Experimental Physics and
Industrial Control System

Marty Kraimer <[email protected]> · Mon, 27 Jun 2005 11:02:48 -0500

I thought about CA client examples, In particular the following two

examples:

a client wants to read a double value and displayLimits where

displayLimits is assumed to have well known propertyIds.

The second example is a V4 version of probe, i.e. given a pvname

display it's propertyIds and the values of the properties.

For the first example part of the implementation is the property

catalog implementation. This will be the propertyCatalog passed to a

read request. It will be used to put values into my container

class ValueDisplayLimits: public propertyCatalog {
    double value;
    displayLimits limits;
    bool   valueModified;
    void traverse (dataViewer &viewer) {
        viewer.reveal(properIdPrimitiveValue,value);
        viewer.reveal(properIdDisplayLimits,limits);
    };
    void traverse (datasManipulator &manipulator) {
        double saveValue = value;
        valueModified = false;
        manipulator.reveal(properIdPrimitiveValue,value);
        manipulator.reveal(properIdDisplayLimits,limits);
        if(saveValue!=value) valueModified = true;
    };
};

I will leave out the calls to the CA client library, since my questions

are about propertyIds.

My questions relate to the arguments passed to the reveal methods.
The two arguments are:

propertyId - something created or located via a character string

objref - a reference to one of the following: primitive type

(char,short, double,etc), string segment, array segment, or a

propertyCatalog

Now for some questions

In my example code I assumed that displayLimits has well known

propertyIds.

What about the double value?

I just made up the name propertytIdPrimitiveValue. What should it be?
The more general question is

For a generic tool that knows how to handle primitive data types how

does it get a propertyId?

It seems that there must be well defined propertyIds for this.

Is the string "primitive value" what we want?
Do we also have "string segment" and "array segment"?

What about the properIds for a propertyCatalog? Do we have "property

catalog"?

Now for questions about developing a V4 version of probe. I assume that

it is just given a pvname

How does it get started, i.e. how does it know how to find properties

associated with the pvname?

Can it ask for a list of properrtyIds?

If so do we have a standard set of propertyIds via which this list can

be obtained?

Given a list of propertyIds how does it obtain the information

associated with each id?

I can guess that one way this could be implemented is via a dataViewer.
It could work like the following:

Ca accepts a request to get data associated with a propertyId. Instead

of putting it into a caller provided container via the caller supplied

propertyCatalog, CA would make the data available via the caller

supplied dataViewer. Does the proposed CA V4 interface allow this?

It would be nice if it was possible to find out info about a propertyId

without needed to request the data associated with the propertId,

Things like: isPrimitive, isStringSegment, isArraySegment,

isPropertyCatalog.

If it is primitive, it would be nice to know the exact type.

Thus should introspection of all dataAccess propertyIds be available?

Marty

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System