Experimental Physics and
Industrial Control System

Marty Kraimer <[email protected]> · Tue, 03 Feb 1998 13:13:45 -0600

Almost as soon as I sent the previous message I was sorry I did. The
argument between:

1) adding a new field to each record
2) adding a new DBR_XXX type to channel access

are the same as for getting the record type name, i.e. VERS and RTYP are
similar.

In each case information about the record type, NOT the record instance
is being requested.

Creating two new character fields for EVERY record instance seems rather
excessive. Thus the version could be stored with the record type
description just like the record type name is now.

>From the responses to the message I sent about RTYP there are objections
to implementing RTYP via a new DBR_type rather than a new field in each
record instance.

Whatever is decided the same method should be used for both RTYP and
VERS, i.e. either both are implemented via new fields in each record
type or both are handled via new DBR_types.

I will respond to the RTYP message before saying more.

Marty Kraimer wrote:
> 
> This finally got to the top of my TODO list.
> 
> There was a lot of discussion about how to implement this feature.
> The main ideas were:
> 
> 1) Provide a field RTYP in every record.
> 2) Provide a new Channel Access Request type.
> 
> I agree with the second alternative.
> 
> Jeff Hill suggested that we have a new request type DBR_CLASS_NAME.
> For example:
> 
> char        classname[80];
> 
> ...
> 
> SEVCHK(ca_get(DBR_CLASS_NAME,mychid,&classname),"ca_get");
> SEVCHK(ca_pend_io(waitTime),"ca_pend_io");
> printf("pv: %s classname %s\n",ca_name(mychid),classname);
> 
> I modifyied db_access.h, db_access.c, and the ca/convert.c to implement
> the change. Jeff will still have to look to see if any other channel
> access changes are necessary if new clients talk to old servers.
> 
> Marty Kraimer

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System