More progress on address bugs in combinations of shared/committed/unshared
components of shared/unshared attributes in compact/dense storage. (Yes, there
are a lot of combinations. :-)
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
Fix object header message code to respect the "do not share" flag, so
that committed datatypes used in attributes are handled properly.
Add more tests for atttributes, checking that they handle shared
attributes in compact & dense storage w/committed datatypes.
Tested on:
FreeBSD/32 6.1 (duty)
Add more tests for attributes with shared components.
Close memory leak in shared message table loading.
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
Add support & tests for using shared datatypes with shared & un-shared
attributes. Involves some fairly icky code to make the "copy on write"
paradigm for shared attributes work.
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
Add support for deleting deleting dense storage when no compact
storage will be used.
Change shared message's "get heap address" routine to return
address in parameter instead of return value, to allow better error
detection.
Start writing tests for shared attributes which use shared
components (datatypes or dataspaces)
Tested on:
Mac OS X/32 10.4.8 (amazon)
FreeBSD/32 6.1 (duty)
Add support for deleting all the attributes on an object, when they
are dense and possibly shared.
Also, add some testing routines, for better error detection.
Should fix daily test failures also...
Tested on:
FreeBSD/32 6.1 (duty)
Creating shared messages and then deleting them should return the file to its
original state.
Added a test for this.
Tested on Windows, smirom, and kagiso.
Minor edits on the way to testing writing to shared attributes in dense
storage.
Tested on:
Mac OS X/32 10.4.8 (amazon)
Too minor to require h5committest
as complete as it could be, so I may add to it later.
Fixed a bug in reference counting messages that are referenced by shared
messages (attribute datatypes and dataspaces).
Tested on mir, smirom, and Windows.
VMS maintenance: updated command file with the new source files;
sohm test was failing on VMS due to upper case/lower case letters in
constant and variable names (i.e. FILENAME vs. filename); modified the test
to avoid confusion.
Platforms tested:
VMS server, kagiso
Add more tests for checking that writing to shared attributes works
correctly. (Looks like it still has a bug when dense storage is used, which
I'll correct shortly).
Add testing routine to shared message code, to retrieve reference count for
a shared message.
Add fractal heap 'op' callback for efficiently computing the hash value of
an existing encoded message in the shared message fractal heap.
Various minor cleanups and warning placations.
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
Add support for opening attributes in dense and/or shared storage by
index.
Move routines for building and operating on tables of attributes into
separate source module.
Fix bug where reverting from "dense" to "compact" storage would 'unshare'
attributes.
Minor code cleanups, etc.
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
Push support for shared attributes into more dark corners of the "dense"
attribute storage support.
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
Add [quite] limited ability to update (ie. write) data for objects in
fractal heap. Limited to just updating objects in managed heap blocks (i.e.
not 'tiny' or 'huge' objects) and must be updated with data of the same length
as the object in the heap. Updating objects in compressed heaps does work
though [as long as the data isn't 'tiny' or 'huge'].
Needed for changing the data value or the name of an attribute that is
stored in dense or shared storage.
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
messages to shared messages and vice versa).
Extended objcopy test to test all "permutations" of shared/unshared and
new/old group format. If this proves to take a long time to test, I'll
go back and have it run fewer tests when HDF5_TEST_EXPRESS is enabled.
Tested on smirom, kagiso, copper, and Windows.
Bug fix.
Description:
Test failed when core and multi VFD is used.
SOlution:
Failure was because the test file was created with the specified VFD access
property but later reopened by default access. That won't work for VFDs such
as multi that produced files that are incompatible with the default sec2
file driver.
Fixed it by using the same VFD access property when it reopens the same test
file.
However, it still fails for the core VFD which does not create any real file.
It is meaningless to try to reopen its file. Fixed it by adding core to
the incompatible list of VFDs so that it would skip the test.
Split is already coded as incompatible VFD but I don't understand why
it is incompatible in this case. Removed it from the incompatible list.
Tested Platforms:
Copper in which the daily test failed.
Set up tests for next phase of testing dense attribute operations, with
shared attributes.
Other minor code cleanups...
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
Add "attribute exists" internal routine to make verifying that an attribute
with the same name doesn't already exist easier.
Tweak "trace" script to produce more whitespace in H5TRACE macros, in order
to make them easier to read.
Minor other whitespace cleanups
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
Switch from using H5L_index_t/H5L_INDEX_<foo> to H5_index_t/H5_INDEX_<foo>
in order to accommodate indices on aspects of attributes as well as links.
Add basic support for deleting attributes in dense storage (needs more
support/tests for shared attributes in dense storage still).
Misc. cleanups, etc.
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
Migrate "internalish" attribute operations into new source code file.
Add test & basic support for opening attributes in dense storage (shared
attributes not tested or supported yet).
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
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)
Add first pass of "dense" attribute storage to objects. Lots of parts of
this are stubbed out, but all the tests are passing and I'll work on the corner
cases soon.
Eliminated several unused parameters from object header message callback
routines.
Other, miscellaneous code cleanups, etc. (and probably some things I've
forgotten about... :-)
Tested on:
FreeBSD/32 4.11 (sleipnir)
Linux/32 2.4 (heping)
AIX/32 5.? (copper)
Finish moving object header message routines into their own source code
module, along with renaming them to have "H5O_msg_" prefix...
Tested on:
Mac OS X/32 10.4.8 (amazon)
FreeBSD/32 4.11 (sleipnir)
Linux/32 2.4 (heping)
AIX/32 5.? (copper)
Migrate more object header routines to use the H5O_msg_ prefix and put
them into the src/H5Omessage.c code module.
Tested on:
Mac OS X/32 10.4.8 (amazon)
FreeBSD/32 4.11 (sleipnir)
Linux/32 2.4 (heping)
AIX/32 5.? (copper)
Refactor object header code to separate process of creating an object
header message from the process of writing to an existing one.
Start renaming operations that deal with object header messages to have
"H5O_msg_" prefix...
Tested on:
Mac OS X/32 10.4.8 (amazon)
FreeBSD/32 4.11 (sleipnir)
Linux/32 2.4 (heping)
AIX/32 5.? (copper)
to make the test pass.
These changes involve changes to the file format of SOHMs, but that's okay
because nobody should have been using SOHMs yet anyway.
Tested on Windows, kagiso, copper, and heping
Fix problem with copying objects when the dataset to copy has an "older"
version of the layout message and needs to compute the size of the dataset's
storage after the dataspace information is available.
Minor random cleanups also...
Tested on:
FreeBSD/32 4.11 (sleipnir)
Too minor/weird to test w/h5committest...
Make the memory size of the max. compact & min. dense attribute storage
variables align with the property list size, instead of the size on disk, to
better handle big-endian machines.
Tested on:
FreeBSD/32 4.11 (sleipnir)
Linux/32 2.4 (heping)
Linux/64 2.4 (mir)
AIX/32 5.? (copper)
Propagate object creation properties up into group, dataset and named
datatype property lists, when those property lists are retrieved for
existing objects in a file.
Also, add H5Tget_create_plist() API routine, to allow named datatype
property lists to be retrieved for named datatypes.
Tested on:
FreeBSD/32 4.11 (sleipnir)
Linux/32 2.4 (heping)
Linux/64 2.4 (mir)
AIX/32 5.? (copper)
Add fields for tracking attributes on objects to object header prefix,
when using the latest version of the format.
Reduce size used to store type of header message from 2 bytes to 1. (If we
have >256 message types, we probably have a different problem... :-)
Also, add "birth" time for object (i.e. creation time) [named to align
with similar fields in file systems]
Correct bug in H5Gget_objinfo() - retrieve the ctime field from object info,
instead of the mtime field.
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
Reduce the size of some of the group information to more reasonable bounds.
(16-bit values instead of 32-bit values).
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
When using the latest version of the file format, move the "modification
time" information into the object header prefix, which is more efficient.
Also add "access time" and "change time" (for metadata) fields, all of which
take about the same space as the previous modification time header message.
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
General cleanups, in preparation for adding some attribute-tracking fields
to the object header prefix.
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
Add # of attributes to the object info struct and deprecate
H5Aget_num_attrs().
Tested on:
FreeBSd/32 4.11 (sleipnir)
Linux/32 2.4 (heping)
Linux/64 2.4 (mir)
AIX/32 5.? (copper)
Finish removing library's internal code that uses H5G_get_objinfo() and
retarget it at either getting the link information or the object information,
as appropriate. (Still need to add user-level tests for H5Oget_info(), but
since several internal components of the library depend on the internal version,
it appears to be working correctly).
Tested on:
FreeBSD/32 4.11 (sleipnir)
Linux/322.4 (heping)
Linux/64 2.4 (mir)
AIX/32 5.? (copper)
Mac OS X/32 10.4.8 (amazon)
Add tests for H5Oopen_by_idx() (in a rare fit of superlative coding I
apparently got the implementation complete correct in my earlier checkin :-)
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
Add tests for H5Literate(), in all combinations.
Sweep up a few minor issues with H5Literate() that the tests exposed.
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
Finish implementation of H5Literate() [still needs real tests]
Clean up datatype macro warnings a bit more.
Unify iterator callback macros and put up in public header file, with the
iterator directions.
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)
Clean up warnings assigned to me from today's meeting - still need
to tackle the datatype conversion ones...
Tested on:
FreeBSD/32 4.11 (sleipnir)
Linux/32 2.4 (heping)
Linux/64 2.4 (mir)
Add src/H5Oalloc.c file forgotten in previous object header checkin
Break out more object header routines into separate file (debugging routines
this time).
Fix "//" style comment in recent ISOHM source changes.
Fix 'size_t' vs. 'unsigned' problem (visible in 64-bit testing) in recent
ISOHM test changes.
Tested on:
FreeBSD/32 4.11 (sleipnir)
Linux/32 2.4 (heping)
Linux/64 2.4 (mir)
version and size information into the superblock to eliminate a read when
loading it.
This is a file format change, and hopefully the last one (knock on wood).
Tested on kagiso and Windows (mostly just a SOHM change).
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)
Added new refreg_name.c file to test H5Rget_name and H5Iget_name for region references.
H5Rget_name tests are skipped for now since they do not work (at least for me :-)
Platforms tested:
heping and copper
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.
Add support for "delete by index" to "old-style" groups, finishing
implmentation of H5Ldelete_by_idx() routine.
Tested on:
FreeBSD/32 4.11 (sleipnir)
Linux/32 2.4 (heping)
Linux/64 2.4 (mir)
Aix/32 5.? (copper)
More progress toward getting H5Ldelete_by_idx() working - removals on
densely stored links should work now (still needs some more tests). Still
need to write code for removals on "old-style" groups.
Tested on:
FreeBSD/32 4.11 (sleipnir)
More progress toward getting H5Ldelete_by_idx to work fully - can now
delete by index in compact group (but not dense groups or "old-style" groups
yet). Need to go write a v2 B-tree routine to delete from the B-tree by index
before the dense storage will work properly...
Tested on:
Linux/32 2.6 (chicago)
Linux/64 2.6 (chicago2)
Straighten out some convoluted code when links were being deleted, which
could cause the "delete" callback for user-defined links to not get called when
the group they were in was deleted.
Had to compromise on the "delete" callback though - only calls the callback
with the ID for the file the link is in, instead of the group, since the group
is being held open upstream in the calling sequence during a group deletion and
this prevents a group and its ID from being created. (This could possibly be
worked around, but would cause a fair bit of havoc in the code and I'm not
entirely certain it's worth it...)
Tested on:
Linux/32 2.6 (chicago)
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)
Finished implementation of H5Lget_info_by_idx for all cases: old vs. new
group formats, compact vs. dense new link storage, increasing vs. decreasing
vs. native iteration order.
Also, refactor symbol table "foo by index" routines to be more generic
and share more code by using a single B-tree iteration callback which makes
callbacks to a specific "get <foo>" callback.
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)
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.
Fix retrieving name for references to root group.
Also, move 'ref' test earlier in testing, right after 'getname' test
Tested on:
Linux/64 2.6 (chicago2)
Implement basic framework for H5Lget_info_by_idx and get it working for
creation order indices on compact groups.
Clean up code a bit.
Close resource link in user-defined link traversal.
Tested on:
Linux/32 2.6 (chicago)
Update H5Iget_name() tests to reflect that the name of the object for an
ID can be found in many more situations now.
Tested on:
FreeBSD/32 4.11 (sleipnir)
Refactor generic property list initialization code to put property list
specific routines in property list modules, instead of scattered to the four
winds. Also, introduce property list class initialization objects, to make
adding new property list classes in the library easier.
Fix daily test failure by using H5Pget_elink_prefix() API routine instead
of looking at the "raw" generic property list information.
Tested on:
Mac OS X/32 10.4.8 (amazon)
FreeBSD/32 4.11 (sleipnir) w/threadsafe
Linux/32 2.4 (heping) w/C++ & FORTRAN
Linux/64 2.4 (mir) w/build-all & 1.6 compat
