Re: lex function for values

[Ben Pfaff]

John Darrington <address@hidden> writes:

> On Fri, May 22, 2009 at 08:03:36AM -0700, Ben Pfaff wrote:
>      John Darrington <address@hidden> writes:
>      
>      > It struck me that the value-rep branch would lend itself to a function
>      > of the form:
>      >
>      > bool lex_match_value (struct lexer* , union value *);
>      >
>      >
>      > which would be useful for matching either a string or a number.
>      
>      (A width would also need to be passed in.)
>
> The width could be infered from the length of the string. 

You mean, from the length of lex_tokstr(lexer)?  I don't think
that would help to distinguish string values from numeric values,
at least.

>      Is this useful very often?
>
> Enough to make it worthwhile I think.  We currently have something
> like this in t-test.q (for the grouping variable).  Similarly, I can
> envisage that any procedure where grouping or categorical values are
> involved, particularly dichotomous ones might need this.

OK.  That's good enough for me.
-- 
Ben Pfaff 
http://benpfaff.org