Experimental Physics and
Industrial Control System

Korhonen Timo <[email protected]> · Thu, 4 May 2000 09:21:09 +0200 (MEST)

On Wed, 3 May 2000, Marty Kraimer wrote:

> Eric Bjorklund wrote:
> 
> > My question here is that since we're mucking with the aao record anyway,
> > wouldn't
> > it be better to have the record support module allocate the buffer during
> > pass 0
> > based on NELM and FTVL?  Then the only device support change would be to
> > *remove*
> > the buffer allocation code.  We would, of course, have to be sure that FTVL was
> > set correctly in each of the records -- for me that would be no problem (and I
> > suspect it wouldn't be for Timor either at this point).

Indeed not. I would prefer this solution.

> Will this work with all device support attached to aao?

I dont see why it would not (with the changes that have to be done
anyway). Referring again to the waveform record, it allocates the buffer
in the record support code using the method that Eric described.

> It sounds like the best solution if it works. Unless someone out there objects I
> say use this solution. It does mean any device support has to be changed, but
> this can be stated in the release notes.

It looks like we have a solution. However, what worries me before going
ahead, is not having heard any comments from JLab (the source for this
code) so far... 

What would you think of adding to the aai/aao records proper conversions
to engineering units (like in ai/ao)? They are not presently there and
would be very useful (at least for our applications.)

Timo

Timo Korhonen  PSI (Paul Scherrer Institut), SLS 
               CH-5232 Villigen PSI 
               tel + 41- 56 3103262  fax + 41 - 56 310 3151 
e-mail:	       [email protected]

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System