Hi,
I suppose you might be interested to see the results of backward binary
compatibility tests for libhdf5 [1]. I've got the request from Debian
maintainers to add this library to the tracker. Hope it may be useful
for you too.
I see the one very dangerous change in the libhdf5 binary interface
between 1.8.5 and 1.8.6 versions [2], but the soname was not bumped: the
value of H5T_ORDER_NONE member has been changed from 3 to 4 in the
H5T_order_t enumeration type (i.e. old apps execute H5T_ORDER_MIXED
instead of NONE).
BTW
The tests were performed by the ACC tool [3].
[1] Test results for HDF5
<http://linuxtesting.org/upstream-tracker/versions/hdf5.html>
[2] libhdf5: 1.8.5 to 1.8.6 binary compatibility report
<http://linuxtesting.org/upstream-tracker/compat_reports/hdf5/1.8.5_to_1.8.6/abi_compat_report.html>
[3] ABI Compliance Checker (ACC)
<http://ispras.linuxfoundation.org/index.php/ABI_compliance_checker>
···
--
Andrey Ponomarenko
Department for Operating Systems at ISPRAS
web: http://www.LinuxTesting.org
mail: aponomarenko@ispras.ru
Hi Andrey,
Thank you for reporting the problem.
Hi,
I suppose you might be interested to see the results of backward binary compatibility tests for libhdf5 [1]. I've got the request from Debian maintainers to add this library to the tracker. Hope it may be useful for you too.
I see the one very dangerous change in the libhdf5 binary interface between 1.8.5 and 1.8.6 versions [2], but the soname was not bumped: the value of H5T_ORDER_NONE member has been changed from 3 to 4 in the H5T_order_t enumeration type (i.e. old apps execute H5T_ORDER_MIXED instead of NONE).
Applications shouldn't use numbers, but you are correct about the soname version. Will fix.
Elena
···
On Apr 4, 2011, at 6:39 AM, Andrey Ponomarenko wrote:
BTW
The tests were performed by the ACC tool [3].
[1] Test results for HDF5
[2] libhdf5: 1.8.5 to 1.8.6 binary compatibility report
[3] ABI Compliance Checker (ACC)
--
Andrey Ponomarenko
Department for Operating Systems at ISPRAS
web: http://www.LinuxTesting.org
mail: aponomarenko@ispras.ru
_______________________________________________
Hdf-forum is for HDF software users discussion.
Hdf-forum@hdfgroup.org
http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
I see the one very dangerous change in the libhdf5 binary interface between
1.8.5 and 1.8.6 versions [2], but the soname was not bumped: the value of
H5T_ORDER_NONE member has been changed from 3 to 4 in the H5T_order_t
enumeration type (i.e. old apps execute H5T_ORDER_MIXED instead of NONE).
Applications shouldn't use numbers...
Even when the application is coded using "H5T_ORDER_NONE" the compiler
ultimately "uses numbers" in the binary. The situation described is
problematic when an application compiled against 1.8.5 headers uses
1.8.6 at runtime. This will happen when the shared library is
upgraded but the application is not recompiled.
Granted, HD5's big scary banner about using the wrong shared library
minor version provides some measure of safety. Still, pesky folks
like me who set HDF5_DISABLE_VERSION_CHECK are at risk.
- Rhys
I see the one very dangerous change in the libhdf5 binary interface between
1.8.5 and 1.8.6 versions [2], but the soname was not bumped: the value of
H5T_ORDER_NONE member has been changed from 3 to 4 in the H5T_order_t
enumeration type (i.e. old apps execute H5T_ORDER_MIXED instead of NONE).
Applications shouldn't use numbers...
Even when the application is coded using "H5T_ORDER_NONE" the compiler
ultimately "uses numbers" in the binary. The situation described is
problematic when an application compiled against 1.8.5 headers uses
1.8.6 at runtime. This will happen when the shared library is
upgraded but the application is not recompiled.
yes, you are correct. I see the problem.
Granted, HD5's big scary banner about using the wrong shared library
minor version provides some measure of safety. Still, pesky folks
like me who set HDF5_DISABLE_VERSION_CHECK are at risk.
Probably we should remove this one 
and be more careful with addressing new features in maintenance releases.
Elena
···
On Apr 5, 2011, at 11:14 AM, Rhys Ulerich wrote:
- Rhys
_______________________________________________
Hdf-forum is for HDF software users discussion.
Hdf-forum@hdfgroup.org
http://mail.hdfgroup.org/mailman/listinfo/hdf-forum_hdfgroup.org
