Three project files and three workspaces added into hdf5/windows/misc
Description:
Add the following files
/windows/misc/H5fortran_detect_gen/H5fortran_detect_gen.dsp
./windows/misc/H5fortran_detect_gen/H5fortran_detect_gen.dsw
./windows/misc/fortrantypegen/H5fortran_detect.dsp
./windows/misc/fortrantypegen/fortrantypegen.dsw
./windows/misc/matchtypegen/H5match_types.dsp
./windows/misc/matchtypegen/matchtypegen.dsw
Solution:
Platforms tested:
Misc. update:
New project files added into all.dsw
Description:
pool.dsp and pooldll.dsp added into all.dsw
Solution:
Platforms tested:
MSVS 6.0 on windows XP.
Misc. update:
New source files added to hdf5.dsp.
Description:
Add new source files H5MP.c and H5MPtest.c to hdf5.dsp
Solution:
Platforms tested:
MSVS 6.0 on windows XP.
Misc. update:
Add new source file into hdfdll project file under hdf5/windows.
Description:
New source files H5MP.c and H5MPtest.c added into hdf5dll.dsp
Solution:
Platforms tested:
MSVS 6.0 on windows XP.
Misc. update:
Add new project file into hdf5/windows
Description:
New C dll test project file pooldll.dsp added into all.dsw
Solution:
Platforms tested:
MSVS 6.0 on windows XP.
Misc. update:
Add new project file to hdf5/windows/all.dsw
Description:
New C library test project file pool.dsp added into all.dsw
Solution:
Platforms tested:
MSVS 6.0 on windows XP.
Misc. update:
Use to generate H5f90i_gen.h and H5fortran_types.f90
Description:
Special project file for windows.
Solution:
Platforms tested:
MSVS 6.0 on windows XP.
Misc. update:
Use to generate H5f90i_gen.h and H5fortran_types.f90 on windows.
Description:
Special workspace for windows.
Solution:
Platforms tested:
MSVS 6.0 on windows XP.
Misc. update:
Use to generate H5fortran_detect.f90 in windows.
Description:
Special workspace for windows.
Solution:
Platforms tested:
MSVS 6.0 on windows XP.
Misc. update:
Use to generate H5fortran_detect.f90
Description:
Special project file on windows XP.
Solution:
Platforms tested:
MSVS 6.0 on windows XP.
Misc. update:
Use to generate H5fort_type_defines.h in windows.
Description:
Special project file for windows.
Solution:
Platforms tested:
MSVS 6.0 on windows XP.
Misc. update:
Special wrokspace for windows.
Description:
Use to generate H5fort_type_defines.h in windows.
Solution:
Platforms tested:
MSVS6.0 on windows XP.
Misc. update:
Added new H5Error APIs.
Description:
Added C entries for
H5Eregister_class H5Eunregister_class
H5Eget_class_name H5auto_is_stack
H5Ecreate_msg H5Eget_msg
H5Eclose_msg
H5Eget_num H5Epop
H5Eget_current_stack H5Eclose_stack
Still missing -- 6 APIs that match existing APIs except that they
include a 'stack' parameter. Those APIs will be added shortly.
Platforms tested:
Firefox
Minor reformatting adn edits.
Added links to C++ RM and Parallel HDF5 sites.
Restructured document lists with the intent of easing discovery.
Description:
Platforms tested:
Firefox
New feature.
Description:
Add "memory pool" internal data structure. This set of routines is
designed to add a way to allocate small pieces of information for a particular
purpose and then free all the pieces at once (i.e. without having to free each
piece individually). Memory pools are also good for localizing lots of small
allocations that logically belong together.
For example, if you were constructing a temporary linked list, you could
create a new memory pool, allocate all the nodes for the list from the memory
pool and when you were done with the list, just destroy the pool instead of
tracking through the list and freeing each block independently.
Platforms tested:
FreeBSD 4.11 (sleipnir)
h5committest
Code cleanup
Description:
Convert H5Eget_num from 'size_t' to 'ssize_t' to allow for correct error
reporting (eventually).
Also, convert a bunch of static routines from "FUNC_ENTER_NOAPI" to
"FUNC_ENTER_NOAPI_NOINIT" or "FUNC_ENTER_NOAPI_NOINIT_NOFUNC"
Platforms tested:
FreeBSD 4.11 (sleipnir) w/pthreads
Too minor to require h5committest
bug fix
Description:
the fortran type integer*1 has become not portable.
define the image fortran datatype as "integer" and make special save, read, and palette functions
that use native integer for a memory type and UCHAR as disk type for the image data
added some more tests with new palette definitions
Solution:
Platforms tested:
linux
solaris
Misc. update:
Description: H5Eget_num used to return an INT type. That's not consistent with
the parameter of H5Epop.
Solution: Changed it to SIZE_T type.
Platforms tested: fuss - very simple change.
Code cleanup
Description:
fortran/src/H5f90i.h used to define different fortran types for every
compiler. Now it only defines three things, and they are the
same for every platform but one.
Solution:
Eliminate duplicate definitions; now there are only two cases
("UNICOS" and "everything else").
This should also let the HDF5 Fortran compile anywhere, even on
unknown platforms.
Platforms tested:
mir, pommier, modi4, copper, Cray X1 (in progress)
A little enhancement for integer datatype filter
Description:
Minimum-bit of the scaleoffset filter should not be less than 0,
However, if the user accidently input a negative value for integer type,
we will reset it to zero.
Solution:
Platforms tested:
too minor to test.
Misc. update:
Minor change for rounding of floating-point value for scaleoffset filter.
Description:
Always round to the bigger absolute integer value if floating-point data is in the middle.
0.5->1, -0.5->-1.
Solution:
Platforms tested:
Too minor to test
Misc. update:
Bug fix
Description:
Add missing #ifdef's to recent code cleanup that fix daily test failures.
Platforms tested:
FreeBSD 4.11 (sleipnir)
Too minor to require h5committest
bug fix 366
Description:
the printing of the array indices was done relatively to the stripmine data (data read
by hyperslabs when its memory requiremnts are too large)
this was causing an incorrect numbering of the array indices (the next read would initialize
the indices to zero)
Solution:
added a field to the print context that keeps track of the stripmine position
and pass to the rendering function the total element position
Platforms tested:
Linux
Misc. update:
Bug fix
Description:
Fix several problems with mounting files on an entry in a group when the
file the group is in has been closed.
Platforms tested:
FreeBSD 4.11 (sleipnir)
h5committest
Update H5Zscaleoffset.c so that it supports negative precision number.
Description:
Solution:
Platforms tested:
Linux 2.4
Misc. update:
too minor to check in other platforms
Description: A user reported conversion of floating-point numbers returned some wrong value. It turned
out a variable in H5T_conv_f_f() wasn't reset after being used for each element. See bug #356 for details.
Solution: Simple reset it.
Platforms tested: fuss; also tested with the program user provided. Very simple change.
Portability feature
Description:
The random() and srandom() functions are not available on all machines.
Set up the configure script to automatically detect them, rather than
requiring their presence or absence to be hardcoded.
Solution:
Added AC_CHECK_FUNCS macro to configure.in and replaced
#ifdef WIN32 conditionals with #ifdef H5_HAVE_RANDOM conditionals.
Platforms tested:
sleipnir, Windows
Cray X1 Port
Description:
Porting 1.7 branch to Cray X1. With these changes, HDF5 builds, but
there are some errors in the tests. Working on the errors.
Solution:
Added nv1-cray file to config directory.
Cleaned up some code in hl/c++ that was causing compiler to complain.
Platforms tested:
Cray X1, mir, sleipnir
Misc. update:
bug fix
Description:
some lines were longer that 132 characters (the "official" max lenght of a line in F95)
the g95 compiler complained about it
Solution:
reduce the line lenght
Platforms tested:
linux (g95, intel)
Misc. update:
bug fix
Description:
the hdf5 file of the test was being generated on the NFS filesystem
Solution:
use the libtest function h5_fixname
Platforms tested:
Linux
Misc. update:
added new fortran parallel tests that replace the old ones
Description:
this first set of tests do a write/read by hyperslabs
Solution:
Platforms tested:
Linux (heping) PGI compiler
Linux (heping) Intel compiler
AIX (copper) IBM compiler
SGI Altix (cobalt) Intel compiler
Xeon Linux cluster (tungsten) Intel compiler
Misc. update:
Remove C99 types from new metadata cache related API calls
(take 2 -- missed one in the previous check in)
Description:
Windows (and perhaps others) don't like int32_t and int64_t. While
we have dealt with the issue internally, it is more of a problem
in API calls.
Solution:
Convert int32_t to int and int64_t to long int in the new metadata
cache related API calls.
Platforms tested:
heping
Misc. update:
Code cleanup
Description:
Migrate MPE macros to H5private.h, in preparation for using the
H5MPprivate.h file for "memory pool" API.
Platforms tested:
FreeBSD 4.11 (sleipnir) w/parallel
Too minor to require h5committest
