H5Pset_btree_ratios().
./src/H5Sall.c
./src/H5Shyper.c
./src/H5Spoint.c
./src/H5Sprivate.h
Added more type checking for the data transfer property list
passed to these functions.
./src/H5D.c
Added a dataset transfer property list as an argument to
H5D_allocate() for the parallel version in order to pass split
ratios down to H5F_istore_lock(). Eventually we won't need
the split ratios for H5D_allocate() because we'll build a
completely full B-tree from the leaves up, but it might be
useful to have other transfer properties at that level
anyway. I always caled H5D_allocate() with &H5D_dflt_xfer.
./MANIFEST
./test/Makefile.in
./test/overhead.c [NEW]
Added an `overhead' test run by `make timings'.
----------------------
./src/H5F.c
./src/H5Fprivate.h
./src/H5P.c
./src/H5Ppublic.h
./test/chunk.c
./test/dsets.c
The number of slots in the raw data cache can be queried or
set with H5Pget/set_cache(), which now take an extra argument.
The default number of slots is 521 and the default maximum
size is 1MB.
./src/H5Fistore.c
./src/H5Fprivate.h
Finished optimizations. The cache is now a hash and a linked
list instead of an array. The cpu time on my machine for
H5F_istore_lock() has been cut by 60% and H5F_istore_unlock() by
35%.
----------------------
./MANIFEST
./bin/release
./bin/snapshot
The format of the MANIFEST file has been relaxed -- comments
and blank lines should no longer confuse the tools that use
it.
As before, this file should contain a list of *all* files that
are part of HDF5, which means all files which are managed by
CVS/SVF. However, files can be marked so they're not part of
a release by adding _DO_NOT_DISTRIBUTE_ after the file name.
I've removed all the *.obj files from the distribution by
adding _DO_NOT_DISTRIBUTE_ after their names. I also removed
the old white-papers that might not apply to the current
version of the library and documentation about chunking
performance since it's wrong now that chunk caching is
implemented. Everyone please check the files you're
responsible for to make sure that they should really be part
of a release.
To get a list of all files use
grep '^\.' MANIFEST | cut -f1
To get a list of files for distribution use:
grep '^\.' MANIFEST | grep -v _DO_NOT_DISTRIBUTE_
(These commands assume tabs in the MANIFEST file. If the file
might contain spaces then run it through `unexpand' or `tr'
first).
./configure.in
./configure
./test/ragged.c
Check for sigaction().
./test/Makefile.in
The `clean' target removes the timing tests.
----------------------
./Makefile.in
Running `make distclean' will not fail if one of the
subdirectories has already been cleaned.
./config/BlankForm
./config/irix5.3
Cleaned it up more. Added better support/documentation for
systems that have more than one compiler.
./config/alpha-dec-osf4.0 [NEW]
Added a new config file as a result of testing on Jim Reus's
machine.
./test/chunk.c
Scaled down the testing range so we can actually run it
interactively.
./tools/h5import.c
Included <unistd.h> to get rid of warning for close().
./src/H5detect.c
Seg-faults on Linux for some reason when NDEBUG is defined, so
I just undef it at the top of the source.
./test/big.c
Added a fflush().
./tools/h5ls.c
The `-d' flag now works even when `-v' isn't specified.
./examples/h5_chunk_read.c
./examples/h5_compound.c
./examples/h5_extend_write.c
./examples/h5_group.c
./examples/h5_read.c
./examples/h5_write.c
Indented according to hdf5 standards.
Fixed compiler warnings
----------------------
./RELEASE
Updated with important changes I made since the second beta.
./src/H5A.c
./src/H5D.c
./src/H5Dprivate.h
./src/H5G.c
./src/H5Gprivate.h
./src/H5R.c
./src/H5Rprivate.h
./src/H5T.c
./src/H5Tprivate.h
Any API function that used to take an `hid_t loc_id' followed
by a `const char *name' can now take any type of object for
the loc_id as long as the object is somehow associated with a
file. Internally, H5G_loc() was modified to return an
H5G_entry_t* instead of an H5G_t* so it's more general.
Among other things, this allows one to retrieve information
about an object like a named type or dataset without knowing
the name of the type or dataset:
int
get_nlinks (hid_t obj)
{
H5G_stat_t sb;
if (H5Gstat(obj, ".", TRUE, &sb)<0) return -1;
return sb.nlink;
}
./test/gheap.c
./test/istore.c
These files needed a couple of changes because they call some
of the internal functions whose H5G_t arguments changed to
H5G_entry_t.
./src/H5A.c
Got rid of all the switch statements for getting symbol table
entries for varous objects and replaced them with a call to
H5G_loc() allowing attributes to automatically apply to any
type of object that belongs to a file.
./test/Makefile.in
Moved the ragged array tests from the normal list of tests to
the `make timings' target.
./test/ragged.c
Added rewrite tests -- rewrite the rows of a dataset changing
the number of rows and the length of each row.
./test/mtime.c
Added a test that checks that H5Gstat() can be called with a
dataset as the first argument.
./src/H5S.c
Added #ifdef HAVE_PARALLEL around code to check for the
HDF5_MPI_OPT_TYPES environment variable because the global
variable that gets set is #ifdef'd.
./bin/release
bzip2 uses .bz2 as the file extension.
----------------------
./src/H5G.c
Fixed a link count on groups. Groups were always created with
a link count of zero instead of one.
./src/H5Gpublic.h
Added H5G_NTYPES so applications can easily declare arrays
which are indexed by object type.
./src/H5O.c
Calling `H5O_link(&ent,0)' will succeed when the file is open
for read only because it's the only way to get the object's
current link count.
./tools/h5ls.c
Can take non-group object names on the command line (and
multiple objects). Reorganized to be more extendible. The
link count is displayed for `-v'.
./src/H5F.c
Dumping of cache statistics is controlled by only the
HDF5_DEBUG environment variable and not the H5F_ACC_DEBUG flag
for H5Fopen() and H5Fcreate() since this makes it similar to
the other debugging options.
./src/H5R.c
./test/ragged.c
./test/Makefile.in
Tests pass but still a little work left to be done.
Fixed errors
H5.c
Fixed core dump with setvbuf() for debugging.
H5Fistore.c
Fixed bug with sparse datasets.
H5Omtime.c
Found a typo in "Posix Programmer's Guide". The `%z' in the
strftime() should be `%Z' instead.
H5T.c
H5Tpublic.h
Improved api tracing for H5Tinsert_array().
h5tools.c
Doesn't dump core for empty datasets.
----------------------
./src/H5Fistore.c
Fixed a bug that caused the wrong number of chunk offsets to
be read or written from the file.
./src/H5Z.c
Fixed a memory bug.
./test/dtypes.c
Removed redundant tests.
./bin/snapshot
Removed Perl code from a shell script. Oops :-)
./src/H5D.c
Removed cvs merge conflicts.
./src/H5Omtime.c
Fixed a typo in the Irix5.3 code.
./src/H5S.c
Fixed a memory leak.
./src/H5Shyper.c
Fixed a bad call to malloc() that didn't allocate enough
memory. Reordered a comparison to eliminate an array bounds
read error.
----------------------
./MANIFEST
Replaced Compression.html with Filters.html
./doc/html/Filters.html
./src/H5Ocomp.c
./src/H5P.c
./src/H5Ppublic.h
./src/H5Z.c
./src/H5Zprivate.h
Added two extra arguments to H5Pget_filter() in order to
retrieve the filter name. The name is the name registered for
the filter with H5Zregister(), but if the dataset creation
property originally came from an existing file then the name
is that which is stored in the file.
./tools/h5ls.c
The `-v' option now prints the names of the filters.
./src/H5B.c
./src/H5Fistore.c
./src/H5O.c
./src/H5Oefl.c
./src/H5Oprivate.h
./src/H5P.c
Plugged a memory leak.
./src/H5MMprivate.h
H5MM_malloc(0) and H5MM_calloc(0) actually allocate a single
byte in order to be sure that we get a valid pointer.
./src/H5S.c
./src/H5Sselect.c
Fixed pointer->integer conversions in error return values in
three places.
Fixed a bug that caused `h5vers -i' to fail with a usage
message. The new behavior is the same as `h5vers -i release'
which increments the release number.
./bin/snapshot
A shell script intended to be run nightly from cron which
compares the current version of hdf5 against the previous
version. If there are differences it creates a tar file in an
ftp directory, increments the hdf5 release number, and commits
the changes back into the cvs repository. The name of the ftp
directory in which to place the release can be given on the
command line (default /hdf3/ftp/pub/outgoing/hdf5/snapshots).
The CVSROOT environment variable must be set.
----------------------
./MANIFEST
./doc/html/H5.format.html
./src/H5O.c
./src/H5Oprivate.h
./src/H5Omtime.c [NEW]
./src/H5private.h
./src/Makefile.in
Added the modification time message. If an object header has
this message then it's value is updated with the current time
whenever anything changes in the object header.
./acconfig.h
./configure.in
Alas, there seems to be no standard way to convert a string
time like 19980727122800 in UTC to a time_t since mktime()
only converts local times to time_t. So I've modified the
configuration to check for various ways of getting the time
zone information:
* Added checks for the `tm_gmtoff' field of `struct tm'.
* Added a check for the `timezone' global variable.
* Added a check for `struct timezone'.
* Added a check for BSDgettimeofday().
* Added a check for gettimeofday() although it doesn't
actually set the timezone argument on some systems.
* Added a check to see if `tm_zone' is a member of `struct tm'.
* Added a check to see if `tzname' is a global variable.
* Added a check to see if `struct tm' is defined in time.h
or sys/time.h.
It's not difficult to get the right UTC modification message
into the object header, but some systems might have problems
getting the right time back out (Irix64 is one) and those
systems will report zero for the H5G_stat_t.mtime from an
H5Gstat() call as if the mtime message isn't even present. It
will, however, continue to be updated as normal.
./src/H5G.c
./src/H5Gpublic.h
The H5G_stat_t struct now contains an `mtime' field which will
hold the object modification time. If the object has no
object modification time then the `mtime' will be initialized
to zero.
Fixed a bug in H5G_stat() that caused the `objno' field of the
H5G_stat_t to be set incorrectly on some machines.
./src/H5D.c
Writing to external datasets fail if the hdf5 file is not open
for writing.
A modification time message is added to the dataset object
header when it's created and H5O_touch() is called from
H5D_write() to update that message.
./src/H5T.c
Fixed a bug in H5Tget_member_dims() that caused a segmentation
fault if one of the output array arguments was the null
pointer.
Relaxed the member dimension checking in H5Tinsert_array() so
it can also be used for scalar members.
./test/Makefile.in
Added additional file names to the `mostlyclean' target.
./tools/h5dump.c
./tools/h5tools.h
Increased the temporary buffer size to 1MB.
Added support for printing compound data types with array
members.
When printing H5T_NATIVE_CHAR and H5T_NATIVE_UCHAR we escape
double quote characters.
./tools/h5ls.c
Changed the output format a little because we were starting to
get too much info to fit on a line. Without `--verbose' each
object occupies one line of output. Otherwise, additional
information is printed below the object name: object file
address, comment, and modification time. If `--dump' is given
then the data is printed after the other information.
./test/cmpd_dset.c
Changed the way the dataset is initialized to be more uniform.
----------------------
./tools/h5dump.c
./tools/h5ls.c
./tools/h5tools.h
Able to handle compound data types now. Moved code that
chooses memory data type based on file data type from h5ls to
libh5tools.a(h5dump.o)
----------------------
./src/H5.c
Handle hid_t of type H5_TEMPBUF, arguments usually called tbuf_id.
Added array tracing where the array rank is stored in a simple
data space. Just use the name of the data space argument when
declaring the array argument:
herr_t
H5Sselect_hyperslab (hid_t space_id, H5S_seloper_t op,
const hssize_t start[/*space_id*/],
const hsize_t _stride[/*space_id*/],
const hsize_t count[/*space_id*/],
const hsize_t _block[/*space_id*/])
and when the program runs you'll see array values printed:
H5Sselect_hyperslab(space=218103813, op=H5S_SELECT_SET,
start=0xbfffef4c {0}, _stride=NULL,
count=0xbfffef44 {64},
_block=NULL) = SUCCEED;
Added more symbolic data types to the tracing output.
./src/H5A.c
./src/H5Apublic.h
./src/H5D.c
./src/H5Dpublic.h
./src/H5F.c
./src/H5Fpublic.h
./src/H5G.c
./src/H5Gpublic.h
./src/H5P.c
./src/H5Ppublic.h
./src/H5S.c
./src/H5Sall.c
./src/H5Shyper.c
./src/H5Spoint.c
./src/H5Spublic.h
./src/H5Sselect.c
./src/H5Ssimp.c
./src/H5TB.c
./src/H5V.c
Changed some API argument names to be more consistent with
other API functions and to produce better tracing output.
Reformatted some long lines. Indented printf statements.
./tools/h5ls.c
Fixed warnings about unsigned vs. signed comparisons.
----------------------
./tools/h5ls.c
If the dataset is of type H5T_NATIVE_CHAR then we print the
value as a string. This is temporary -- I plan to add better
control of this later but needed something now for debugging.
./src/H5Fistore.c
Squashed a bug in the chunk caching code that caused the wrong
chunk to be returned.
./bin/trace
./src/H5.c
Added support for printing values of array arguments when the
size of the array is supplied by some previous argument. You
must declare the argument as an array in order for the
automatic tracing stuff to work. For instance, the third
argument of H5Pset_chunk() is an array whose size is
determined by the second argument `ndims'. Here's how you
should declare it:
herr_t
H5Pset_chunk(hid_t plist_id, intn rank, hsize_t dims[/*rank*/])
The comment inside the `[]' is the name of some previous
integer argument (int, unsigned, size_t, ssize_t, hsize_t,
hssize_t). The trace output will look something like:
H5Pset_chunk(plist=1234567, rank=2, dims=0x112233 {11, 22})
Changed tracing so that data types are printed out
symbolically when possible.
Changed tracing so data type initializations are not printed.
This used to be confusing because lots of H5Tcopy() and
H5Tregister...() calls showed up before the applications first
explicit API call.
./src/H5Ipublic.h
Changed the file atom group from zero to one so printing of
atoms during tracing is more consistent -- they're all big
numbers now.
./src/H5A.c
./src/H5E.c
./src/H5F.c
./src/H5G.c
./src/H5Sselect.c
./src/H5T.c
./src/H5TB.c
./src/H5Z.c
Accidently modified these when working on the tracing, but
nothing should have changed.
./src/H5P.c
Changed the definition of H5Pset_chunk() for tracing.
./src/H5S.c
./src/H5Spublic.h
Changed the definitions of H5Sset_extent_simple() and
H5Screate_simple() for tracing. Changed the FUNC_ENTER() name
for H5Screate_simple() so tracing shows the correct name.
----------------------
./bin/release
The archive names have only three parts to a version number.
Instead of `hdf5-1.2.3d.tar.gz' they will be named like
`hdf5-1.2.3.tar.gz'.
./src/H5public.h
./src/H5.c
Dropped the patch level from the version number. The version
number is now a major number, a minor number, and a release
number.
./doc/html/Version.html
Updated to get rid of patch numbers.
./tools/h5repart.c
Added a `-V' flag like for `h5ls' that prints the version number.
----------------------
./doc/html/H5.format.html
./src/H5F.c
./src/H5Gprivate.h
./src/H5Oshared.c
Added file alignment fields: the boot block has an extra
reserved address field. The symbol table entry scratch pad
space was reduced from 24 bytes to 16 bytes. The index
permutation was moved earlier in the data type message for
compound data types and extra padding was added. Four bytes
of padding was added to the shared message format.
----------------------
./doc/html/H5.format.html
./src/H5Gent.c
./src/H5Gprivate.h
./src/H5Oattr.c
./src/H5Oprivate.h
./src/H5Oshared.c
./src/H5HG.c
./src/H5HGprivate.h
Added padding fields in symbol table entries, attribute
messages, shared messages, and global heap objects to insure
that things are aligned on 8-byte boundaries in the file, and
thus in memory. Otherwise some little endian machines
complain (DEC Alpha) during encoding/decoding of file meta
data. I chose to add alignment to the file rather than
rewriting the ENCODE/DECODE macros for the little endian case.
Completely rewrote the section on attribute messages.
More alignment stuff will follow.
./src/H5detect.c
Fixed a typo `nd'->`dn'
./test/dtypes.c
Commented out conversion tests to/from `long double' on
machines where it's the same size as `double' to get rid of
compiler warnings.
./doc/html/Big.html
Fixed a couple typos.
