Experimental Physics and
Industrial Control System

Andrew Johnson <[email protected]> · Tue, 12 Oct 1999 15:09:36 -0500

Luedeke Andreas wrote:
> 
> we observed a strange "mis-feature" with our database files. Actually it
> seems that the configuration of a VME_IO hardware link depends on the
> order of DTYP and INP/OUT definition in the .db file. Is this meant to
> be that way? If not, what have we done wrong?

The order of the definitions of DTYP and INP/OUT fields is obviously
important as you have discovered, and I can believe this makes the work in
dbLoadRecords much easier to do (with the existing link type definitions).
This is the one "hidden" dependency between record fields, and it's
something which we want to get rid of (see my Link Support proposal, which
unfortunatly hasn't moved very far since the SLAC meeting due to time
pressure from elsewhere).

> I don't know what's the default setting of the field type,
> but it seems to be not VME_IO.

I think the default value for a record's DTYP field is determined by the
first device definition for that record read in by dbLoadDatabase.  For
most of the built-in types those are likely to be the soft record support.

> I neither found anything written down in the application developers
> guide nor in the record reference manual and searching for some keywords
> in tech-talk didn't gave useful results. I can't believe that I'm the
> first one stumbling over that. Maybe the one who has written the stuff
> can comment on it.

I hope Marty is enjoying the wine and weather in Italy after ICALEPCS, but
in his absence I think I can safely guess that this was an omission from
the IOC Application Developers Guide, probably from Pages 35 or 36.  Maybe
it should also appear in the Database Concepts chapter of the Record
Reference Manual as well, although that doesn't get updated very often.

- Andrew
-- 
Complexity comes for free, Simplicity you have to work for.

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System