Moved the source of H5S_extent_npoints before H5Sextent_npoints to
reduce a compiler complaint.
H5Shyper.c:
H5Sprivate.h:
H5Spublic.h:
H5Sselect.c:
Changed the parameter declarations of H5S_hyper_add
and H5Sselect_hyperslab as originally intended as const hsize *.
The purpose of the change is to correct the H5 filenames declared in
testphdf5.c when used on the Intel TFLOP PFS system.
Problem:
The ROMIO library complains of the H5 filenames having no "piofs:",
"pfs:", "ufs:", or "nfs:" prefixes on the filenames. The only type
currently recognized are the "pfs:" prefixes, and the files must
reside in the Intel TFLOP PFS system: /pfs.
Solution: (REQUIRED)
The compile flag -DPOOMA_ARCH=TFLOP is added to the list of CFLAGS in
Makefile.ascired located in the parallel test directory. This compile
flag will allow a special set of filenames to be used when compiling
testphdf5.c on the Intel TFLOP. In addition, instructions to editing
the filenames in testphdf5.c to reside in the user's own /pfs directory
are included in INSTALL.ascired.
Platform tested:
Intel TFLOP
----------------------
./bin/trace
./src/H5.c
./src/H5private.h
./src/H5A.c
./src/H5D.c
./src/H5F.c
./src/H5G.c
./src/H5P.c
./src/H5Pprivate.h
./src/H5S.c
./src/H5T.c
Output-only arguments have their addresses printed during
tracing and added symbolic output for the H5F_driver_t
arguments. That's another reason that we should be careful to
add `/*out*/' after arguments that are output-only and
`/*in,out*/' after arguments that are used for both input and
output values.
No internal function calls H5Pget_class() anymore.
./src/H5T.c
./src/H5Tconv.c
./src/H5Tpkg.h
./src/H5Tpublic.h
Added H5Tget_overflow() and H5Tset_overflow() so the
application can query or set a function that will be called
whenever an overflow occurs. Implemented as documented in
previous e-mail except the overflow handler gets two buffers:
one that contains the source value and one to receive the
optional destination value.
./test/dtypes.c
Tests overflow handler.
./src/H5.c
We have to declare fdopen() because I'm getting errors when
compiling on Irix64 even though we include <stdio.h> as
documented in the fdopen() man page.
----------------------
./src/H5public.h
If sizeof(size_t)==sizeof(long long) then we define `hsize_t'
in terms of size_t to get rid of warnings on systems that
treat `long long' the same as `long'.
----------------------
./src/H5T.c
Prints statistics for no-op conversions also.
For now, hardware floating point conversions may or may not
raise SIGFPE depending on the hardware, operating system,
etc. Software conversions never raise SIGFPE and use +Inf or
-Inf for overflow.
./test/dtypes.c
Catches SIGFPE and causes the test to be skipped. Better test
for NaN.
./config/irix5.3
Removed the -U_POSIX_SOURCE because it was removed from the
main makefiles.
./bin/trace
./src/H5S.c
Fixed a typo that prevented tracing info from being added to
new API functions.
API defined in the html/Dataspaces.html document. This code does not include
support for strides, merging selections, or permutations of coordinates yet,
but it's a drop-in replacement for the existing API with the same features.
----------------------
./MANIFEST
Added new files from testpar directory.
./config/commence.in
Removed -D_POSIX_SOURCE because hdf5 uses a couple of
non-Posix things when certain debugging flags are switched on.
to cleanup temporary files used by that particular module.
Added a cleanup() routine to testhdf5 that calls all the other
cleanup_xxx() routines to do cleanup.
Cleanup action can be disabled either by -c (no cleanup) option
or HDF5_NOCLEANUP environment variable.
----------------------
./MANIFEST
./configure.in
./examples/Makefile.in [NEW]
Added a Makefile.in for the examples directory. To generate
the Makefile run configure again.
changed to use system supported MPI library. The path to it is
changed slightly too. Updated all the paths variables.
Also updated to reflect the new program source files structure.
Platform tested: Intel Red
big testphdf5.c file. Broke it into modules of related routines.
testphdf5.c -- main routine and global variables initialization plus
some ulitility routines.
t_file.c -- tests of file operations using parallel I/O.
t_dset.c -- tests of datasets operations.
H5D.c:
If a dataset, using chunk storage, is created/opened and if its data
file is accessed via MPIO, allocate file space for the chunks immediately.
Also do the allocation when a dataset is extended. Must do it at these
points because H5Dcreate, H5Dopen and H5Dextend are collective calls
in MPIO access mode and are the only chance to get all MPI processes
to synchonize their meta-data.
Also disallow opening/creating datasets with compression option if
MPIO is used. Not easy to support concurrent compression writes in
an efficient way. But read-only access should be fine. Will implement
that later.
H5F.c:
Disable file cache if MPIO is used by forcing cache size and number
of elements to zero.
H5Fistore.c:
Added two new routines, H5F_istore_get_addr and H5F_istore_allocate.
H5F_istore_allocate supports allocating file space for the chunks
of a dataset. H5F_istore_get_addr returns the file space address
of a chunk that has been allocated.
Changed H5F_istore_read and H5F_istore_write to bypass the chunk
cache read/write mechanism. Instead, they do read/write directly
with the file for only the elements requested.
(Some debug statements, enclosed by #ifdef AKC, are still in the
file. I am still working on some code. They would be cleaned
out soon.)
----------------------
./src/H5Tbit.c
./src/H5Tpkg.h
Fixed a bug in H5T_bit_copy(). Added H5T_bit_get_d() and
H5T_bit_set_d() which treat portions of a bit vector as an
unsigned integer. Added H5T_bit_inc() that increments part of
a bit vector and returns an indication of overflow.
./src/H5Tconv.c
./src/H5Tpkg.h
./test/dtypes.c
Added a slow general floating point conversion which works so
far on Intel, MIPS, and DEC but the test is turned off because
a cast from double to float will cause a SIGFPE on some
systems if an overflow occurs.
Added fast hardware conversions between native floating point
types. This function is also subject to the SIGFPE problem.
./src/H5detect.c
Removed the exponent bias adjustment when the significand
isn't normalized. This is now handled in the conversion
functions instead.
./src/H5T.c
Register new conversion functions.
Plugged a memory leak in the library termination code.
./RELEASE
Added a list of major changes since the first alpha.
./src/H5.c
./src/H5private.h
./src/H5A.c
./src/H5AC.c
./src/H5D.c
./src/H5Ffamily.c
./src/H5Fistore.c
./src/H5G.c
./src/H5Gprivate.h
./src/H5HG.c
./src/H5O.c
./src/H5T.c
./src/H5Tbit.c
./src/H5Tconv.c
Fixed various compiler warnings on Irix64.
./src/H5MM.c
Added PABLO_MASK to this file.
./test/chunk.c
Removed a warning about memcpy().
./INSTALL_MAINT
./README
./RELEASE
Partially updated for second alpha, but haven't updated
version numbers yet.
./src/H5.c
./src/H5A.c
./src/H5AC.c
./src/H5B.c
./src/H5D.c
./src/H5F.c
./src/H5Fcore.c
./src/H5Ffamily.c
./src/H5Fistore.c
./src/H5Fmpio.c
./src/H5Fsec2.c
./src/H5Fsplit.c
./src/H5Fstdio.c
./src/H5G.c
./src/H5Gnode.c
./src/H5HG.c
./src/H5HL.c
./src/H5I.c
./src/H5MM.c
./src/H5MMprivate.h
./src/H5O.c
./src/H5Oattr.c
./src/H5Ocomp.c
./src/H5Ocont.c
./src/H5Odtype.c
./src/H5Oefl.c
./src/H5Olayout.c
./src/H5Oname.c
./src/H5Osdspace.c
./src/H5Oshared.c
./src/H5Ostab.c
./src/H5P.c
./src/H5S.c
./src/H5T.c
./src/H5Tconv.c
./src/H5detect.c
./test/hyperslab.c
./test/istore.c
Changed memory allocation functions so they fail instead of
dumping core. The `x' was removed from the name to remind us
of that: H5MM_xmalloc() -> H5MM_malloc(), etc.
H5MM_calloc() takes one argument like H5MM_malloc() instead of
two like calloc() because we almost always called it with `1'
for one of the arguments anyway. The only difference between
the two functions is that H5MM_calloc() returns memory which
is initialized to zero.
./src/H5Gent.c
./src/H5Gprivate.h
Removed H5G_ent_calloc() since it wasn't used.
./src/H5Fistore.c
Fixed a bug found by Albert. Thanks, Albert! This fix
combined with the changes to memory allocation prevent the
library from failing an assertion if the application uses an
unreasonable size for chunks (like Alberts 10000x10000x4).
./src/H5MF.c
./src/H5MFprivate.h
Changed H5MF_free() to H5MF_xfree() since calling it with an
undefined address is allowed.
----------------------
./html/Datatypes.html
./src/H5T.c
./src/H5Tconv.c
./src/H5Tpkg.h
./src/H5Tpublic.h
./src/H5detect.c
./test/cmpd_dset.c
./test/dsets.c
./test/dtypes.c
./test/external.c
./test/tattr.c
./test/th5s.c
Organized the predefined data types.
./html/H5.intro.html
NO MODIFICATIONS! This file is a mess (obviously generated
from something else) so I skipped updating it!
./src/h5ls.c
./test/dsets.c
./test/tattr.c
./test/th5s.c
Added extra argument to H5Sget_dims() calls per Albert's last
update.
Added maxdims agrument to H5Sget_dims. Also removed the check
of dims since H5Sget_dims can still return ndims even if both
*dims and *maxdims are NULLs.
----------------------
./src/H5T.c
Fixed a potential division by zero error.
./src/H5Fistore.c
Fixed an assertion failure that I accidently introduced with
the previous changes.
----------------------
./html/tracing.html NEW
This entire update is to make it possible for the library to
print the name, arguments, and return value of every API call
without requiring any extra work from developers or app
programmers. This file describes how this all works.
./configure.in
Added the `--enable-tracing' switch. If you use it then the
library will include code to print API function names,
argument names and values, and function return values.
However, you must then turn on the tracing by setting the
HDF5_TRACE environment variable to a file descriptor number.
The default is `--disable-tracing' since enabling it causes a
slight increase in library size and a slowdown resulting from
an extra function call for each API function call (I couldn't
even measure the slowdown :-)
./bin/trace NEW
A perl script that synchronizes the H5TRACE() macro calls in
the *.c files with the function return type and formal
argument names and types. If you use GNU make and gcc then
this will be done automatically, otherwise just invoke this
script with the names of one or more .c files. You could do
it by hand to, but encoding argument types is a little tricky
at first.
./config/commence.in
Added the $(TRACE) macro, which defaults to the no-op.
Added -D_POSIX_SOURCE to the compiler command line.
./src/Makefile.in
Override the default for $(TRACE).
./config/depend.in
Automatically calls $(TRACE) to synchronize the H5TRACE()
macros in any source file that changed. As with makefile
dependencies, one way to force synchronization of all files is
to remove the `.depend' file.
./MANIFEST
Added new files.
./src/H5Eprivate.h
Modified HRETURN_ERROR() and HRETURN() for tracing.
./src/H5.c
./src/H5private.h
This is where the real tracing work really happens, in
H5_trace().
./src/H5A.c
./src/H5D.c
./src/H5G.c
./src/H5P.c
./src/H5S.c
./src/H5Z.c
Added H5TRACE() calls to all API functions. You don't really
need these changes if you don't want to merge your stuff
because they can be generated automatically by going to the
hdf5/src directory and saying ../bin/trace *.c
./src/H5T.c
Added H5TRACE() calls. Other stuff below.
./src/H5E.c
./src/H5Epublic.h
Added H5TRACE() calls. Created a type H5E_auto_t for the
`func' argument of H5Eset_auto() and H5Eget_auto() to make
those arguments easier to parse for tracing. It should also
make it clearer for users that don't know how to read
complicated ANSI data types.
./src/H5F.c
Added H5TRACE() calls. Changed a couple `uintn' argument
types in API functions to `unsigned int' since `uintn' part of
the API. Changed a few "can't" and "cant" error messages to
"unable to".
./src/H5Ipublic.h
Removed H5_DIRECTORY from the H5I_group_t enum. It wasn't
used anywhere.
./src/H5Tconv.c
Removed an unused label.
./src/H5Fistore.c
./src/H5Oattr.c
./src/H5Odtype.c
./src/H5T.c
./test/dsets.c
./test/dtypes.c
Fixed a warning about a variable possibly used before it's
initialized. Added __unused__ to turn off some unused
argument warnings that pop up when debugging is turned off and
optimizations are turned on.
----------------------
./src/H5Tbit.c
./MANIFEST
./test/Makefile.in
./test/bittests.c NEW
Finished the bit vector operations and added test cases.
./src/H5Tconv.c
./test/dtypes.c
Finished integer->integer general conversion and added test
cases. Overflows and underflows are handled by substituting
the closest possible value. Examples:
(unsigned)0xffff -> (unsigned) 0xff
( signed)0xffff -> (unsigned)0x0000
(unsigned)0xffff -> ( signed)0x7fff
( signed)0x7fff -> ( signed) 0x7f
( signed)0xbfff -> ( signed) 0xbf
( signed)0x8000 -> ( signed) 0x80
./src/H5private.h
Added definitions for MIN and MAX that take 3 or 4 arguments:
MIN3(), MIN4(), MAX3(), MAX4(). Also added MIN2() and MAX2()
as aliases for MIN() and MAX().
./test/tattr.c
Removed some redundant `&' operators.
./configure.in
./src/H5config.h.in [regenerated]
./src/H5.c
Fixed warnings on DEC where long double is the same as
double.
----------------------
THIS CHECKIN IS FOR QUINCEY -- NOT EVERYTHING WORKS (but it compiles)
MOST OF THE CHANGES ARE FOR BETTER TYPE CONVERSION IN THE NEXT ALPHA
./MANIFEST
./src/H5Tbit.c NEW
./src/Makefile.in
Bit vector operations (not done yet)
./configure.in
Added -lm to the library list, needed by bit-vector operations
and conversion functions.
Removed vestiges of PARALLEL_SRC no longer used by the
makefiles. Albert came up with a better way (that actually
works :-)
./src/H5D.c
No code changes. Split a couple of long lines, refilled a
couple multi-line comments.
./src/H5T.c
./src/H5Tpublic.h
Fixed a bug reported by Jim Reus regarding conversion of
compound data types whose members require conversions which
are satisfied by as-yet unregistered soft conversion
functions.
Added H5T_IEEE architecture, but the funny-looking integer
types will be changed to H5T_BE_ and H5T_LE_ architectures
with the type names changed to match the H5T_NATIVE_ integers.
Added an H5Tconvert() but it hasn't been documented or tested
yet.
./src/H5Tconv.c
./src/H5Tpkg.h
Registered conversion functions integer->integer (a general
case) and integer->float (for a specific case). The
integer->integer conversion depends on the bitvector
operations which aren't finished yet and the int->float
conversion hasn't been retested since it was borrowed from
AIO. Don't look at them yet, they're ugly :-)
./src/H5detect.c
Fixed a typo which caused the msb_pad field of an atomic type
to not be initialized.
./test/dtypes.c
Added a test for number conversions but it's commented out
until the conversion stuff is truly working.
which can be set by passing to H5Fcreate and H5Fopen
a string of characters (one character for each flag to turn on)
as the value of MPI_Info with key H5F_MPIO_DEBUG_KEY.
