Fixed a bug (1822) in the H5DSdetach_scale function when 0 bytes
were allocated after the last reference to a dim. scale
was removed from the list of references in a VL element of the
DIMENSION_LIST attribute; modified the function to comply
with the Spec: DIMENSION_LIST attribute is deleted now when no
dimension scales left attached.
Platforms tested: jam with electric fence, amani amd linew
Add bug fixes for Tools.
Description:
Added list from bugzilla:
1726 - NPOESS: h5repack loses attributes for datasets of type H5T_REFERENCE
1814 - NPOESS: h5repack doesn't handle references to the groups as an
element of a dataset
1817 - h5copy fail to copy dangling link by specifying link path directly
1793 - h5ls on a non-existent file gives 0 return code
Tested:
not need, just text update.
Update RELEASE.txt
Description:
Remove outdated known issue (should have been removed
from 1.8.4, whoops) from RELEASE.txt
Tested:
doc only, none needed.
Fix memory leaks
Description
Added a routine to free memory which addresses a memory leak
when variable length strings are used as fill values.
Also added some minor tweaks to the H5I 'save ID structures' mechanic.
Tested:
h5committest and valgrind (on jam/amani) to confirm freed memory.
Added note about use of volatile for variables.
Added blurb to release notes about fixing strict-aliasing problems.
Tested on:
None - trivial change to comment.
Improve configure's large-file support control.
Description:
Modified configure to now attempt to add defines necessary for
supporting largefiles on all systems, instead of solely on linux. This
is in response to user requests to enable largefile support on Solaris
by default, as well as to give extra control on AIX (instead of just
jamming the necessary flag into the config files).
The old --enable-linux-lfs flag is being deprecated in favor of the
--enable-largefile flag (enabled by default), which can be used on all
platforms. --disable-linux-lfs can still be used to disable largefile
support (on linux) when the --enable-largefile flag is not specified.
On systems where large files cannot be supported in this manner,
configure will report as such.
Tested:
h5committest
AIX (NCSA's blue_print machine)
duty, liberty, and linew.
Autotool Upgrade / Libtool Bug Fix
Description:
Updated autotools referenced in bin/reconfigure as follows:
Autoconf 2.64 --> Autoconf 2.65
Automake 1.11 --> Automake 1.11.1
Libtool 2.2.6a --> Libtool 2.2.6b-mcg
The referenced libtool version is a custom version of 2.2.6b. It
has been tweaked to fix a bug in libtool that occurs
when using PGI 10.0 compilers. A check incorrectly categorizes
the C++ compiler as version 1.0 instead of 10.0, and the link
line is subsequently set up incorrectly and fails to compile.
A patch has been made available and will be included in the next
release of libtool, but in the meantime I've applied the patch to a custom
installation as indicated above. This bin/reconfigure now references
the custom installation, and the resulting configure script will correctly
categorize the PGI 10.0 C++ compiler.
Ran bin/reconfigure to regenerate configure and makefiles.
Tested:
jam and amani with PGI 10.0 compilers.
h5committest
Description:
Attempting to copy an object with NULL references (all bytes zero) with the
H5O_COPY_EXPAND_REFERENCE_FLAG flag set would cause a failure or an assertion
(depending on whether it was in debug mode). Changed copy routine to detect
NULL references (object and region) and avoid attempting to expand the reference
in this case.
Tested: jam, linew, amani (h5committest)
Fix Bug 1730 (in 1.8)
Description:
This is a fix for Bug 1730. Note that this fix is only going into 1.8,
as there is a more desired fix prepared for the trunk, but we don't
want that one in 1.8 for compatibility reasons.
This change adds a mechanism to the H5I interface to save returned
ID structures for immediate re-use if needed. This addresses a potential
performance issue by delaying the case when the next ID to be
registered has grown so large that it wraps around and needs to be
checked to see if it is available for distribution again.
Tested:
h5committest, and manually on jam, amani, liberty, others.
Description:
Modified algorithm for extending a dataset with early allocation so it only
deals with the new chunks. Formerly, it would loop over all chunks, checking
to see if each existed in cache and on disk, causing major performance issues
with large numbers of chunks.
Tested: jam, linew, amani (h5committest)
Description:
The core file driver previously matched files only based on the name used to
open them, even when a backing store is used. This caused problems in the
failing test when it traversed an external link to itself, and did not recognize
that it was the same file. Similar tests in links.c were skipped when the core
file driver was used. The failure was previously masked by inappropriate file
flushes caused by external link traversal. Added code to the core file driver
to keep track of low level file information of the backings store to allow for
accurate recognition of duplicate files. Also re-enabled all links.c tests for
the core file driver.
Tested: Fedora (core and sec2 drivers; no time for full commit test, will
monitor daily tests)
Description:
Support for expanding external links was not implemented in H5Ocopy, even though
a flag existed for H5Pset_copy_object to enable this. This patch implements
this feature.
Tested: jam, amani, linew (h5committest); Fedora
Description:
Previously, the library would reopen the source file when traversing an external
link if unable to find the target file otherwise. This has been corrected.
Also moved the call to H5F_flush from H5F_try_close to H5F_dest, so the file is
only flushed when the last identifier for the file is closed. This prevernts
situations where the library could attempt to flush a file with protected
metadata.
Tested: jam, amani, linew (h5committest); Fedora
Work for bugzilla 1754: h5diff: support comparing through links.
Description:
merged from hdf5 trunk (18164, 18165, 18166, 18167, 18170, 18266, 18273)
Tested:
h5committest (jam, amani and linew)
Description:
In certain cases, removal of an object in a v1 b-tree would cause the leftmost
key in the right neighbor to be overwritten. While this did not pose a problem
for group b-trees, with chunked datasets it would overwrite the offset value
of the neighbor's leftmost child, causing corruption. Reworked the code to
differentiate between b-trees whose children are fundamentally associated with
their left key and those who are associated with their right key.
Tested: jam, linew, amani (h5committest)
Bring r17939 from trunk to 1.8 branch:
Handle external links from symlinked files by adding another check to
look for "child" files for links from the actual location of the "parent" file,
instead of from the location of the symlink.
Tested on:
FreeBSD/32 6.3 (duty)
(h5committested on trunk)
Fix Bug1672 - Display garbage value on LE machine for BE data.
Description:
Casuing by calling diff_dataset twice when -r or no option was given.
Change to call it once which fix the problem.
It also improves the performance. (diffing time in half)
According further test, it also occurred on BE machine as well and it
seems occruing only with the BE data attached to Bugzilla.
Don't know how the file was created. Anyway this fix will prevent
from diffing with garbage values in similiar potential case.
Tested on:
linux32 (jam) , linux64 (almani), solaris (linew)
