| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
one more fix related to bug 551. the printing of the subsetting parameters was not using the new hsize_t printing format
correct output is now
[pvn@kagiso ad]$ /home1/pvn/build_hdf5/tools/h5dump/h5dump -d /a -s "0,8899716890" -c "1,10" jasp_char.h5
HDF5 "jasp_char.h5" {
DATASET "/a" {
DATATYPE H5T_STD_I8LE
DATASPACE SIMPLE { ( 20, 8899716900 ) / ( 20, 8899716900 ) }
SUBSET {
START ( 0, 8899716890 );
STRIDE ( 1, 1 );
COUNT ( 1, 10 );
BLOCK ( 1, 1 );
DATA {
(0,8899716890): 1,
(0,8899716891): 1,
(0,8899716892): 2,
(0,8899716893): 2,
(0,8899716894): 3,
(0,8899716895): 3,
(0,8899716896): 4,
(0,8899716897): 4,
(0,8899716898): 5,
(0,8899716899): 5
}
}
}
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
fix for bugzilla bug #551
several programming errors contributed to this bug
1) the parsing of subsetting was using atoi to convert the parameter to an int,
which caused problems for numbers greater that int. Substitute with atof
2) several index counters were declared as int, use hsize_t instead
3) the numerical format passed for printf was %lu, defined one compatible with
hsize_t instead (unsigned long long)
|
|
|
|
|
|
|
|
| |
Fix display order of attributes.
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fix several bugs
1) the parsing of subsetting was using atoi to convert the parameter to an int, which caused problems for numbers greater that int. Substitute with atof
2) the printing of indices in the subsetting case was not being done. Solution: calculate the element position at the start of the subsetting using the algorythm
Given an index I(z,y,x) its position from the beginning of an array of sizes A(size_z, size_y,size_x) is given by
Position of I(z,y,x) = index_z * size_y * size_x
+ index_y * size_x
+ index_x
And pass that position to the function that dumps data, h5tools_dump_simple_data.
3) several index counters were declared as int, use hsize_t instead
4) modified the test generation program so that it includes test cases for subsetting of 1d, 2d, 3d, and 4d arrays and add these tests to the shell script
|
|
|
|
|
| |
added a size of array information when printing differences
modified print_pos and diff_datum for having one extra argument *dims
|
|
|
|
|
|
|
|
| |
Take out separate memory type in the file for SOHM objects and create
aliases for existing memory types for SOHM use.
Tested on:
FreeBSD/32 4.11 (sleipnir)
|
|
|
|
| |
avoid usage to call exit(1), it makes processes to stop in parallel.
|
|
|
|
| |
introduced a new macro PDIFF to calculate differences for unsigned types
|
|
|
|
| |
introduced more ABS macro use
|
|
|
|
|
|
| |
warnings clean
../../../hdf5/tools/h5repack/h5repack_opttable.c:240: warning: passing arg 2 of `aux_tblinsert_layout' as signed due to prototype
|
|
|
|
|
|
|
|
| |
warnings clean
../../../hdf5/tools/h5repack/h5repack_copy.c:615: warning: passing arg 3 of `print_dataset_info' as `float' rather than `double' due to prototype
introduced double precision arithmetic
|
|
|
|
|
|
|
|
| |
fixed warning
../../../hdf5/tools/h5diff/h5diff_common.c: In function `usage':
../../../hdf5/tools/h5diff/h5diff_common.c:346: warning: function might be possible candidate for attribute `noreturn'
|
|
|
|
| |
fix some memory allocation problem with info->prefix
|
|
|
|
|
|
| |
compiler warning clean
../../../hdf5/tools/lib/h5tools_ref.c:257: warning: old-style parameter declaration
|
|
|
|
|
|
|
|
| |
fixed 2 initializations of char* with HDstrdup and HDcalloc
info->prefix = HDcalloc(1, 1);
fname = HDstrdup(argv[opt_ind]);
some were exposed by compiler warnings
|
|
|
|
|
|
| |
more compiler warning
../../../hdf5/tools/h5diff/h5diffgentest.c:111: warning: passing arg 1 of `test_hyperslab' discards qualifiers from pointer target type
|
|
|
|
|
|
| |
more compiler warning
../../../hdf5/tools/lib/h5diff_attr.c:214: warning: passing arg 4 of `diff_array' with different width due to prototype
|
|
|
|
|
|
| |
cleaned more warnings
../../../hdf5/tools/lib/h5diff_attr.c:214: warning: passing arg 4 of `diff_array' with different width due to prototype
|
| |
|
|
|
|
|
|
|
|
| |
Cleaned warnings
h5diff_array.c:804: warning: passing arg 1 of `fabs' as floating rather
than integer due to prototype
introduced double precision arithmetic when possible instead of single
precision
|
|
|
|
|
|
| |
floating point comparison
../../../hdf5/tools/lib/h5diff_array.c: warning: comparing floating point with == or != is unsafe
|
|
|
|
| |
h5diff_array.c:127: warning: declaration of 'per' shadows a global declaration
|
|
|
|
|
|
| |
cleaned compiler warning
../../../hdf5/tools/lib/h5diff_array.c:139: warning: declaration of 'per' shadows a global declaration
|
|
|
|
| |
terminator is not correct
|
|
|
|
| |
remove and insert these files, somehow their end of line terminator is not correct
|
|
|
|
|
|
|
| |
Added a relative error formula to deal with floating point uncertainty
in the comparison of floats and double types.
Added new tests for this feature to the file generator program and to
the shell script
|
|
|
|
|
| |
changed the usage message relative to the relative error formula to a more
common and easy to understand form | (b-a) / a |
|
|
|
|
| |
open/close in the cheking of objects when there are none requested by the user
|
|
|
|
|
| |
do not call H5Ocopy only when there is a user request for
filters/layout
|
|
|
|
| |
Removed all but one warnings for h5stat on Linux (heping)
|
|
|
|
| |
make h5repacktst clean a big file which name was changed to "h5repack_big_out.h5", do not use H5Ocopy only when the original DCPL has filters or a request is made for such, more code cleaning
|
|
|
|
|
|
|
|
|
|
|
|
| |
Basic support for H5Literate() routine. Still needs to be fleshed out and
refactored to simplify. Also, needs tests. :-)
Tested on:
FreeBSD/32 4.11 (sleipnir)
Linux/32 2.4 (heping)
Linux/64 2.4 (mir)
AIX/32 5.? (copper)
Mac OS X/32 10.4.8 (amazon)
|
|
|
|
|
|
|
|
|
|
|
| |
The version of libtool used by HDF5 isn't directly affected by the reconfigure
script; instead, libtoolize --force must be used by hand. Libtool was the
source of the problem, so rolling its version back to 1.5.14 should solve the
issue (at least temporarily).
Reconfigure should still work on both heping and kagiso.
Tested on heping, kagiso, and tg-login3.
|
| |
|
|
|
|
| |
removed whitespace
|
|
|
|
|
|
|
|
| |
h5repack revision:
1. added a new test due to the introduction of H5Ocopy in the copy of objects (compressed dataset with references, that still must go a second sweep of the file to be regenerated).
2. Moved all the source files from the h5repack test program to a new file h5repacktst.c and removed the old ones (testh5repack*.c).
3. Renamed the binary files from test*.h5 to h5repack*.h5 for easy reference.
4. Modified the shell script to use variables for file names instead of hard coded names
|
|
|
|
|
|
|
|
| |
Hopefully this will fix
issues on tg-login3.
bin/reconfigure should still work on both heping/mir and kagiso.
|
|
|
|
|
|
|
|
|
|
|
|
| |
This feature is still in progress; Shared Object Header Messages are not
complete as a feature and are not thoroughly tested. There are still
"TODO" comments in the code (comments with the word "JAMES" in them,
so as not to be confused with other TODO comments).
Hopefully this checkin will reduce the liklihood of conflicts as I finish
implementing this feature.
All current tests pass on juniper, copper (parallel), heping, kagiso, and mir.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add new H5Lget_val_by_idx() routine & tests.
Also includes most of changes for H5Ldelete_by_idx() routine.
Tested on:
Mac OS X/32 10.4.8 (amazon)
FreeBSD/32 4.11 (sleipnir)
Linux/32 2.4 (heping)
Linux/64 2.4 (mir)
AIX/32 5.? (copper)
|
|
|
|
|
| |
Introduced the second sweep of the file for a case a reference is present and H5Ocopy was not used.
Moved the code from file h5repack_refs.c to h5repack_copy.c and removed the first file
|
|
|
|
|
|
|
|
|
|
|
|
| |
Should disable linking against shared libraries in Fortran for compilers that
don't support shared libraries.
Should also fix problem when the wrong Fortran file extension was specified.
If these changes don't solve the Daily Test issues, I'll look at backing out
the autotool version change until I have time to fix them.
Tested on heping, kagiso, juniper.
|
|
|
|
|
|
|
|
|
|
|
|
| |
h5repack support for H5Ocopy in the copy of objects. The old method
for recreating references was dropped (references recreated in a second
traversal of the file)
The logic for using H5Ocopy or not is
if the input DCPL has filters or non default layout OR these are
requested by the user THEN
use the old h5repack read / write
ELSE
use H5Ocopy
|
|
|
|
|
|
| |
h5dump bug 701. Symptom: The creation of a hardlink pointing to the root group "/" causes h5dump to display it as a link pointing to itself.
Cure: the root group was not being inserted in the table that keeps track of object names and links.
Added a test for this in the test generation program, the creation of a hardlink to the root
|
|
|
|
|
|
|
|
| |
Linux machines.
Updated to the latest versions of autotools.
Tested on kagiso, heping, and juniper. Let me know if you have any problems.
|
|
|
|
|
|
|
| |
Clean up a couple of warnings...
Tested on:
Linux/64 2.6 (chicago2)
|
|
|
|
|
|
|
|
| |
Rename new H5Gcopy() routine to H5Ocopy() as discussed in last Friday's
design discussion.
Tested on:
Linux/32 2.6 (chicago)
|
|
|
|
|
|
|
|
| |
Changed H5Lget_linkval() to H5Lget_val(), per our design discussion last
Friday afternoon.
Tested on:
Linux/32 2.6 (chicago)
|
|
|
|
|
|
|
|
|
| |
Added a framework to display information about a particular object.
This option (-O object_name) is not available to the users yet.
Currently only name of an object (or objects) is displayed.
Platforms tested:
sol, kagiso and copper.
|
|
|
|
|
|
|
|
|
| |
Change H5Lget_linkinfo() to H5Lget_info() to better align with coming API
changes.
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
|