The HDF5 1.8.16-pre2 pre-release candidate has 2 changes from HDF 1.8.16-pre1:
* The default production mode has been switched from development to production.
* The three issues that prevented building HDF5 1.8.16-pre1 with Visual Studio 2015 were addressed; Visual Studio 2015 will build the pre2 release candidate.
Thank you to Marco Atzeri and Sean McBride who reported test results from Cygwin 64 bit and on Mac 10.8 - 10.10 and Windows 8. The Mac errors seem to involve flags that we don't test. A Jira issue will be created specifying the flags with which the tests fail or produce errors to be addressed as possible. Any patches to be considered will be welcomed.
If you have time to test this release candidate, we would greatly appreciate it. We try our best to test HDF5 on a variety of platforms and with multiple compilers, but there is always a system that we couldn't test on. Your feedback is critical. If there are no problems reported next week that will seriously impair the operation of HDF5 1.8.16, we will proceed with the final 1.8.16 release expected by November 15.
CMake fans will need CMake 3.1 or greater to build HDF5-1.8.16.
The shared library version numbers for lib.so.* files in HDF5-1.8.16-pre1 have been updated according to libtool's versioning system and the changes in symbols from HDF5-1.8.15-patch1. The API compatibility report is at this link: https://gamma.hdfgroup.org/ftp/pub/outgoing/hdf5/hdf5-1.8.16-pre2/interface_compatibility_report.
The effect on the numbers from HDF5-1.8.15-patch1 to HDF5-1.8.16 was to increase libhdf5.so.10.0.1 to libhdf5.so.10.1.0, libhdf5_cpp.so.10.0.1 to libhdf5_cpp.so.11.0.0, and all other wrapper library files from libhdf5*.so.10.0.1 to libhdf5*.so.10.0.2.
Your reports are invaluable and very much appreciated. Thank you and happy testing!
Larry
···
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Larry Knox The HDF Group http://hdfgroup.org
1800 So. Oak St., Suite 203, Champaign IL 61820
217.531.6122
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Did you also take a look at the dashboard warnings? Many of them are trivial to fix. If you're not interested in some of the warning flags, I can turn them off. It would be nice to get the dashboard into a green state so that it can serve as a tool to catch _new_ problems at least. I find having a dashboard like that is only useful if you keep it green.
Attached is a patch that fixes several warnings from the Rogue14 bot.
On Fri, 30 Oct 2015 20:12:51 +0000, Larry Knox said:
The Mac errors seem to involve flags that we don't test. A Jira issue
will be created specifying the flags with which the tests fail or
produce errors to be addressed as possible. Any patches to be
considered will be welcomed.
--
____________________________________________________________
Sean McBride, B. Eng sean@rogue-research.com
Rogue Research www.rogue-research.com
Mac Software Developer Montréal, Québec, Canada
The effect on the numbers from HDF5-1.8.15-patch1 to HDF5-1.8.16 was to
increase libhdf5.so.10.0.1 to libhdf5.so.10.1.0, libhdf5_cpp.so.10.0.1
to libhdf5_cpp.so.11.0.0, and all other wrapper library files from
libhdf5*.so.10.0.1 to libhdf5*.so.10.0.2.
Question:
- is it really needed to implement a new API bump for C++ interface ?
The continuous API bump forces to maintain several legacy version
of the library and/or to recompile every time the software depending
on the library.
Library API should be as stable as possible.
Only ImageMagick is worst than HDF5 for the packages that I am
managing for cygwin.
Thank you for the patch. We will do our best to include it into the release; it does look straightforward. We will also review the flags and will get back to you.
Thanks again for your contribution to release testing!
On Fri, 30 Oct 2015 20:12:51 +0000, Larry Knox said:
The Mac errors seem to involve flags that we don't test. A Jira issue
will be created specifying the flags with which the tests fail or
produce errors to be addressed as possible. Any patches to be
considered will be welcomed.
Did you also take a look at the dashboard warnings? Many of them are trivial to fix. If you're not interested in some of the warning flags, I can turn them off. It would be nice to get the dashboard into a green state so that it can serve as a tool to catch _new_ problems at least. I find having a dashboard like that is only useful if you keep it green.
Attached is a patch that fixes several warnings from the Rogue14 bot.
The global constants were changed to references to remove multiple memory leaks, therefore we need to bump the C++ interface number. I'm sorry for the inconvenience but it can't be avoided.
Thank you for your long-time support,
Binh-Minh
···
--------------------
The HDF Group
_______________________________________
From: Hdf-forum <hdf-forum-bounces@lists.hdfgroup.org> on behalf of Marco Atzeri <marco.atzeri@gmail.com>
Sent: Monday, November 02, 2015 12:11 PM
To: hdf-forum@lists.hdfgroup.org
Subject: Re: [Hdf-forum] HDF 1.8.16-pre2 release candidate is available for testing
On 30/10/2015 21:12, Larry Knox wrote:
Hello everyone,
A second pre-release candidate version of HDF5 1.8.16 is available for
testing and can be downloaded at the following link:
The effect on the numbers from HDF5-1.8.15-patch1 to HDF5-1.8.16 was to
increase libhdf5.so.10.0.1 to libhdf5.so.10.1.0, libhdf5_cpp.so.10.0.1
to libhdf5_cpp.so.11.0.0, and all other wrapper library files from
libhdf5*.so.10.0.1 to libhdf5*.so.10.0.2.
Question:
- is it really needed to implement a new API bump for C++ interface ?
The continuous API bump forces to maintain several legacy version
of the library and/or to recompile every time the software depending
on the library.
Library API should be as stable as possible.
Only ImageMagick is worst than HDF5 for the packages that I am
managing for cygwin.
