Experimental Physics and
Industrial Control System

Mark Rivers <[email protected]> · Thu, 4 Oct 2018 12:38:48 +0000

Folks,

I decided that the only way to really find problems with EPICS 7 was to put it in production on our 4 beamlines at the APS.  This includes 7 VME/vxWorks IOCs and roughly 70 IOCs running on Linux and Windows, including ~30 areaDetectors.

I'm happy to say that it is working.  There were a few issues that we have encountered so far and that have now been fixed.

1) There was a bug in the soft motor device support that was exposed when switching to EPICS 7.  This is fixed on the master branch of motor.

2) There was a bug in the mbbi and mbbo records in EPICS 7 that prevented the *VL fields from being modified at run-time.  This is fixed in the 7.0 branch of epics-base.

3) There was a problem with autosave that was exposed with a change in dbAccess.c::dbGet() in EPICS 7.  dbGet() was changed so that it now returns the current length of long string fields, rather than the maximum length.  This caused problems for autosave with PVs that are long strings and whose value is empty ("") .  Rather than an empty string being written to the autosave file, the string "0" was written.  This caused serious problems with the calculation fields in the transform record, for example, where there is an important difference between an empty string and  the string "0".  This is fixed in the master branch of autosave.

This is a pretty short list, and I want to thank the EPICS 7 developers for all the effort they have put into this project!

Mark

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System