and replace it with new url for COPYING file.
Fix 2 lines in java error test expected output file where messages
include line numbers changed by reducing the copyright header by 2
lines.
HDFFV-9049
Intel 15.0 15.0.1 Fails with --enable-fortran2003
Added BIND to the Fortran 2003 functions.
Tested: windows (tested concept)
linux (--enable-fortran and --enable-fortran2003)
Prep work for adding a memory free function, part of HDFFV-8551.
Tested on:
32-bit LE linux (jam)
* with C++, Fortran
* with thread safety on
These are minor changes (almost entirely renaming foo functions
to HDfoo) that don't change functionality. This should not
require a full commit test.
- Fixed a typo in return value of the nh5dread_f_c function ( was 1
instead of 0 on success); fixed the return value to make it consistent
with other Fortran functions; cleaned the code from debug statements.
Platforms tested: jam with GNU and Intel compilers, fortran 2003 feature.
Items merged: fortran directory,
src/libhdf5.settings.in
configure.in configure
MANIFEST
Tested: (all platforms used by daily tests, both with --enable-fortran and --enable-fortran2003)
Remove trailing whitespace from C/C++ source files, with the following
script:
foreach f (*.[ch] *.cpp)
sed 's/[[:blank:]]*$//' $f > sed.out && mv sed.out $f
end
Tested on:
Mac OS X/32 10.5.5 (amazon)
No need for h5committest, just whitespace changes...
Platforms tested: kagiso with g95 and Intel compilers; more testing will be done after checking in a fresh
copy from the trunk. New code itself was tested with all Fortran compilers available at THG
Make H5Dopen versioned and change all internal usage to use H5Dopen2
Add simple regression test for H5Dopen1
Tested on:
FreeBSD/32 6.2 (duty) in debug mode
FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe,
in debug mode
Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN,
in production mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, in production mode
Mac OS X/32 10.4.10 (amazon) in debug mode
Deprecate H5Dextend in favor of H5Dset_extent (without using API
versioning, due to changed behavior) and switch internal usage to H5Dset_extent
Tested on:
FreeBSD/32 6.2 (duty) in debug mode
FreeBSD/64 6.2 (liberty) w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (kagiso) w/PGI compilers, w/C++ & FORTRAN, w/threadsafe,
in debug mode
Linux/64-amd64 2.6 (smirom) w/default API=1.6.x, w/C++ & FORTRAN,
in production mode
Linux/64-ia64 2.6 (cobalt) w/Intel compilers, w/C++ & FORTRAN,
in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, in production mode
Mac OS X/32 10.4.10 (amazon) in debug mode
Tested platform:
Kagiso only since it is only a comment block change. If it works in one
machine, it should work in all, I hope. Still need to check the parallel
build on copper.
Previous implementation failed to compile on MAC OS X Tiger 10.4 and AIX 5.2
In previous implementation, each overloaded Fortran function called the same
C function. Since declaration of parameters in INTERFACE block for
C Function differed, some compilers issued warnings but g95 on MAC and AIX
failed to compile.
Solution: Add C wrappers corresponding to each overloaded Fortran function
to call C stub function. This fix should also reduce number of warnings
on IRIX systems.
Platforms tested: MAC OS and AIX 5.2 (copper) with g95;
sol (with Sun compilers), heping with PGI compilers
Code cleanup
Description:
Fix a bunch of warnings flagged by Windows compilers.
Platforms tested:
FreeBSD 4.11 (sleipnir)
Too minor to require h5committest
Code cleanup
Description:
Trim trailing whitespace, which is making 'diff'ing the two branches
difficult.
Solution:
Ran this script in each directory:
foreach f (*.[ch] *.cpp)
sed 's/[[:blank:]]*$//' $f > sed.out && mv sed.out $f
end
Platforms tested:
FreeBSD 4.11 (sleipnir)
Too minor to require h5committest
Description:
When compiler flag was used to set the size of Fortran integer
to 8 bytes, library would fail.
Solution:
Cleaned up the code; added detection of Fortran INTEGER type size
and appropriately defined int_f type for C-stubs routines.
Platforms tested:
Solaris 2.8 32 and 64-bit, AIX 5.1 64-bit parallel
PGI Fortran with -i8 flag on heping
Absoft Fortran with -i8 flag on heping
g95 on mir (Fortran integer is 8 bytes by default that cannot be
changed - compiler bug ;-)
AIX Fortran with -qintsize=8 32 and 64-bit modes on copper
Misc. update:
Bug Fix/Code Cleanup/Doc Cleanup/Optimization/Branch Sync :-)
Description:
Generally speaking, this is the "signed->unsigned" change to selections.
However, in the process of merging code back, things got stickier and stickier
until I ended up doing a big "sync the two branches up" operation. So... I
brought back all the "infrastructure" fixes from the development branch to the
release branch (which I think were actually making some improvement in
performance) as well as fixed several bugs which had been fixed in one branch,
but not the other.
I've also tagged the repository before making this checkin with the label
"before_signed_unsigned_changes".
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel & fphdf5
FreeBSD 4.10 (sleipnir) w/threadsafe
FreeBSD 4.10 (sleipnir) w/backward compatibility
Solaris 2.7 (arabica) w/"purify options"
Solaris 2.8 (sol) w/FORTRAN & C++
AIX 5.x (copper) w/parallel & FORTRAN
IRIX64 6.5 (modi4) w/FORTRAN
Linux 2.4 (heping) w/FORTRAN & C++
Misc. update:
Bug fix
Description:
Correct typedef for dataset region references to avoid struct alignment
issues on Crays.
Solution:
Change the typedef for hdset_reg_ref_t from a struct to an array of
unsigned char's of the correct size and propagate the appropriate adjustments
around the code.
Platforms tested:
FreeBSD 4.10 (sleipnir)
Linux 2.4 (verbena) w/fortran
Cray T90 (subzero) w/fortran
Cray SV1 (wind) w/fortran & parallel
Cray T3E (cyclone) w/fortran & parallel
Purpose: Maintenance
Description: h5dwrite/read_f and h5awrite_read_f functions were overloaded
with dims argument being of type INTEGER and INTEGER(HSIZE_T).
We promised to remove overloading in 1.4 release. It was
removed from documentation but not from the source code.
Solution: Finish code cleanup.
Platforms tested: Solaris 2.7 32 and 64 bit, and copper with --enable-
parallel. Windows are broken, so Kent and Xuan
will test it with their changes later.
Misc. update:
Code cleanup
Description:
Cleaned up various compiler warnings
Platforms tested:
FreeBSD 4.9 (sleipnir)
Linux 2.4 (verbena) w/FORTRAN
too minor for h5committest
Description: On Linux systems valdrind tool complained about memroy leaks in the
following statements like
if(!a) free(a);
Solution: replaced the statements with
if ( a != NULL) free(a);
Platforms tested: eirene (too small for committest)
Misc. update:
Code cleanup
Description:
Switched 'hobj_ref_t' from funny structure with array inside to just be
'haddr_t', since that was equivalent and less confusing.
Platforms tested:
h5committested
Purpose: Windows maintenance and code cleanup
Description: There were a lot of places in the source code where varibale
was declared but never used.
Solution: Cleaned the source code to avoid warnings on Windows adn Unix
platforms.
Platforms tested: arabica, modi4, Windows
Probably my environment is wrong on Linux, cannot
compile at all when fortran is enabled
If daily tests fails tomorrow, this will be my fault
but I want to go home now :-)
Misc. update:
Purpose: Added copyright statement; cleaned code and fixed bugs for
Windows and Cray T90IEEE
Description: Fortran files did not have copyright statement; VL types
did not work on T90IEEE since I never brought the correct code
from 1.4 branch; there were compilation warnings on Windows;
some character parameters were not passed correctly to C stubs
causing tests to fail on Windows.
Solution: Added copyright statement and clean the code.
Platforms tested: CRAY T90IEEE, arabica, modi4 with parallel, Windows 2000
Only static tests (both debug and release) passed on Windows.
DLLs have multiple problems ;-)
Misc. update:
Bug fix.
Description:
The macro MAX defined here is very commonly defined, e.g., AIX systems.
The compiler did not like duplicated macro defines.
Solution:
H5_MAX is already defined in H5f90.h and used in other code.
Remove MAX and change its use to H5_MAX.
Platforms tested:
"h5committested", plus tested at LLNL Blue.
Misc. update:
Purpose: Catching up with the C library
Description: Added the follwoing new fortran functions
h5iget_name_f
h5tis_variavle_str_f
h5zunregister_f
h5zfilter_avail_f
h5pset_shuffle_f
h5pset_fletcher32
h5pset_edc_check_f
h5pget_edc_check_f
h5dfill_f
Solution:
Platforms tested: arabica(C and F90), burrwhite (pgcc and pgf90), modi4 (F90 and parallel)
Misc. update:
Purpose:
Maintenance
Description:
* Added support for generic properties.
* Added support for time allocation properties.
* Added support for variable length datatypes (only datatypes based
on INTEGER, REAL and CHARACTER Fortran types are supported).
* added some missing functions
Solution:
I am checking in new Fortran APIs and their man pages to support
1.5 features listed above.
Not all APIs have tests yet. APIs were written in Fall 2002, and I am afraid
that I will loose the code or totally forget what I did if I wait longer. ;-)
Platforms tested:
arabica (fortran), eirene (fortran), modi4 (parallel, fortran)
Code cleanup
Description:
While working on the "External test" failure I restructured and
cleaned up some C stub code.
Platforms tested:
dangermouse (Linux 2.4)
Purpose:
Windows port
Description:
Multiple changes:
* Windows platforms require special compiler directives
in order to create DLLs.
* In read/write subroutines data arrays were passed by descriptor.
This worked on UNIX but did not work on Windows.
Solution:
* added compiler directives.
* read/write APIs have been changed. There is an additional
parameter (array that contains the sizes of data buffer dimensions)
and regular arrays are used instead of assumed-shaped arrays.
Platforms tested:
* Currently this feature does not work. Common blocks are not exported correctly
from one F90 module to another.
I am checking this in so I can ask DEC for help.
* For static library tests passed on Windows 98 ( except flush2_fortran)
All tests passed on Linux, Solaris 2.7, O2K and T3E
Purpose:
New feature
Description:
Now all Fortran flags such as H5F_ACC_RDONLY_F (previously defined in the
H5fortran_flags.f90 file) are generated at runtime when h5init_fortran_f
subroutine is called. All flags have now the same value as
corresponding C flags. This change affects user's programming model:
Every Fortran program that uses F90 HDF5 Library has to call
h5init_fortran_f(error) subroutine before the first call to the
F90 HDF5 Library and h5close_fortran_f(error) after the last call to the
Library. h5init(close)_types_f calls are not needed anymore since
F90 datatypes are intialized(closed) with the h5init(close)_fortran_f
calls.
Platforms tested:
O2K and Solaris2.7
Purpose:
Reimplemented references to the objects and dataset regions.
Description:
Previous implementation was not portable. This implementation
should work on UNIX workstations and Crays, but is very inefficient
since it uses memcpy to repack Fortran buffers with references
to C buffers and vice versa.
Solution:
I used fortran derived datatype with integer fields. h5dwrite_f and
h5dread_f have extra parameter when references are written or read.
This parameter describes size of the buffer that holds references.
Platforms tested:
J90 and Solaris 2.6
H5Pf.c
Some of the functions do not exist now in the development branch.
Commented those out, so one does not need to apply patch in order
to build Fortran Library.
Adding the Fortran interface to the HDF5 library
Description:
Fortran is now a subdirectory of the HDF5 library tree.
Platforms tested:
Solaris and IRIX (O2K)
