Harald Bräuning, Ludwig Hechler, Vitaliy Rapp, Solveigh Matthies, Alexander Schwinn (Protokoll)
Topics
1.0.0 release experiences/status
According to the users, there is are no major issues with the latest version. Class development works fine.
Harald Braüning reported a bug which is related to multible use of GSI-specific fields.
Cosylab will start to use 1.0.0 when they have the usability "ok" of Harald Bräuning
When doing the next major release we will introduce two weeks of "commissioning" on the central development-envoronment(asl73x) before delivery to other systems. This will allow us to fix crucial release-bugs on the fly for two weeks, without increasing the version-number.
CMW-Installation RPM/script
Currently there are two possibilities to install CMW ( See attachment )
Via the RPM-Installer ( global-installation)
Dependency-links between the different cmw-packages are located in the installation-directory of fesa (via softlinks)
Via a script from Vitaliy Rapp ( local-installation )
Each cmw-package defines it's dependencies to other packages in the lib/include folder (via softlinks)
The current global installation makes it difficult to link e.g. cmw-clients ( or later UNIKOS-applications). Since they dont depend on any fesa-version, they should not need to use the links in the fesa-directory. However if they want to use the /opt/cmw directly, they dont know which versions of the cmw-packages belong together in the current global-installation.
The FE-Int team agreed that we have to provide the softlinks in the cmw-folders as well in the global-installation of cmw for the next release. (like in the local-installation)
The dependency-links between the different cmw-packages in the installation-directory of fesa probably can be removed.
FE-Int meeting Timing-Interface
Together with Dietrich Beck we will discuss the FESA - API of the Timing receiver (Which properties, etc. ).
Alexander Schwinn will make some proposals, in order to have a base for the discussion.
The date for this will be the 14.08.
Implementation of the Beam-Process concept in FESA
Some work here has to be done from the Timing-Team ( How to mask the data of the timing-events )
For the next FE-Int meeting Alexander Schwinn will provide some proposals, how the FESA-clyce-selector for clients (when using properties) could look like
SCU's with 64Bit processor
According to Vitaliy Rapp and Ludwig Hechler, the hardware team does some effort to provide SCU's which run a 64-Bit processor.
Ludwig Hechler will ask if the 64-Bit SCU's already will be usable for the MCS