| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
| |
Refactored free space manager to use metadata cache for serialized free
space sections. This speeds up the fractal heap test considerably...
Tested:
FreeBSD 4.11 (sleipnir)
Linux 2.4/32 (chicago)
Linux 2.4/64 (mir)
Mac OS X (amazon)
|
|
|
|
|
|
| |
the DT from
dying again.
|
|
|
|
|
|
|
| |
Added command files to compile and test h5copy tool.
Platfroms tested:
VMS server
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
using
the $srcdir properly. It is not right to chdir into testfiles and write
files there because in real srcdir mode, one should not changes things
in the srcdir area which could be shared by multiple builds simultanteously.
Solution: added the proper $srcdir components to the source file name.
Also clean up the indentation by cb.
Tested: only by hand in heping.
|
|
|
|
|
|
|
|
|
| |
ttsafe test failed on the tts_error call. The error was due to the outdated
expected error stack; fixed.
Test was also moved to use new error APIs.
Platforms tested: Linux 2.4 (heping) and Solaris 2.9 (shanti)
|
|
|
|
|
|
| |
added a call to h5tools_close
tested: linux
|
|
|
|
|
|
|
|
|
|
| |
"make check-vfd" will now run all tests in the test directory with different
file drivers (at least, all of those tests that use the testing framework's
FAPL). Tests that fail will be skipped.
This is not a perfect fix, but is better than nothing.
Along with this change, check-vfd should be added to the Daily Tests.
|
|
|
|
|
|
|
| |
1. changed the -F flag option names to "BE and "LE" for big and little endian
2. added a more verbose usage message for these options
3. add a new test
4. add a make clean instruction to *.bin
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Fix the last scattered bunch of problems with the object deletion code,
which appears to be completely working now (for objects that are stored
within heap blocks - standalone objects aren't implemented yet).
Also, re-work the regression test to speed up some of the existing tests
and add in 100-200 more combinations of tests - overall, its probably even
slower than it was... :-/
Tested:
FreeBSD 4.11 (sleipnir)
Linux 2.4 (chicago)
Mac OS X (amazon)
|
|
|
|
|
|
| |
updated the usage message with new info for the binary output option
tested: linux
|
|
|
|
|
|
| |
the tests.
This command script cannot tell reliably if tests passed. More work is needed.
|
|
|
|
|
| |
new feature
h5dump binary output. use little and big endian type conversions
|
| |
|
|
|
|
|
|
| |
build the tool.
Test command file is not quite working yet.
|
|
|
|
| |
test: linux
|
|
|
|
| |
All tests passed on the VMS server.
|
| |
|
|
|
|
| |
Now delete them. Sorry for the mistake.
|
|
|
|
|
|
|
|
|
|
|
|
| |
bug fix
calling h5tools_get_fapl running the mpio driver caused H5FD_pl_copy to
fail silently in some conditions. a MPI call was made before MPI_Init.
solution: corrected the MPI call to be made after MPI_Init and added error return
conditions to H5FD_pl_copy and h5tools_get_fapl
tested on copper parallel, mir, shanti
|
|
|
|
|
|
| |
This will generate h5copy on windows.
The previous check-in is under the wrong directory.
|
|
|
|
| |
This will generate h5copy on windows.
|
|
|
|
|
|
|
|
|
|
| |
h5ls uses -S flag to display datatype in a machine-readable form. Since
VMS is not a case sensitive system, triple quotes around S are needed; fixed.
h5ls -w80 -v -S tattr2.h5
test still reports failure due to line 311 in the tattr2.ls file "Modified: XXXX..."
Document it for now in the command file.
|
| |
|
|
|
|
|
|
| |
Updating all.dsw,all.sln(for command-line testing only) and binread.dsp.
New projects have been added into.
|
|
|
|
|
|
|
|
|
|
| |
except for
the last one
h5ls -w80 -v -S tattr2.h5
h5ls prints "native" for all types instead of the complete datatype definition.
I will look into the problem next.
|
|
|
|
|
|
| |
included.
Forgetting updating MANIFEST may be the reason why alpha release doesn't include binread.dsp and causes the problem on windows.
|
|
|
|
|
|
|
|
|
|
| |
Checkpoint further work on object deletion code for fractal heaps. I
believe everything is finally working for this sub-task. More regression tests
soon, but I don't foresee any problems.
Machines:
FreeBSD 4.11 (sleipnir)
Linux 2.4 (chicago)
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Refactored object deletion code for fractal heap to fix scaling problems.
Passing all previous tests again now.
Tested:
FreeBSD 4.11 (sleipnir)
Linux 2.4 (chicago)
Mac OS X (amazon)
|
|
|
|
| |
commented h5copy calls that are failing with mpicc compiler
|
| |
|
| |
|
|
|
|
| |
updated some comments
|
|
|
|
| |
tested: shanti, mir
|
|
|
|
|
|
|
|
|
|
| |
Upgrade
Description:
chnaged it from doing CVS-managed files to SVN-managed files.
(The use of ed, is real ugly, but ed exists in all Unix systems.
May want to consider using awk which is a more powerful and will have
a cleaner implementation.)
|
|
|
|
|
|
| |
his line, and those below, will be ignored--
M h5copy/h5copy.c
|
|
|
|
|
|
|
|
| |
it uses the API function H5Gcopy to copy groups and datasets from file
to file, read from a command line
note: the test script is not finished yet
tested on heping, copper, shanti
|
|
|
|
|
|
| |
hdf5-1.8.0-alpha4 release
with "hdf5_1_8_0_alpha4" tag.
|
|
|
|
| |
hdf5-1.8.0-alpha4
|
|
|
|
| |
the libraary.
|
|
|
|
|
|
|
|
| |
Clean up some compiler warnings.
Tested on:
FreeBSD 4.11 (sleipnir)
Too minor to require h5committest
|
|
|
|
| |
file from "h5dump -b".
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Updates on windows
Tested with
VS C/C++ 6.0 and Compaq fortran 6.6c
VS .Net 2003
Because
1. New source files added to the library
2. New program to test "h5dump -b" option
all.dsw and all.sln need to be updated. all.sln is used only for hdf5bt_net.bat and hdf5bt_intelc.bat.
The previous check-in forgot including the dependency of a new project file.
|
|
|
|
|
|
|
|
|
|
| |
Updates on windows
Some new source files are added to support new implementation of external links, hdf5dll.dsp needs to updated.
Tested with
VS C/C++ 6.0 and Compaq fortran 6.6c
VS .Net 2003
|
|
|
|
|
|
|
|
| |
Some new source files are added to support new implementation of external links, hdf5.dsp needs to updated.
Tested with
VS C/C++ 6.0 and Compaq fortran 6.6c
VS .Net 2003
|
|
|
|
|
|
|
|
| |
Because
1. New source files added to the library
2. New program to test "h5dump -b" option
all.dsw and all.sln need to be updated. all.sln is used only for hdf5bt_net.bat and hdf5bt_intelc.bat.
|
|
|
|
| |
clash on Windows
|
|
|
|
|
|
|
|
| |
with merged
object header messages properly.
Too minor to require h5committest
|
|
|
|
|
|
| |
be "CRLF"
instead of "native".
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
1) Added trace file support to the metadata cache. This allows capture
of all metadata cache calls in trace files for purposes of optimization
and debuging.
2) Added an expunge entry function. This allows an entry to be deleteded
from the cache without writing it to disk even if it is dirty.
3) Added a function call to resize pinned entries.
4) Added code to deal with entries that are dirty on load. This is
needed in support of a bug fix which can alter object headers on
load to repair files.
5) Added progress reporting code to the "MDC API smoke check" test in
cache_api.c. To enable the progress reporting, set report_progress
to TRUE in mdc_api_call_smoke_check().
Tested with h5committest, and a parallel test on phoenix (dual athelon
linux box).
|