| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
Description:
Changed Skip list package to use a deterministic skip list. This allows the
skip list package to avoid calling rand() and srand(), even on machines without
rand_r(). There is no longer a p-value or maximum level for skip lists.
Tested: jam, smirom, linew (h5committest)
|
|
|
|
|
|
|
|
|
|
| |
#1538 (B2) Problems with Dim Scale APIs reported by Mathworks
ISSUE: When some functions are called with an invalid dimension index (for example, greater than the rank of the dataset) the function does not immediately fail, rather asserts down the code.
SOLUTION: added this check for H5DSdetach_scale, H5DSset_label, H5DSget_label (the other functions have this check) . H5DSiterate_scales also added a check for an invalid dimension scale index
TEST: added some test cases with calls to invalid indices
DOCS: Nothing added
STATUS: done for 1.9, 1.8. closed bug
tested: windows, linux
|
|
|
|
|
|
|
|
|
|
|
| |
Description:
Previously, H5Lcopy and H5Lmove would (through H5L_move) improperly apply the
"create intermediate groups" property to the source path traversal, and not the
destination. Fixed it to apply the property to the destination and not the
source. Also fixed H5Lcreate_ud to reject internal link classes without
throwing an assertion.
Tested: linew, jam, smirom (h5committtest)
|
|
|
|
|
|
|
|
|
| |
Fixed warnings from absoft's compiler for !DEC$ statements.
Solution: There should not be a space after !DEC$ statements, removed the spaces.
Platforms tested:
Jam with gcc and f95
|
|
|
|
|
|
|
|
| |
#1521 (B2) H5DSdetach_scale memory leak
ISSUE: Purify (Windows Visual Studio 6) complains of a memory leak in H5DSdetach_scale call regarding a H5Aread call (stack is H5A_read, H5T_convert, H5T_conv_vlen, H5T_vlen_seq_mem_write, H5MM_malloc).
SOLUTION: When a scale is detached from a dataset, the variable length structure length field is decreased in one entry. The associated pointer must be reallocated with the new length.
DOCS AND TEST: Nothing added
tested: linux
|
|
|
|
| |
fix compiler error on a missing "static" declaration
|
|
|
|
| |
Tested: Fedora 10 gcc
|
|
|
|
|
|
| |
Added new test to h5repack.bat.
Tested on XP64-VS2005--IVF91
|
|
|
|
|
|
|
|
| |
note : lots of weird warnings of the type
../../../../hdf5/tools/h5repack/h5repacktst.c:3919: warning: will never be executed
appear on the same code as 1.9 for the 1.8 branch
tested: linux
|
|
|
|
|
|
| |
is generated during build and no longer exists in the windows/src folder.
Tested: XP64 VS2005
|
|
|
|
|
|
|
|
| |
#1402 (B1) h5repack does not preserve creation order indexing.
ISSUE : h5repack does not handle group creation order flags.
ACTION: call H5P(g)(s)et_link_creation_order functions when handling groups, add new groups with these flags to the test generation program, and verify results in the test program.
TEST: in the test program, function that compares property lists, added code to verify groups
DOCS: Added the note to RELEASE.txt: "h5repack now handles group creation order. PVN - 2009/4/2 - 1402".
|
|
|
|
|
|
|
|
|
| |
ID was passed in. I put
some argument check in the internal function H5I_find_id and took out the assertion check.
I also removed the argument check in H5Iis_valid because it's in H5I_find_id now.
Tested on jam and linew - simple change.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
1. #1501 (B1) tools bug if dataset is larger than H5TOOLS_BUFSIZE limit.
ISSUE : the tools use the following formula to read by hyperslabs: hyperslab_size[i] = MIN( dim_size[i], H5TOOLS_BUFSIZE / datum_size) where H5TOOLS_BUFSIZE is a constant defined of 1024K. This is OK as long as the datum_size does not exceed 1024K, otherwise we have a hyperslab size of 0 (since 1024K/(greater than 1024K) = 0). This affects h5dump. h5repack, h5diff
SOLUTION: add a check for a 0 size and define as 1 if so.
TEST FOR H5DUMP: Defined a case in the h5dump test generator program of such a type (an array type of doubles with a large array dimension, that was the case the user reported). Since the written file commited in svn would be around 1024K, opted for not writing the data (the part of the code where the hyperslab is defined is executed, since h5dump always reads the files). Defined a macro WRITE_ARRAY to enable such writing if needed. Added a run on the h5dump shell script. Added 2 new files to svn: tools/testfiles/tarray8.ddl, tools/testfiles/tarray8.h5. NOTE: while doing this I thought of adding this dataset case to an existing file, but that would add the large array output to those files (the ddls). The issue is that the file list is increasing.
TEST FOR H5DIFF: for h5diff the check for reading by hyperslabs is H5TOOLS_MALLOCSIZE (128 * H5TOOLS_BUFSIZE) or 128 Mb. This makes it not possible to add such a file to svn, so used the same method as h5dump (only write the dataset if WRITE_ARRAY is defined). As opposed to h5dump, the hyperslab code is NOT executed when the dataset is empty (dataset is not read). Added the new dataset to existing files and shell run (tools/h5diff/testfiles/h5diff_dset1.h5 and tools/h5diff/testfiles/h5diff_dset2.h5 and output in tools/h5diff/testfiles/h5diff_80.txt).
TEST FOR H5REPACK: similar issue as h5diff with the difference that the hyperslab code is run. Added a run to the shell script (with a filter, otherwise the code uses H5Ocopy).
FURTHER ISSUES: the type in question ("double") has a different output cross platforms (e.g on liberty some garbage number is printed at some array locations)
SOLUTION: defined an "int" type for this test. However the printing of such an array has a bogus output at least in one platform (FreeBsd), so eliminated the test run altogether and filed a bug report on this
|
|
|
|
|
|
|
|
|
|
|
| |
Description:
Due to a bug in H5F_super_read, every time a file with a user block was opened
it would grow by the size of the user block. The bug has been fixed, and
comments have been added to clarify when an eoa address should be relative and
when it should be absolute.
Tested:
jam, linew (h5committest)
|
|
|
|
|
|
|
|
| |
necessary as Quincey pointed out.
I took it out and used the driver address instead.
Tested on jam - simple change.
|
|
|
|
|
|
|
|
|
|
|
|
| |
is opened and closed with
v1.8 library, v1.8 library writes the driver info block in the superblock. But v1.6
doesn't write it. This caused the data after the superblock to be overwritten. The
solution is to use a flag to indicate when the original file doesn't have the driver
info, v1.8 library doesn't write it either.
Tested on jam. I have some trouble to access smirom and linew using h5committest.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
#1047 (B1) The h5lt info routines are returning data in transposed form. ISSUE : the dimension array of the function h5ltget_dataset_info_f was not transposed (because of C-FORTRAN storage order, the FORTRAN function calls the C function with the same name). SOLUTION: transpose the array. ACTION: add a test case.
added the note
Fortran High-Level APIs:
------
- Lite: The h5ltget_dataset_info_f function (gets information about a dataset)
was not correctly returning the dimension array (PVN - 2009/3/23)
tested: linux
|
|
|
|
|
|
|
|
|
|
|
| |
Description:
Some files apparently exist in the wild which have corrupt symbol table messages
on the root group. These files can be opened by 1.6 (which uses the cached
information in the superblock) but not by 1.8. This patch fixes 1.8 and 1.9 so
they can now open these files, and will correct them if necessary. Also fix
some potential (rare) problems with array datatype versions.
Tested: jam (parallel; h5committest not working on linew or smirom)
|
|
|
|
|
|
| |
1. #1487 (B1) DS memory leaks ISSUE 1: On a "go to" error condition, previously allocated buffers were not freed. NOTE: these are "potential" memory leaks because typically the error conditions do not occur, so the potential memory leaks also do NOT occur. ISSUE 2: A function used to read dimension scales realistic data (topography of the North Atlantic, latitude and longitude) was being called without the data buffer being freed. SOLUTION FOR ISSUE 1: added "free" calls for the allocated buffers on the error sections. SOLUTION FOR ISSUE 2: added a "free" call after the read_data function, after using the buffer (on a H5Dwrite and H5Screate_simple). The read_data function reads both data and dimensions from the ASCII data files. DOCS: no docs added. Done for 1.9 and 1.8 STATUS: Closed bug
tested: h5committest
|
|
|
|
| |
regarding a AIX failure
|
| |
|
|
|
|
| |
Actually set the exit value for the test (*sigh*)
|
|
|
|
|
|
|
|
|
|
|
| |
Bring r16574 back from trunk:
Clean up many compiler warnings and make test return non-zero exit code
when a failure is detected.
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
(other platforms tested on trunk)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
was being selected.
When reading the compression parameter keyword, the compression type read flag was incorrectly set to read, removed this line of code
in->configOptionVector[COMPRESS] = 1;
Modified one configuration file to have the COMPRESSION-TYPE GZIP
Keyword.
Entered a bug description fix of
- h5import: By selecting a compression type, a big endian byte order was being
selected (PVN - 2009/11/3)
tested: linux
|
|
|
|
|
|
|
|
| |
Bring r16560 back from trunk.
Tested on:
Mac OS X/32 10.5.6 (amazon) debug & production
(Following up with tests on more platforms)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug Fix
Description:
Fixing BZ #1381. The --includedir=DIR configure option, which is used to
spceify installation location of C header files, did not work correctly as
the path was hard-coded into config/commence.am. I'm presuming this is
because an older version of automake didn't know where to put C header
files. In any case, removing this line now defaults the includedir to the
same directory that it is currently hard-coded to, and also fixes the
configure flag which allows for customization of this value.
Tested:
jam, liberty
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Bug Fix
Description:
Removing the code from configure which strips the '-g' flag from CFLAGS
when in production mode. The current default CFLAGS in production mode
does not include '-g', as intended, but we should allow users to
override this and enable '-g' by setting the CFLAGS environment
variable if desired. Note that this applies to FCFLAGS and CXXFLAGS
as well.
Tested:
kagiso, linew, liberty
|
|
|
|
|
|
|
|
|
| |
string ptr's belonging to buffer in H5DSset_label and H5DSget_label. Also added free of buffers in error section of both functions. Potential memory leaks may exist elsewhere, and this will not close the bug.
Tested:
h5committest
vista 32 VS2008
|
|
|
|
| |
since section.
|
|
|
|
| |
been added with r16490.
|
|
|
|
|
|
| |
H5G_dense_iterate
Tested: Fedora 10 (too minor for full committest)
|
|
|
|
|
|
|
|
|
|
| |
test_big_chunks_bypass_cache,
test the correctness of the data when the fill value is defined or not. The
library should let the chunks bypass the cache depending on the size of the
chunks and whether to write fill value to the chunks.
Tested on jam - simple change.
|
|
|
|
|
|
|
|
|
|
|
| |
Bring v16532 back from trunk to v1.8 branch:
Pass the chunk "user data" to H5D_chunk_unlock(), so that chunks with
an address already aren't reallocated.
Tested on:
FreeBSD/32 6.3 (duty) in debug mode
(h5committest not needed here, performed on trunk)
|
| |
|
|
|
|
|
|
|
|
|
| |
Description:
In some situations it was possible for the fill value to not be written to parts
of a chunked dataset, particularly when extending and/or shrinking. Prior to
the fix for the chunk cache (1015) these bugs would have been exceedingly rare.
Tested: jam, smirom, linew (h5committest)
|
|
|
|
|
|
|
| |
windows OS are always LE on all platforms, simply added le to the two locations where these files are used.
Tested:
Vista 32 VS2008
|
| |
|
|
|
|
|
|
| |
have 2 expected outputs for 2 h5ls runs depending if run on a big or little endian machine. Configure.in was modified to export a variable carrying endianess information to testh5ls.sh. This script then compares the current run with 2 expected outputs, one for a big-endian machine (linew was used to generate the output), other for little endian (jam was used to generate the output)
the way h5ls prints types, it starts searching for NATIVE types first. One solution would be h5ls not to detect these native types, using for example the same print datatype function that h5dump does, that would make the output look the same on all platforms ("32-bit little-endian integer" would be printed instead). Drawback, this "native" information would not be available. Other solution is to have not one but 2 expected outputs and make the shell script detect the endianess and compare with one output or other
tested: jam, linew
|
|
|
|
|
|
|
|
|
|
|
| |
#1404). When the index was set to
creation order in querry function but there's no creation order indexed in the file, the library tried
to build and sort a table of all links. To optimize it, let the library use the B-tree for names of
links.
Tested with h5committest.
|
|
|
|
|
|
|
| |
bug fix: TESTING2 macro introduced recently for the test suite (it adds an extra space at the beginning of output, for indentation) was already available for HL and caused compilation problems on AIX. Replaced the name with TESTING_2
tested: h5committest
|
|
|
|
| |
Platforms tested: Mac OSX and AIX (by Ed) (minor fix)
|
|
|
|
|
| |
bug fix: a comma accidently left on a conditional typedef was causing compilition error on AIX
tested: linux
|
|
|
|
|
|
|
|
|
|
| |
Bring r16499 back from trunk:
Cache chunk info for newly created chunk.
Tested on:
FreeBSD/32 6.3 (duty)
(Tests included in upcoming revise_chunks branch changes)
|
|
|
|
|
|
|
|
|
|
|
| |
Bring r16488 back from trunk:
Clean up code and eliminate resource leaks. Also avoid "null" I/O when
chunk doesn't exist and we can skip it.
Tested on:
Mac OS X/32 10.5.6 (amazon)
(too minor to require h5committest)
|
|
|
|
|
|
|
|
|
|
|
| |
Bring r16483 back from trunk:
Clean up (i.e. remove) more internal calls to H5E_clear_stack(), along with
some other minor code cleanups.
Tested on:
Mac OS X/32 10.5.6 (amazon)
(too minor to require h5committest)
|
|
|
|
| |
by rev 16490.
|
|
|
|
|
|
|
|
|
| |
replacing all instances with long long.
Tested:
h5comittest
fedora 10 x64
XP32, VNET
|