Purpose:
a windows bug fix
Description:
support > 2G file on windows NTFS
Solution:
using fsetpos to seek the address
Platforms tested:
windows 2000, confirmed on arabica and eirene.
Bug Fix
Description:
If the library was compiled with 'NDEBUG' defined, but the user's program
(the fortran API in this case) was compiled with 'NDEBUG' not defined, the
IDs returned from the library would appear to be incorrect.
Solution:
Removed 'ifndef NDEBUG' from around "H5I_TEMPLAX_MAX" definition and error
checking.
Platforms tested:
Linux 2.2 (eirene) & FreeBSD 4.2 (hawkwind)
Prepare for 1.4.0beta1 release
README:
src/H5public.h:
Updated with 1.4.0beta1 version information.
configure.in:
Changed default setting to --enable-production --disable-debug.
configure:
Derived from configure.in via autoconf.
Platforms tested:
Eirene (linux) running just the configure command and compared
difference with previous configure output to verify expected
difference.
Bug Fix
Description:
More descrimination is necessary for the type of background buffer needed.
Compound types need H5T_BKG_TEMP and vlen & array types need H5T_BKG_YES.
Solution:
Make the internal routine be more specific about the background type set.
Platforms tested:
FreeBSD 4.2 (hawkwind)
Bug fix
Description:
Non-optimized conversions have a memory overwrite bug when the destination
size of a compound datatype is greater than the source size.
Solution:
Corrected direction of walking through the destination array for final
copying.
Platforms tested:
FreeBSD 4.2 (hawkwind) & Cray J90 (killeen)
Bug fix
Description:
In H5FD_mpio_open, there was a mem-xfree of the file variable
before it ever gets alloc'ed. This is inside a failure condition
that rarely happened. But if it happens, it would be freeing
a null pointer.
Solution:
Removed the error code.
Platforms tested:
Modi4 -64 parallel. (Everything still works. :-)
Bug Fix
Description:
Compound datatypes weren't reading background data properly when used
with array or vlen fields.
Solution:
Changed to always read background information from file in those situations.
Platforms tested:
FreeBSD 4.2 (hawkwind)
Feature
Description:
H5Eprint used to print "thread 0" even in MPI-application.
Added code to print MPI-process rank ID (as in MPI_COMM_WORLD) to
give more information of which process is printing the messages.
Platforms tested:
IRIX64-64, both serial and parallel.
Bug fix.
Description:
Cray T3E (and J90?) was failing on partial I/O tests. The bug was in
the optimized hyperslab routines for cases where a hyperslab filled the
temporary datatype conversion buffer and the optimized hyperslab routine
needed to start in the middle of a hyperslab.
Solution:
Corrected code to detect the situation and read in the correct amount of
elements in the hyperslab.
Platforms tested:
Cray T3E (mcurie.nersc.gov)
Purpose:
Bugfix
Description:
Under SunOS 5.5 the symbol FIONBIO wasn't known.
Solution:
Include <sys/filio.h> which defines this symbol under Solaris.
Also put the UNUSED qualifier in the right place in a function
argument list (gcc 2.7.2 didn't like it before the type name).
Platforms tested:
SunOS 5.5 (hatteras), SunOS 5.6 t(thor.sistec.kp.dlr.de)
Bug fix
Description:
Compiler on Cray T3E has a weird problem with operations on size_t
variables.
Solution:
Change comparison to make Cray T3E happy.
Platforms tested:
Cray T3E (mcurie)
Bug fix
Description:
"UNUSED" macro was in wrong place for older versions of gcc
Solution:
Moved the "UNUSED" macro to the right of the H5FD_t type declaration and
the compiler was happy...
Platforms tested:
Solaris X86 2.5 (hatteras)
Purpose:
fix a bug
Description:
add <windows.h> for WIN32
Solution:
Add the following lines at the beginning of the file
#ifdef WIN32
#include <windows.h>
#endif
Platforms tested:
NT4.0,5.0, and LINUX
Code addition
Description:
The dumper needs to know when a datatype needs to be reclaimed, so I
added a small helper function to detect if a particular datatype is or
contains a particular class of datatypes.
Platforms tested:
Linux 2.2.16-3smp (eirene)
Patch on the patch of my code cleanup... :-)
Description:
Fix compiler problems with previous code patching...
Platforms tested:
Linux 2.2.16-3smp (eirene)
Bug fixes
Description:
Patch things back up from incompletely changing some variables in
yesterday's code cleanup, this should fix the compiler errors that Kent
and Bill have found in the CVS version of the code.
Also, fix a bug buried in the datatype conversions when compound datasets
were nested inside an array or variable-length datatype. Internally
require that any datatype which contains a compound datatype sets the
"need_bkg" conversion flag so a background buffer to pass to the compound
datatype conversion is allocated.
Platforms tested:
FreeBSD 4.2 (hawkwind) & Solaris 2.6 (baldric)
Bug fix
Description:
When the v1.2 compatibility code was turned on, internal functions in the
library were getting confused.
Solution:
Separated guts of H5Pget_driver call into an API function (the definition
of which depends on the compatibility switch) and an internal function
which always behaves like the v1.3/4 function. Replaced API function calls
in the library code with the internal function.
Platforms tested:
FreeBSD 4.2 (hawkwind)
Code cleanup
Description:
Several places in the code were using -2 as a default value for various
features. However, when a default value is returned from a function that
is supposed to return negative on failure, it was causing confusion and
extra work for users to check for the special value.
Solution:
Replaced hard-coded -2 values in the code with symbolic names, then changed
symbolic names to 0 instead of -2.
Platforms tested:
FreeBSD 4.2 (hawkwind)
Bug fix
Description:
Fixed some typo that cause compiling errors in parallel mode.
Also added a default case in the H5Pget_xfer code to catch
potential future added data transfer mode.
Platforms tested:
IRIX64-64, parallel.
Backward compatibility code
Description:
Add in code to allow the library to emulate the v1.2 API and behavior.
Platforms tested:
FreeBSD 4.2 (hawkwind)
Bug fix
Description:
The optimized MPI-IO calls, H5S_mpio_spaces_write/H5S_mpio_spaces_read,
are changed for collective data transfer only since they
call H5FD_mpio_setup to do setup to eveually call MPI_File_set_view
in H5FD_mpio_read or H5FD_mpio_write. MPI_File_set_view is a
collective call. Letting independent data transfer use this
route would result in hanging.
Solution:
For now, the checking is being done in
H5D_write and H5D_read before H5S_mpio_spaces_write/H5S_mpio_spaces_read
is called because the checking code in H5S_mpio_spaces_xfer, though with
the right idea, is not correct yet.
Platforms tested:
IRIX64-64 parallel.
Purpose:
Bug fix
Description:
H5S_mpio_spaces_xfer() did not do clean up all the times especially
the MPI_type freeing.
Solution:
Include the MPI_type freeing as part of cleanup code. Replaced
most HRETURN_ERROR(...) with HGOTO_ERROR so that cleanup code
will be executed.
Platforms tested:
IRIX64-64.
Purpose:
make mtime test available on windows
Description:
calculating time zone difference on windows
Solution:
1.Use /#if WIN32 #endif/ to make sure the modification will not affect other platforms.
2. use <sys/types.h> and <sys/timeb.h> on NT to obtain time zone difference.
Platforms tested:
windows 2000
Bug fix (feature, kind of)
Description:
The library used to hang if a collective dataset read/write request
does not have the same number of eventual MPIO request. Part of the
reason is that H5FD_read/H5FD_write immediately returns succeess
if it sees the request size is 0. This caused problem since other
processes with I/O to do would be hanging by waiting for the early
returned process(es).
Solution:
H5FD.c: disable the early return code in parallel mode. Make it
go on even with "nothing" to transfer.
H5D.c: the optimized MPIO xfer routines can handle collect calls correctly
when the condition is right (e.g., no conversion). When the COLLECTIVE
request cannot be handled correctly without the risk of hanging, the
COLLECTIVE is changed to INDEPENDENT calls for the eventual MPIO
calls.
Platforms tested:
IRIX64 parallel (-64, n32), IRIX64 -64 sequential, Linux sequential.
Small code optimization
Description:
Statistics on the amount of times a datatype conversion required alignment
were being kept in all case, even though they are only used when H5T_DEBUG
is turned on.
Solution:
Build some extra macros so that the statistics are only kept when the
H5T_DEBUG macro is defined.
Platforms tested:
Solaris 2.6 (baldric)
Bug fix
Description:
Previous re-arrangement of FUNC_ENTER macro was a _little_ too agressive
about not re-checking for library or interface re-initialization. Although
it happens very rarely that the library is shut down and then re-started
the library & interface initialization functions weren't getting run in
that case.
Solution:
Separated out the library, interface & function checks again. This version
is still a little faster than the previous way things were being done.
Platforms tested:
FreeBSD 4.1.1 (hawkwind)
Code optimization
Description:
Re-vamped FUNC_ENTER macro to remove as many of the if's as possible and
also to only check once if any given function is an API function.
This improves the performance of the hyperslab I/O benchmark (h5hypers)
that I've been testing with by another 5%. All library functions should
be 5-15% faster, depending on how many times they are called and what
percentage of the function's time was spent in the FUNC_ENTER macro vs.
the percentage of time in the main body of the function.
Platforms tested:
Solaris 2.6 (baldric)
Bug fix
Description:
When performing I/O on the entire dataset for datasets with the maximum
number of dimensions, H5S_MAX_RANK, (32 currently), the dimension
information was getting corrupted internally and causing no data to be
actually written to the file.
Solution:
Changed the internal array sizes from H5S_MAX_RANK to H5O_LAYOUT_NDIMS,
which is one element larger to accomodate the size of the elements
themselves.
Platforms tested:
Solaris 2.7 (arabica)
Bug fix
Description:
The optimized MPIO code was not working with the TFLOPS machine.
It kept reporting "can't convert from size to size_i" errors in
H5FD_mpio_write() and H5FD_mpio_read().
Solution:
The error was actually in H5S_mpio_spaces_xfer() in H5Smpio.c
in which the variable mpi_count was declared as size_t (only
4 bytes big in TFLOPS) when it should be hsize_t (8 bytes big in TFLOPS).
Corrected and also changed to use the (hsize_t) to case size_i in
H5FD_mpio_write() and H5FD_mpio_read() since that is what it should
be.
Platforms tested:
TFLOPS and modi4 (-64, parallel)
Buglet fix
Description:
Get rid of compiler warnings. The HGOTO_RETURN was returning NULL
when it should have returned FAIL.
Solution:
Swapped the two.
Platforms tested:
Linux
Bug fix and code checkpoint
Description:
HP's compilers didn't seem to like comparing function pointers with '<'
& '>', so I've changed how the function pointers are compared to aviod
those operations. This should fix last night's daily test failures on
opus and sangamon.
Also, more code development for generic property lists.
Platforms tested:
FreeBSD 4.1.1 (hawkwind)
