| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
| |
Description: Frank and I noticed that h5jam and h5unjam were using
old names when printing help message.
Solution: Fixed
Platforms tested: heping
Misc. update:
|
|
|
|
|
|
|
|
|
| |
Description: On SGI Altix(cobalt) Linux, wrong values were printed out with enum data type.
No apparent reason was found out.
Solution: Use another pointer to the buffer containing enum data type values to do print.
Platforms tested: cobalt - very simple change.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
bug fix
Description:
in 1 case when there was not a requested layout, the original chunk was not preserved
Solution:
Platforms tested:
linux (small change)
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
bug fix
Description:
a string with 2 characters was declared as having size 2. changed to size 3 (+ null character)
this was causing compiler warnings with g++
Solution:
Platforms tested:
linux
free bsd (g++)
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
bug fix, new test file
Description:
h5dump was not properly displaying array indices > 3D
Solution:
added the same algorythm and data structure that h5diff uses to calculate the array index
from a element number position
Platforms tested:
linux
solaris
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
| |
Description: A little refactoring in tools/h5dump/h5dump.c for the bug fix of
wrong byte order of 1-byte integer types.
Platforms tested: No test needed, too simple.
Misc. update: Added information about the bug fix of wrong byte order of
1-byte integer in RELEASE.txt
|
|
|
|
|
|
|
|
|
|
|
| |
Description: H5T_NATIVE_SCHAR, H5T_NATIVE_UCHAR were always considered as little
endian due to the algorithm to detect their order in H5detect.c. This error
didn't affect data but didn't look right.
Solution: In H5detect.c, use native int instead to detect order if type size
is only 1 byte.
Platforms tested: h5committest, fuss and arabica
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug
Description:
Refinement to bug fix for bug #282
Solution:
See other checkin
Platforms tested:
verbena
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix, #282
Description:
Not honoring input byte order when input file is binary.
Solution:
Do the byte order even in the default case.
Platforms tested:
verbena, shanti
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix (#56)
Description:
Bug 56, h52gif crashes if data is not 1 byte.
Solution:
Check data type and exit with error when data is bad type.
Platforms tested:
shanti,verbena.
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
bug fix
Description:
changed a close function test failure from < 1 to < 0
Solution:
Platforms tested:
linux
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix (#181)
Description:
Attributes written wrong in output.
Solution:
Do the right thing. Note: followed H5IM.c as a model.
Platforms tested:
verbena. output checked on windows.
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix
Description:
In an earlier bug-fix, I inadvertently inverted the meaning of the
"% utilization" in h5ls.
Solution:
De-invert it. :-)
Platforms tested:
FreeBSD 4.10 (sleipnir)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Remove feature
Description:
Retire threaded, balanced binary tree code from HDF5 use. Requiescat in
pace...
Also, regenerate dependencies files.
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel
Too minor to require full h5committesting (the code is already
disconnected from everything except its tests)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix.
Description:
h5import.c includes the following scripts especially for CodeWarrior:
#if defined __MWERKS__
argc = ccommand(&argv);
#endif
However, ccommand() function is not available with CodeWarrior 8.0 .
Solution:
Remove the above scripts from h5import.c.
Platforms tested:
CodeWarrior 8.0 on Windows XP.
(I discussed with Pedro about this update. He suggested to deleter those scripts. And these scripts only work with CodeWarrior, so it is not necessary to test on other platforms.)
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix.
Description:
When building h5dumpgentest.c with Visual C++ 6.0, I got the following warning message:
h5dumpgentest.c(4528) : warning C4013: 'h5tools_can_encode' undefined; assuming extern returning int
Solution:
Add the header file h5tools.h into h5dumpgentest.c.
Platforms tested:
Linux 2.4 (Heping)
Visual C++ 6.0 on Windows XP/2000.
.NET on Windows XP.
(Tested with SZIP-with-encoder and SZIP-without-encoder).
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix.
Description:
creat(name, (mode_t)0777) is used to create a new file. However, creat() is not available in Visual Studio.
Solution:
Use a Win32 macro so _creat() will be used on Windows and creat() will be used on Unix. #ifdef WIN32 fd = _creat(name, _S_IREAD | _S_IWRITE); #else /* WIN32 */ fd = creat(name,(mode_t)0777); #endif /* WIN32 */ (Note: At first, I tried to define HDcreat() as _creat() for WIN32 and creat() for other platforms in H5private.h. But _creat() and creat() use different parameters for file permission settings. So, I added the WIN32 macro in h5jamgentest.c.)
Platforms tested:
Linux 2.4 (Heping)
Visual C++ 6.0 on Windows XP/2000.
.NET on Windows XP.
(Tested with SZIP-with-encoder and SZIP-without-encoder).
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:
For definitions related to the file stat utilties are included in hdf5/src/h5private.h as following:
#ifdef WIN32
#ifdef __MWERKS__
#define HDfstat(F,B) fstat(F,B)
typedef struct stat h5_stat_t;
typedef off_t h5_stat_size_t;
#else /*MSVC*/
#define HDfstat(F,B) _fstati64(F,B)
typedef struct _stati64 h5_stat_t;
typedef __int64 h5_stat_size_t;
#endif
#else
#define HDfstat(F,B) fstat(F,B)
typedef struct stat h5_stat_t;
typedef off_t h5_stat_size_t;
#endif
So, on CodeWarrior, these file stat utilities were defined as off_t.
However, in hdf5/tools/misc/h5repart.c, the above file stat utilities on CodeWarrior were defined as __int64 due to missing the Macro for CodeWarrior:
#ifdef WIN32
_int64 left_overs=0; /*amount of zeros left over */
_int64 src_offset=0; /*offset in source member */
_int64 dst_offset=0; /*offset in destination member */
_int64 src_size; /*source logical member size */
_int64 src_act_size; /*source actual member size */
_int64 dst_size=1 GB; /*destination logical memb size */
#else
off_t left_overs=0; /*amount of zeros left over */
off_t src_offset=0; /*offset in source member */
off_t dst_offset=0; /*offset in destination member */
off_t src_size; /*source logical member size */
off_t src_act_size; /*source actual member size */
off_t dst_size=1 GB; /*destination logical memb size */
#endif
Solution:
Add the CodeWarrior Macro for above codes in h5repart.c, so those file stat utilties match with the definitions in h5private.h.
#if defined(WIN32) && ! defined (__MWERKS__)
_int64 left_overs=0; /*amount of zeros left over */
_int64 src_offset=0; /*offset in source member */
_int64 dst_offset=0; /*offset in destination member */
_int64 src_size; /*source logical member size */
_int64 src_act_size; /*source actual member size */
_int64 dst_size=1 GB; /*destination logical memb size */
#else
off_t left_overs=0; /*amount of zeros left over */
off_t src_offset=0; /*offset in source member */
off_t dst_offset=0; /*offset in destination member */
off_t src_size; /*source logical member size */
off_t src_act_size; /*source actual member size */
off_t dst_size=1 GB; /*destination logical memb size */
#endif
Platforms tested:
Heping.
Visual Studio 6.0 on Windows XP.
CodeWarrior 8.0 on Windows XP.
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Using HDunlink at this file to make it portable on windows.
Description:
Solution:
Platforms tested:
Linux 2.4(too minor to use h5committest)
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
bug fix
Description:
added a condition that checks if the hl library is enabled to add it to the list of linking libraries
Solution:
Platforms tested:
linux
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
modified h5cc to link with the high level library (libhdf5_hl)
Description:
added libhdf5_hl.a to the list of libarries to link with
Solution:
Platforms tested:
linux
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
new feature
Description:
when requesting a filter for all datasets and the filter cannot be applied for one dataset
use the original dcpl withou filter and print a warning
this check was done previously with buit in functions for szip that are now in the library
Solution:
Platforms tested:
linux
solaris
aix
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
bug fix
Description:
same bug fix as the 1.7 branch, the logic for dealing with no input layout was wrong
Solution:
Platforms tested:
linux
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix (#264)
Description:
h5dump did not print attribute data in ASCII format when
-r is used.
Solution:
Added the ability to print in ASCII for Attributes Data also.
Added a test for printing Attributes with -r option.
tall-2B.ddl is the standard output for printing attributes with -r option.
Platforms tested:
tested in heping.
Misc. update:
Update MANIFEST.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Code cleanup & optimizations
Description:
Clean up some of the code in attributes to avoid allocating memory and
performing type conversions when the conversion is a noop.
Avoid memory allocations of attribute data structures by switching to use
library's free list memory allocator routines.
Avoid memory allocations of object header continuation data structures by
switching to use library's free list memory allocator routines.
Rearrange threaded, balanced binary tree macros slightly to avoid some
overhead.
Platforms tested:
FreeBSD 4.10 (sleipnir) w/parallel
Solaris 2.7 (arabica)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix
Description:
Change H5G_obj_t -> int when 1.4 compatibility is turned out
Platforms tested:
FreeBSD 4.10 (sleipnir) w/1.4 compat
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
enhanced performance for h5diff
Description:
the comparison cycle for the 2 buffers read from file was using a compare function
where the type must be checked for each datum (e.g due to recursive calls in compound types)
for atomic types this checking can be avoided
Solution:
introduced cycles for each one of the atomic integer and float class types
Platforms tested:
linux
aix
solaris
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix
Description:
Correct values used in "space utilization" equation: the 'used' and
'total' values were backwards.
Platforms tested:
FreeBSD 4.10 (sleipnir)
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
bug fix
Description:
moved the option handling algorythm to just one file
Solution:
Platforms tested:
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
bug fix
Description:
one case was not handled in the combination of input options (layout and filters)
Solution:
redo the algorythm that handles all cases
Platforms tested:
linux
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
bug fix
Description:
when specifying both an input object e.g -f mydset:GZIP=1 and a defined chunk -l CHUNK=20x20
the filter used a defined default chunk instead
Solution:
add a check for the input chunk
Platforms tested:
linux (small change)
Misc. update:
|
|
|
|
|
|
|
|
|
| |
The script does not use $RUNSERIAL to execute programs.
This failed badly in parallel systems like Tflops.
Patched $RUNSERIAL in to make it work.
Tested in tflops only.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Updated three h5dump --xml test outputs.
Description:
three h5dump xml tests failed on windows and we finally figured out it was due to
the update of HDF5 library. Several test outputs needed to be regenerated.
Solution:
Generated the three new xml output files.
Platforms tested:
windows XP?(xuan did), h5committest with xml tests turned on.
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix
Description:
Correct a couple of array bounds issues exposed by the PGI compiler
Platforms tested:
Linux 2.4 (verbena) w/PGI compilers
Too minor too requie h5committest
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix
Description:
Include "h5tools.h" header for native type conversion function prototype
to unbreak CC=g++ builds.
Platforms tested:
FreeBSD 4.10 (sleipnir)
Not tested by h5committest
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Putting 'h5jam' in the 1.6 branch
Description:
Solution:
Platforms tested:
verbena, arabica, cobalt
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
| |
Code cleanup/bug fix
Description:
Change some 'hsize_t' index variables to 'int' variables to avoid problems
with PGI compilers.
Platforms tested:
Linux 2.4 (tungsten) w/PGI C
Too minor to require h5committest
|
|
|
|
|
|
|
|
|
| |
Description: In last checkin, for a function call lookup_ref_path(refbuf) on line 5039
was accidentally changed to lookup_ref_path(*refbuf).
Solution: changed it back.
Platforms tested: no test needed for such a simple change.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Refactor code
Description:
Refactor common code for determining the native type for using in the
tools into separate routine.
Also, reduce diffs between the two branches and bring back some fixes from
the development branch to the release branch.
Platforms tested:
FreeBSD 4.10 (sleipnir)
too minor to require h5committest
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Purpose: change feature
Description: Back up supporting bitfield and time datatypes in H5Tget_native_type. Leave it for future support. Simply returns "not supported" error message for now.
Platforms tested: h5committest
Misc. update: RELEASE.txt
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Purpose:
Feature
Description:
(Same change to release branch)
Datatypes and groups now use H5FO "file object" code that was previously
only used by datasets. These objects will hold a file open if the file
is closed but they have not yet been closed. If these objects are unlinked
then relinked, they will not be destroyed. If they are opened twice (even
by two different names), both IDs will "see" changes made to the object
using the other ID.
When an object is opened using two different names (e.g., if a dataset was
opened under one name, then mounted and opened under its new name), calling
H5Iget_name() on a given hid_t will return the name used to open that hid_t,
not the current name of the object (this is a feature, and a change from the
previous behavior of datasets).
Solution:
Used H5FO code that was already in place for datasets. Broke H5D_t's, H5T_t's,
and H5G_t's into a "shared" struct and a private struct. The shared structs
(H5D_shared_t, etc.) hold the object's information and are used by all IDs
that point to a given object in the file. The private structs are pointed
to by the hid_t and contain the object's group entry information (including its
name) and a pointer to the shared struct for that object.
This changed the naming of structs throughout the library (e.g., datatype->size
is now datatype->shared->size). I added an updated H5Tinit.c to windows.zip.
Platforms tested:
Visual Studio 7, sleipnir, arabica, verbena
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug fix.
Description:
The patch for running XML test script had undefined variables that
pmake (IRIX64 native make) would not tolerated. It is bad to duplicate
codes too.
Solution:
Replaced all the duplicated code with recursive call of make with
a redefined TEST_SCRIPTS. This has the problem of not calling
the XML tests by default but the XML tests failed in some platforms
including the IRIX64. This solution has the same effect as before
but cleaner. Long term fix should be done in h5dump or the XML test
scripts.
Platforms tested:
H5committested. Also in modi4 using pmake.
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
bug fix
changed some names of the datsets for the h5repack tests
Description:
the dataset region reference data was not portable between the Cray T3E and other machines
Solution:
this was due to the fact that the buffer to store the heap ID and index was using a sizeof(int) for its
size
4 is used instead of sizeof(int) to permit portability between
the Crays and other machines (the heap ID is always encoded as an int32 anyway)
Platforms tested:
Cray T3E (read data from linux)
linux
solaris
aix
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
h5repack tests
Description:
modified a test file generation contents, for more easy debugging
(generated just one reference dataset and one region reference )
Solution:
Platforms tested:
linux
solaris
aix
Misc. update:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
new test for h5repack
Description:
added a test that generates and copies a file with a dataset with fill value
(this is to test the property list function H5Pequal)
Solution:
Platforms tested:
linux
solaris
aix
Misc. update:
|