Experimental Physics and
Industrial Control System

[email protected] (Tim Mooney) · Mon, 17 Aug 1998 12:22:27 -0500

re...

> > > But if SHFT were allowed to be set by GDCT, wouldn't it just be
> > > overwritten by the device support layer in the cases where the device
> > > support sets this field, thus making it work for all cases?
> > 
> > Yup. I neither see any problems with the SHFT value made DCT accessible.
> > What do you think, Marty?
> > 
> 
> I dont see a problem except that it may be confusing.
> 
> 
> > Another workaround (as SHFT is DB accessible):
> > Put a couple of dbpf commands at the end of the startup.command script
> > (after iocInit) to set the SHFT fields.
> 
> 
> I think I prefer this as a solution because using raw soft support will
> not be used very often. 

I think using signal to specify SHFT is arcane and confusing to begin with, and
it makes an awful mess in the database when you have several parallel ports on the
same card.  I also think requiring that SHFT be specified in the startup file (for
soft support only?) is actually more confusing than allowing it to be specified in
a .db file would be, and makes it impossible to distribute a database that you know
is going to work.  

My two cents go toward allowing both signal and SHFT to specify SHFT--using SHFT
if it is nonzero, and falling back on signal only if SHFT is zero.  Every existing
database will work with this choice, and it gives future device support an obvious
prescription for handling more than one parallel port per card.

Tim Mooney

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System