At 04:08 PM 8/4/00 -0400, Chaim Frenkel wrote:
>I've never used it, but PL/I' preprocessor and ASMG's preprocessor ,
>If I recall correctly, both worked this way. The text of the arguments
>and the actual input stream were available for manipulation. The
>return value (or emitted strings) were used as input to the parser.
>
>This would be quite close to a BEGIN time subroutine with the rest
>of the input stream supplied as a FILEHANDLE. This could allow those
>folks who want to turn perl into a different language, enough rope.
Yup. I misunderstood what he was talking about at first. Now I understand.
I don't think this'll be much of a big deal over and above BEGIN. (How much
of a pain that is is a separate issue)
> >>>>> "DS" == Dan Sugalski <[EMAIL PROTECTED]> writes:
>
>DS> At 09:38 PM 8/4/00 +0200, Jean-Louis Leroy wrote:
> >> Dan Sugalski <[EMAIL PROTECTED]> writes:
> >>
> >> > This does complicate the job of the parser/lexer rather
> >> > considerably.
> >>
> >> Why? Isn't it 'just' a matter or making the lexer read from a
> >> hot-redirectable input stream?
>
>DS> If it is, then it's not that big a deal, nor would it be all that useful.
>DS> Wiht substitutions it's more than that, and if we're going to do this, we
>DS> ought *not* essentially duplicate C's #define stuff. Macros mean the
>parser
>DS> needs to keep source around, possibly snapshot its current state and then
>DS> do an on-the-fly rework of the input buffers. (Maybe. Depends on what
>this
>DS> does)
>
>DS> The cleverer the preprocessor, the more work the lexer/parser needs to
>do.
>DS> Which is dandy but, as I said, if you're going to do it, do it once
>and do
>DS> it with some force.
>
>
>
>
>
>--
>Chaim Frenkel Nonlinear Knowledge, Inc.
>[EMAIL PROTECTED] +1-718-236-0183
Dan
--------------------------------------"it's like this"-------------------
Dan Sugalski even samurai
[EMAIL PROTECTED] have teddy bears and even
teddy bears get drunk
