MathGroup Archive 1994

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: Options and Contexts

  • To: mathgroup at yoda.physics.unc.edu (mathgroup)
  • Subject: Re: Options and Contexts
  • From: John Lee <lee at math.washington.edu>
  • Date: Tue, 19 Apr 94 11:44:26 -0700

Daryl S. Reece <gt9397a at prism.gatech.edu> writes:

> I have a function that I want to define in its own context.
> I want to assign options to this function within its own context.
> I then want to be able to call the function from the Global
> context with options and not have to differentiate between
> the context of options assigned at the global level
> and the default values.  An example will help.

> BeginPackage["Test`"];
> Tester::usage = "Tester[opts___] acts based on options.";
> Opt1::usage = "Option for Tester.";
> Begin["Test`Private`"];
> Options[Tester] = {Opt1->A};
> Tester[opts___] := 
> Module[{opt1},
>   opt1 = Opt1/.{opts}/.Options[Tester];
>   Print["Equality of option and A ", opt1===A];
>   Print["Equality of option and B ", opt1===B];
> ]
> End[];
> EndPackage[];

> Tester[]
> Equality of option and A True
> Equality of option and B False

> Tester[Opt1->B]
> Equality of option and A False
> Equality of option and B False

> Tester[Opt1->Test`Private`B]
> Equality of option and A False
> Equality of option and B True

> These last two examples show that any options assigned at
> the global level cannot be compared within the Private contexts
> without specifically declaring their contexts.  This is a problem 
> since I would like to be able to call the package from any context
> and have it behave the same.  Thanks ahead of time for the help.

All you have to do is to define in the "Test`" context, with usage
messages, any symbols that can legitimately be passed as options.  For
example, if you add the following lines after the third line above,

  A::usage = "Default value for the Opt1 option of Tester.";
  B::usage = "Value for the Opt1 option of Tester.";

then you should get the behavior you want.

Jack Lee
Dept. of Mathematics
University of Washington






  • Prev by Date: Laplace transform
  • Next by Date: integrating interpolations
  • Previous by thread: Options and Contexts
  • Next by thread: Mathematica World - April