Expressions are evaluated when getToken() or validate() is called on the Parameter, but only if the expression has changed. It is a documented feature that random() in an expression yields one random number, not a new one each time you read the value of the parameter.
I believe this is the correct behavior. Reading the value of a parameter should never cause its value to change. This would usually be an unexpected side effect. Reading is reading, and is distinct from writing. Edward At 07:17 PM 3/25/2002 +0000, G Dring wrote: >This has been posted to com.soft-sys.ptolemy as well - sorry if you see this >twice, but as a newbie, I'm not sure which is the better way to go... > > >I have a query about how and when expressions are evaluated... I'm coding a >class in the DE domain which allows you to provide an expression as a >parameter. This defines an inter-firing period. I don't think that the >expression is being evaluated each time. The relevant bits of code are: > >=> variable declaration > > public Parameter interArrival; > >=> within the constructor method > > interArrival = new Parameter(this, "interArrival", new >DoubleToken(1.0)); > >=> within the fire() method > > _delay = ((DoubleToken)interArrival.getToken()).doubleValue(); > > _nextMsgTime = ((DEDirector)getDirector()).getCurrentTime() + _delay; > >If I use this as code as a message source, I get a constant inter-arrival >time for the entire session. This appears to be setting the value of the >expression at initialization time, wheras I'd like it set each time I >evaluate interArrival.getToken(). Any advice? > >(I also posted a reply to myself, which I include here for completeness...) > >A bit of additional information that I have dug out of the classes... > >Parameter is derived from Variable, which (very sensibly, in most cases) >sets a _needsEvaluation flag. However, when using an expression such as >random() or gaussian(), this causes the same token to be presented each >time, unless somehow the value is changed... It seems that the variable >class could do with another private variable _alwaysEvaluate, so that this >behaviour can be controlled... > >giles > >*** A million monkeys were given a million typewriters. It's called the >Internet > >---------------------------------------------------------------------------- >Posted to the ptolemy-hackers mailing list. Please send administrative >mail for this list to: [EMAIL PROTECTED] ------------ Edward A. Lee, Professor 518 Cory Hall, UC Berkeley, Berkeley, CA 94720 phone: 510-642-0455, fax: 510-642-2739 [EMAIL PROTECTED], http://ptolemy.eecs.berkeley.edu/~eal ---------------------------------------------------------------------------- Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
