Remove LDOUBLE_TO_INTEGER_WORKS macro/define from configure/library/tests,
since it was added to support non-standard behavior on SGI IRIX & HP-UX, which
we don't support any longer.
Tested on:
Mac OSX/64 10.10.2 (amazon) w/serial & parallel
(h5committest not required on this branch)
Remove the LDOUBLE_TO_INTEGER_ACCURATE macro, since it was added to work
around non-standard behavior on SGI IRIX, which we are no longer supporting.
Tested on:
Mac OSX/64 10.10.2 (amazon) w/serial
(h5committest not required on this branch)
unsupported Code Warrior compiler. Also removes the H5_NO_SHARED_WRITING
symbol from the pubconf file.
Fixes: HDFFV-9080
Tested on: jam only, after bin/reconfigure
(all platforms should see no source changes)
- Remove Infering parallel compilers (C and Fortran) from configure.ac
- Remove restriction to build shared with parallel
- Cleanup parallel sections in configure.ac
- remove large file support checks
- MPE fixes.
tested with h5committest.
Developers will now have to run bin/reconfigure (deprecated, will be removed
soon) or autogen.sh after checking out code.
Part of: HDFFV-9120
Tested on: jam with Fortran and C++
Change AC_TRY_RUN TO AC_TRY_COMPILE
Changed all the instances of AC_TRY_RUN for the Fortran and C++ tests to use
AC_COMPILE_IFELSE and moved all the tests to new files in m4 directory:
aclocal_cxx.m4 (for C++ tests)
aclocal_fc.m4 (for Fortran tests)
tested: jam
CodeWarrior as a work-around to broken open() functionality (it's not clear
which version). The only effect of this #define was to cause test_file_close()
in test/tfile.c to be skipped.
Fixes HDFFV-9080
Tested on: jam
We don't have CodeWarrior, so I just ensured the deletion didn't adversely
affect a typical build.
H5O_MSG_FLAG_FAIL_IF_UNKNOWN_AND_OPEN_FOR_WRITE. Will add a new
H5O_MSG_FLAG_FAIL_IF_UNKNOWN_ALWAYS with the new avoid_truncate patch
coming later.
tested h5committest
Revert r25273, 25283 & 25439 (the hyperslab improvement changes). They
are buggy and it's taking me a long time to correct the problem. I'll check
in a revised form of the changes when I've got them straightened out.
Tested on:
Mac OSX 10.10.0 (amazon) w/gcc 4.9.x, C++, FORTRAN
Linux 2.6.x (jam) w/parallel
Merge 64-bit ID changes from branch to trunk. (Plus a few minor cleanups
that aren't on the branch)
Tested on:
Mac OSX/64 10.9.4 (amazon) w/C++ & FORTRAN
(h5committested on branch already for a week)
testfrome.c has incorrectly exit code (using -1). Fixed by replacing them
with C defined EXIT_FAILURE. Verified by temporary decreaing MAXNUMTESTS
to introduce failure in testhdf to see the changed code did work.
Also replaced all other correctly coded exit codes with the C defined
EXIT_SUCCESS and EXIT_FAILURE. This would help portability for non-Unix
systems.
Tested: h5committest.
Correct error in new hyperslab selection test to use the right # of
dimensions for start/stride/count/block arrays. Also, switch H5Sselect_select
to H5Smodify_select.
Tested on:
MacOSX/64 10.9.3 (amazon) w/gcc 4.9.1, OpenMPI, C++, FORTRAN & threadsafe
(Too minor to require h5committest)
Bring in Chao/Neil/my changes to optimize hyperslab selection operations
further, along with 3 new public API routines: H5Scombine_hyperslab(),
H5Sselect_select() and H5Scombine_select(), along with many minor cleanups to
the code and fixing a few compiler warnings.
Tested on:
Mac OSX/64 10.9.3 w/gcc 4.9.x and parallel w/OpenMPI
(h5commttest forthcoming)
ib files. The fix also removed dependencies on libhdf5, etc. when the --disable-sharedlib-rpath co
nfigure option was invoked.
Added instead configure variable hardcode_into_lib=no. This removes rpath from lib files on Linux
and solaris machines.
Tested with h5committest on jam, koala, ostrich and platypus (cmake), and with configure option --disable-sharedlib-rpath on emu, platypus and quail.
Make progress toward moving from DXPL IDs to property list structures
within the library. Also move the signature location code from the H5F
package to the H5FD package, where it's a better fit. Also, clean up some
more compiler warnings along the way.
Tested on:
Mac OSX/64 10.9.2 (amazon) w/C++, FORTRAN & parallel
(h5committest forthcoming)
More migration to using H5F_io_info_t pointers and away from using
property list IDs internally. Also, clean up some compiler warnings in the
cache code.
Tested on:
Mac OSX/64 10.9.2 (amazon) w/C++, FORTRAN & parallel
(h5committest forthcoming)
Begin process of migrating from using property list IDs internally to the
library to using the internal generic property list data structure.
Tested on:
Mac OSX/64 10.9.2 (amazon) w/C++, FORTRAN & parallel
(h5committest forthcoming)
Adds write tracking to the core VFD, which can be configured
via the H5Pset/get_core_write_tracking() API call. When enabled
and writing to the backing store is enabled, this feature will
track writes and only write out the changed bytes on flush/close.
Tested on:
32-bit LE linux (jam) w/ Fortran and C++
64-bit BE linux (ostrich)
64-bit Darwin 12.5.0 (kite)
All were tested using the core VFD w/ paging on (core_paged
in the VFD list). Make check-vfd was also tested on jam.
functions that return library-allocated memory. The test and tool
code were modified to use this new function where it's appropriate.
Fixes HDFFV-7710, HDFFV-8519, and HDFFV-8551
Tested on:
64-bit Windows 7 w/ VS2012
32-bit LE linux w/ parallel and fortran (jam)
32-bit LE linux w/ fortran and C++ (jam)
Clean up more compiler warnings, plus merge a few Coverity bug fixes from
the hdf5_1_8_coverity branch back to the trunk:
r20877:
Purpose: Fix coverity issue 1723
Description:
Modified test_generate in hl/test_image to close file "f" before exit, even if
an error occurs.
r20879:
Issue 63: change check of return of H5Tget_nmembers to <=0. No need to go
futher if call fails as well as empty.
r20881:
Coverity #659 in Run 46: I changed the Line 442 where it tries to check whether
FLAG_PRINTED is TRUE. But it had just been set to FALSE. I took out the
condition check in the print statement.
Tested on:
Mac OSX/64 10.9.2 (amazon) w/C++, FORTRAN & Parallel
(too minor to require h5committest)
Remove all traces of MPI-POSIX VFD and GPFS detection/code.
Remove remaining traces of stream VFD.
Remove testpar/t_posix_compliant test (it's not actually verifying anything).
Clean up H5D__mpio_opt_possible() further.
Moved environment variable that disables MPI collective operations into
MPI-IO VFD (instead of it being in src/H5S.c).
A few other small code cleanups.
Tested on:
Mac OSX/64 10.9.2 (amazon) w/parallel & serial
KitWare recommendations: Change CMake commands to lower case.
Bump up CMake minimum versions.
Split test file into separate file for h5repack test.
Tested: local linux
Brought changes from Coverity branch back to trunk, and cleaned up misc.
other warnings & formatting issues:
r20833:
Fixed Coverity 667 and 668 with real integer overflow tests this time.
r20834:
Use HDstrncpy and HDstrncat. --gh
r20835:
Change to use strncpy - use base_len + 1 for line 156, use HDstrlen(path) + 1 for line 159
r20836:
Fixed coverity 585 by casting output of fgetc() to a char.
r20837:
Changed sprintf calls to snprintf with size 1 less than the allocated buffer to address coverity issue #967.
Tested on:
Mac OSX/64 10.9.2 (amazon) w/C++, FORTRAN & parallel
(too minor to require h5committest)
Rename GCC_DIAG_OFF/ON macros to H5_GCC_DIAG_OFF/ON and move from
src/H5private.h to src/H5public.h. Wrap typedef of hsize_t and hssize_t
in DIAG_OFF(long-long) macros.
Clean up a bunch of "macro '-' is unused" warnings.
Tested on:
Mac OSX/64 10.9.2 (amazon) w/gcc 4.8.2, C++, FORTRAN, parallel
(too minor to require h5committest)
Revert some earlier usage of strncpy, which was incorrect.
Bring Coverity changes from branch back to trunk:
r20821:
Use HDstrncpy. --gh
(Fixed already, with strdup)
r20822:
(Not merged, incorrect use of HDstrncpy())
r20823:
(Not merged, incorrect use of HDstrncpy())
r20824:
Maintenance: Bug fix: addressed CID 666.
Value stored at *expression_len should be used in the call to
HD5packFstring to avoid overflow (and unnecessary arithmetic
calculation and casting)
r20825:
Issue 642: Added check for error and handler with print to stderr and exit.
r20826:
Undo revision 20818, as that issue has already been fixed in the 1.8 branch
and trunk (but not coverity branch)
r20827:
(Not merged, incorrect use of HDstrncpy())
r20828:
Use HDstrncpy. --gh
(Corrected use of strncpy())
r20829:
Check return of H5Lget_val(print_udata->fid, path, targbuf,
linfo->u.val_size + 1, H5P_DEFAULT) and if error set trgbuf[0] to 0.
Check if H5Lunpack_elink_val(targbuf, linfo->u.val_size, NULL,
&filename, &objname) was successful and allow print. Otherwise filename
and objname are not created. (init those to NULL)
r20830:
resolved coverity issues 939, 940, 941, 944, and 947.
all were complaints about use of sprintf, and in all cases, the
buffers used were large enough for all eventualities.
Resolved issue by replacing calls to sprintf with calls
to snprintf.
r20831:
Maintenance: Addressed CID 852
Replaced sprintf with snprintf
r20832:
Purpose: Fix valgrind issues with hl/examples/ex_image2
Description:
Modified hl/examples/ex_image2 to free global "gbuf" before exit.
Tested on:
Mac OSX/64 10.9.1 (amaon) w/C++, FORTRAN & Threadsafety
(too minor to require h5committest)
view objects. The addition of view objects in the fastforward project
is expected to be brough into the trunk sometimes in the future, which
is why we need to make this change.
Tested Manually on Jam and Ostrich.
Tested with h5commitest - Koala with intel compilers failed, but nothing had to do with those changes.
error on Koala: error while loading shared libraries: libirng.so
Fix for HDF5/HDFFV-8620 H5Rget_name with NULL name parameter fails.
The name parameter now excepts NULL, in which case the length of then
name is returned.
Tested: jam (intel and gnu)
Remove duplicate "intent" flags from top & shared file structures, leaving
them only in the shared file structure.
Tested on:
Mac OSX/64 10.9.0 (amazon) w/gcc 4.8, both parallel & serial
(too minor to require h5committest)
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.
receiving review comments from people. I put the safeguard in both H5F_get_objects and H5F_get_objects_cb to prevent
overfill the list.
tested with h5committest.
The previous behavior was to execute a sec2, stdio, and family
test everytime it ran, which was very inefficient under check-vfd.
Tested on:
32- and 64-bit LE linux (jam and koala)
64-bit BE linux (ostrich)
64-bit OS-X 10.6 (fred)
The reason was typo in the index variable in the newly added test_get_obj_ids().
Solution: fixed the typo.
Tested: AIX machine. (Did not run h5committest because the fix is in a remote
machine and the fix is quite obvious.)
Description:
There is a bug in the current big.c tests. In a nut-shell,
the writing part randomly picks 50 locations and write 4KB
from that location. Random location range is from byte 0
to end of dataset. There is the problem--if the random location
is greater than (end of dataset - 4KB), then error.
Solution:
Adjust the random number to be at least 4KB less than the end of the
dataset. That allows it write 4KB more from the generated starting
position.
Tested: h5comittest plus jam serial build and test.
Clean up warnings, enable new compiler warning flag(s) and bring back
changes from Coverity branch:
r20813:
Remove the dead code as listed for coverity bug #1722. h5committested.
r20814:
Issue 69: Check return value and throw error if negative return. Also free datatype id on error
r20815:
Use HDstrncpy. --gh
r20816:
Replaced one last HDstrcat call with HDstrncat to resolve coverity issue 832.
r20817:
Use HDstrncpy and HDstrncat. --gh
r20818:
Purpose: Fix valgrind issues with h5jam
Description:
Modified h5jam to free strings strdup'd in parse_command_line before exit. Note
that they may still not be freed in case of error, due to the widespread use of
exit().
r20819:
Issue 80: change loop to use int as loop index.
r20820:
Maintenance: Fixed the bug found by coverity CID 788
There were two problems with this function:
1) it tried to unnecessary free NULL pointer
2) it tried to allocate c_name buffer that is done by H5Pget_class_name
Tested on:
Mac OSX 10.8.4 (amazon) w/gcc 4.8.1, C++ & FORTRAN
(too minor to require h5committest)
Changes applied to eliminate several compiler warnings in the test directory in
files:
dt_arith.c
filter_fail.c
big.c
dynlib2.c
dynlib3.c
Also an HDassert(0 && "Unknown type"); statement was added to all 29 combination
s of
case OTHER:
default:
See crucible review HDF5-145 for more details.
Tested with h5committest.
Clean up warnings, switch library code to use Standard C/POSIX wrapper
macros, remove internal calls to API routines, update checkapi and checkposix
scripts.
Tested on:
Mac OSX/64 10.8.3 (amazon) w/C++ & FORTRAN
Big-Endian Linux/64 (ostrich)
Clean up warnings in H5Tconv.c (down to _only_ 9000 lines of output now!)
Merge changes from Coverity branch back to trunk:
r20684:
Fix for coverity bug #1721 which was due to the fix for coverity bug #943.
r20685:
Use HDstrncpy. --gh
r20761:
Purpose: Fix valgrind issues
Description:
Free image_data and data as appropriate in test_image.
r20762:
Purpose: Fix coverity issue 600
Description:
Add check for return value of H5O_close in H5Ocopy. Also cleaned up various
warnings.
r20763:
Purpose: Fix valgrind issues with h5stat
Description:
Modified h5stat to free "iter" before exit, and free "hand" before exit if
parse_command_line exits directly.
r20764:
fixed coverity issues:
69, 327, 614, 684, 685, 696, 697, 1681, 967, 826, 660, 80
r20765:
Fixed coverity bug 668
Pulled x * y * z multiply out of malloc operand into a separate n_elements variable to quiet integer overflow warning.
No actual integer overflow tests are performed since it's just a test program but I did add a check that n_elements is >= 1.
I also changed an error condition from doing its own close and returning -1 to "goto out;" like the rest of the program.
r20766:
Fixed coverity bug 667
Pulled x * y * z multiply out of malloc operand into a separate n_elements variable to quiet integer overflow warning.
No actual integer overflow tests are performed since it's just a test program.
Tested on:
Mac OSX/64 10.8.3 (amazon) w/C++ & FORTRAN
(h5committest upcoming)
reviewed my previous commit and suggested me to use H5I_iterate to iterate through
all opened objects in H5Z_unregister. I changed it in this commit.
Tested with h5committest.
Stop aliasing property to indicate internal collective metadata operations
with property to perform collective raw data operations from the application.
Tested on:
Mac OSX/64 10.8.3 (amazon) w/paralllel
to copy library file names in the form of libdynlib{123].*.
Another problem: Even when copy failed in copying the libray files (was in the form of *.so*),
the test still passed. That was because .libs was included in $HDF5_PLUGIN_PATH and the needed
plug in libraries were generated there.
Fixed by copying the plug in library files to separate directories and seetup HDF5_PLUGIN_PATH
to include them but NOT .libs.
Tested: used desycommittest (duck, emu, jam, koala, ostrich) plus cmakehdf5 in jam.
The cmakehdf5 in jam reported a failure in
Test #594: H5REPART-h5repart_test ....................................***Failed
Since this failure could not be caused by my change in test_plugin.sh, I reported it to Allen
and proceed to check this in.
Test was not checking error result. It basically return succeess to make check
all the time.
Fixed it so that it does check the return code of the test (plugin) and set
nerror accordingly. It then exit 1 if there is any error detected.
Test: use the desy committest to pass on jam, koala, ostrich, duck and emu.
Also hand test by "rm test/.lib/libdynlib1* to verify the test script can
indeed response to errors properly.
Some systems, like Mac, strings inspect library files and older versions
of strings maybe know newer library format, resulting in errors.
Make it read the file as stdin, avoiding this problem.
Tested: h5committest. all passed.
The ERROR macro had a typo of missing a separator comma. When used, it seg.
fault. Added the comma back in.
Fix is reviewed in Crucible HDF5-107.
Tested: h5committest passed. Also tested using the ERROR macro in the
Mac (Owl) platform. Before fix, it seg. fault. After fix, it prints the
message as intended.
Description:
Adds 2 new API functions, H5Dscatter and H5Dgather. H5Dscatter retrieves data
from a specified callback function and scatters it into a selection, defined by
a supplied dataspace, within a supplied memory buffer. H5Dgather gathers data
from a selection within a supplied memory buffer and passes it in a contiguous
form to a supplied callback function. Added tests for these functions
Tested: jam, ostrich, koala (h5committest); ummon
Description:
Adds 2 new API functions, H5Dscatter and H5Dgather. H5Dscatter retrieves data
from a specified callback function and scatters it into a selection, defined by
a supplied dataspace, within a supplied memory buffer. H5Dgather gathers data
from a selection within a supplied memory buffer and passes it in a contiguous
form to a supplied callback function. Added tests for these functions
Tested: jam, ostrich, koala (h5committest); ummon
Bring reviewed changes from Coverity branch back to trunk (QK & JK):
r20457:
Coverity issue 691: return of H5duo could be negative. Fixed by using
STDOUT_FILENO and redesign parse_command_line and main to cleanup file
allocations. The output_file var is null when using stdout. In cleanup do not
close output_file if NULL.
r20510:
Initialize ufid = -1 and predicate HDclose call on ufid != -1
r20511:
Purpose: Fix coverity issue 1715
Description: Free "file" and nested data on failure in H5FD_core_open.
r20512:
Initialize ifid = -1 and predicate HDclose call on ifid != -1
r20514:
Initialize h5fid = -1 and predicate HDclose call on h5fid != -1
r20516:
Added else branch to the if (ret_value < 0) check.
r20522:
Addressed coverity issues 930-933, 850, 836, 835, 1307. All minor
potential buffer overwrite bugs, or coverity errors. Fixed by replacing
strcpy and sprintf with strncpy and snprintf.
r20523:
fixed coverity issues 68, 1120, 1116i
r20524:
Check H5Z_SZIP->encoder_present < 1 assuming 0 represents absence.
r20601:
Purpose: Fix coverity issues 1703-1705
Description: Modified the cleanup code in test_free in accum.c to reset
allocated buffers to NULL after they are freed, and modified the error cleanup
code to check if these buffers are NULL before freeing them. Also fixed some
unrelated warnings in accum.c.
r20602:
Use HDsnprintf and HDstrncat
r20603:
Purpose: Fix coverity issues 808-809
Description: Modified test_core in vfd.c to check the returns from malloc, and
keep track of whether points and check are allocated by setting them to NULL
when they are not. Added code to free points and check on error if they are
not NULL. Also fixed unrelated warnings in vfd.c.
r20604:
Use HDstrncpy.
r20605:
Use HDstrncpy and HDstrncat.
r20606:
Purpose: Fix coverity issue 807
Description: Modified long_compact in stab.c to keep track of whether objname is
allocated by setting it to NULL when it is not. Added code to free objname on
error if it is not NULL.
r20607:
Changed string function calls to use versions that specify the string length
to fix coverity issues 832 and 839.
Tested on:
Mac OSX/64 10.8.2 (amazon)
(Too minor to require h5committest)
Have free space manager use temporary address space for storing the
section info, until the file is flushed or closed.
Tested on:
Mac OSX/64 10.7.4 (amazon) w/debug, gcc 4.7.x, C++, FORTRAN & threadsafe
(h5committest forthcoming)
Changes resulting from Klocwork static analysis tool, from Mark Miller
@ LLNL (miller86@llnl.gov).
Tested on:
Mac OS X/64 10.7.4 (amazon) w/debug, C++ & FORTRAN, using gcc 4.7.x
(too minor to require h5committest)
Description:
When jumping out from between H5_BEGIN_TAG and H5_END_TAG macros using
HGOTO_ERROR or HGOTO_DONE, the previous metadata tag is not reset on the dxpl.
This could cause problems when, for example, calling H5Ocopy within an
H5Literate callback.
Added new HGOTO_ERROR_TAG and HGOTO_DONE_TAG macros which must be used in place
of the above between H5_BEGIN_TAG and H5_END_TAG.
Tested:
jam, koala, ostrich (h5committest), durandal
Refactor ID class registration to be more like how other "class"s in the
library are registered, and clean up compiler warnings.
Tested on:
Mac OSX/64 10.7.4 (amazon) w/debug
(Too minor to require h5comittest)
Copy the file image test file from the source directory to the build
directory, instead of trying to open the [in-memory] file R/W.
Also, clean up a few other test files that were left lying around from
other tests.
Tested on:
MacOSX/64 10.7.4 (amazon) w/debug
Too minor to require h5committest
(2) fix to extend into the aggregator based on threshold
(3) fix for the "flavor" logging problem
(4) fix for test failures due to the above fixes
(5) bug fix for h5stat when h5stat against a non-existing file
Clean up more FUNC_ENTER/FUNC_LEAVE macros and move H5D & H5T code toward
the final design (as exemplified by the H5EA & H5FA code).
Tested on:
Mac OSX/64 10.7.3 (amazon) w/debug & parallel
Correct several errors in fractal heap code: root indirect block was
getting pinned/protected more than once, "single" free space sections weren't
getting "re-parented" correctly when the heap transitioned between having a
root indirect block and a root direct block, and several related issues. Also
cleaned up some warnings in library/tests.
Tested on:
FreeBSD/32 8.2 (loyalty) w/gcc4.6, w/C++ & FORTRAN, in debug mode
FreeBSD/64 8.2 (freedom) w/gcc4.6, w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-amd64 2.6 (koala) w/Intel compilers, w/default API=1.6.x,
w/C++ & FORTRAN, in production mode
Mac OSX/64 10.7.3 (amazon) w/debug
Description:
When copying an object with attribute creation order tracked, the attribute
creation order was not copied correctly to the destination file, causing an
error if the creation order was also indexed (due to attempting to insert
duplicate keys) or incorrect creation orders otherwise. Fixed to copy the
creation order correctly.
Also fixed the attribute character set not being copied, and fixed an issue
where an attribute opened with H5Aopen (or similar, but not by_idx), from an
object using the latest format but without creation order being tracked, would
always report the creation order as 0 (and marked as valid).
Tested: jam, koala, ostrich (h5committest), durandal
Bring "merge committed datatypes during H5Ocopy" feature from branch to
trunk. (Also has some minor bugfixes with it)
Tested on:
Mac OSX/64 10.7.3 (amazon) w/debug
(h5committest coming up)
Minor code safety issue in test/fheap.c and whitespace in other files.
Tested on:
Mac OSX/64 10.7.3 (amazon) w/debug
(Too minor to require h5committest)
This occurs due to the istore test creating very large files on systems which
do not have POSIX-like sparse file semantics. The large amount of I/O causes
the test to run for a very long period of time.
The fix was to copy the "big" test's sparse file check and only run the
largest sparse file test when POSIX-like sparse file semantics are found.
Tested on:
jam (nfs)
ostrich (nfs)
loyalty (ufs)
64-bit linux VM (ext4)
64-bit OS-X Lion (hfs, detected as not POSIX sparse)
64-bit Windows 7 (NTFS, detected as not POSIX sparse)
The OS-X failure to pass the sparse check is likely an error and will be
entered as a new bug.
The parser mistaked "-" as substraction. I fixed it and also fixed another problem
with some special cases like 100-x and 2/x.
Tested on jam, koala, and ostrich.
Description:
When using the new object header format, it was possible for corruption to occur
if the first object header chunk changed size such that the lenght of the "chunk
0 size" field changed. This only occurred if there were messages that had not
been decoded. The original algorithm that changed the object header chunk size
marked all messages as dirty, causing those that had not been decoded to have
both the raw and native form invalidated. Changed the algorithm to avoid
marking messages dirty and added assertions to catch the case where messages
are dirtied without being decoded (or recently created) first.
Tested: jam, koala, ostrich (h5committest), durandal
Description:
When using the new object header format and adding an attribute with a size near
64K, it was possible for file corruption to occur. This happened only if the
first object header chunk was smaller than 256 bytes and then grew to larger
than 64K after the attribute was added.
Tested: ostrich, jam, koala (h5committest), durandal
Correct corner case for creating a contiguous dataset with a zero-sized
dataspace, when the allocation time is set to early.
Also clean up a few compiler warnings in the dataspace code.
Tested on:
Mac OSX/64 10.7.3 (amazon) w/debug & parallel
Remove some leftover uses of the __FUNCTION__ macro, replacing them with
FUNC macro, as used everywhere else.
Tested on:
Mac OSX/64 10.7.3 (amazon) w/debug, production & parallel
(too minor to require h5committest)
Add FUNC_ENTER macros for package-private routines and begin process of
switching package routines to use them. All H5G routines are currently
finished.
Tested on:
Mac OSX/64 10.7.3 (amazon) w/debug, production & parallel
Description:
When shrinking a chunked dataset, the library fills in the unused parts of
chunks that have been shrunk. The fill value buffer allocated for this purpose
had a maximum size of 1 MB, but the fill was performed in a single operation.
Therefore, if the amount of unused space in a chunk after being shrunk was
greater than 1 MB, the library would read off the end of the fill value buffer.
Changed the maximum fill buffer size to be equal to the chunk size.
Tested: durandal; jam, koala, heiwa (h5committest)
Don't check dataset storage size for compressed datasets with region
reference datatypes. (The address of the region reference type in the file
varies and affects the compressed size)
Tested on:
Mac OS X/32 10.7.2 (amazon) w/debug & production + check-vfd
Recalculate the size of destination attribute message when the source and
destination versions are different during an object copy operation.
(Jira: HDFF-7718)
Tested on:
Mac OS X/32 10.7.2 (amazon) w/debug
(h5committest upcoming)
Description:
Added new H5SL_TYPE_GENERIC skip list type, which uses void *'s as keys and a
client-supplied callback for key comparison. This was added to support the
upcoming "merge named datatype" feature for H5Ocopy, but may be used in other
places as well. Also added testing.
Also fixed a potential bug with the H5SL_TYPE_OBJ implementation, and added
testing for that.
Tested: jam, koala, heiwa (h5committest), durandal
Correct error in loading local heap prefix & data block from the file.
Sometimes the local heap's prefix could be loaded before the data block (e.g.
using H5Oget_info), but then when the data block was loaded later, the free
list information would get lost, causing the heap's size to grow larger than
necessary. This is Jira bug #HDFFV-7767
Tested on:
Mac OS X/32 10.7.2 (amazon) w/debug
(h5committest coming up)
Description:
H5Ocopy could get confused when copying a named datatype containing an
attribute which used that named datatype as its datatype. This happened
because H5Ocopy would recurse into the attribute's datatype before the object
the attribute was in was fully copied (i.e. before the "post-copy" pass).
Modified H5Ocopy to avoid recursing before the post-copy step in this case.
Required many changes, including to how non-committed shared messages are
copied.
Tested: jam, koala, heiwa (h5committest); durandal
Some machines, like LLNL udawn, a blue-gene machine, requires all executables,
be launched by some command like mpirun.
Solution:
Added $RUNSERIAL to launch the executable.
Tested: LLNL uDawn.
them is in the INIT_INTEGER macro definition in dt_arith.c. It complained about line 150 where it tried to subtract 1 from
the negative minimal value of "int", causing it to overflow (or underflow). So I revised the code to avoid it.
Tested on jam, koala, linew, and Mac OS Lion with CLANG compiler.
temp_point->l = (unsigned long long)((i * 100 + j * 1000) * n);
The value can overflow the signed int before being converted to unsigned long long. So I changed it to
temp_point->l = (unsigned long long)((i * 40 + j * 400) * n);
to keep it under the maximal value.
Tested on jam. Simple change.
Ran bin/reconfigure to update the Makefile.in in directories not part of the fortran directory check=in. Updates Makefile.in due to changes made in configure.in for the Fortran 2003 additions.
Tested on all platforms run under daily tests.
i_mask = ~((unsigned)~0 << (precision[0] + offset[0])) & ((unsigned)~0 << offset[0]);
in line 3680 left shifted a 32-bit integer for 32-bit. The result is undefined by C language. A user
discovered it using clang compiler with -fcatch-undefined-behavior option (see Issue 7674
in Jira). So I changed it in a funny way to avoid it.
Tested on jam, koala, and heiwa.
Further patches for Jira 7638, to ensure that H5D/Aget_type will
be attached to the correct top file. (This change to the library should
reverted when Jira 7638 is finally finished)
Tested on:
FreeBSD/32 8.2 (loyalty) w/gcc4.6, w/C++ & FORTRAN, in debug mode
FreeBSD/64 8.2 (freedom) w/gcc4.6, w/C++ & FORTRAN, in debug mode
Linux/32 2.6 (jam) w/PGI compilers, w/default API=1.8.x,
w/C++ & FORTRAN, w/threadsafe, in debug mode
Linux/64-amd64 2.6 (koala) w/Intel compilers, w/default API=1.6.x,
w/C++ & FORTRAN, in production mode
Solaris/32 2.10 (linew) w/deprecated symbols disabled, w/C++ & FORTRAN,
w/szip filter, w/threadsafe, in production mode
Linux/PPC 2.6 (heiwa) w/C++ & FORTRAN, w/threadsafe, in debug mode
Mac OS X/32 10.6.8 (amazon) in debug mode
Mac OS X/32 10.6.8 (amazon) w/C++ & FORTRAN, w/threadsafe,
in production mode
1. Quincey changed H5P_DATASET_ACCESS_DEFAULT to H5P_DEFAULT for the tests that call H5Rdereference. Inside H5R_dereference, I assigned H5P_DATASET_ACCESS_DEFAULT to OAPL_ID if it is H5P_DEFAULT.
2. I added the test for invalid OAPL_ID in trefer.c.
Tested on jam, koala, heiwa.
Switch from H5P_DATASET_ACCESS_DEFAULT to H5P_DEFAULT for calls to
H5Rdereference2().
Tested on:
Mac OS X/32 10.6.8 (amazon) w/debug
(too minor to require h5committest)
Check in ugly version of fix for Jira #7638. I'm going to leave the issue
open in Jira and when the library has been refactored to use shared file
pointers (instead of top file pointers) for file operations, I'll uncomment
the check in src/H5Oshared.c
Tested on:
Mac OS X/32 10.6.8 (amazon) w/debug
(too small for h5committest)
super block. We took it out in v1.8 library because it's meaningless
for the MULTI file. v1.8 library saves the EOA for the metadata file,
instead. But this caused some backward compatibility problem.
v1.8 library couldn't open the file created with v1.6 library. I
fixed the problem by checking the EOA value to detect the file
created with v1.6 library.
Tested on jam, koala, and heiwa.
