Re: Adding a "plugin" parser to ifile

On Thu, 06 Mar 2003 15:36:18 -0500, Jason Rennie <[email protected]> said:

Jason> [email protected] said:
>> - I've been messing about with various ifile options "-h -w -k"[1] and
>> I'm coming to the conclusion that it would be quite useful to
>> customize the parser for ifile. While you can do this in front of
>> ifile, I think as a first pass it would be convient to provide a
>> "subprocess parser". 

Jason> Let me propose something here.  How about we break ifile into
Jason> two parts: the parsing/lexing part and the classifier.  To run
Jason> ifile, you'd give the message to the first program and pipe the
Jason> output to the second program.  The first program could output
Jason> one token per line.  Then the lexing part of the classifier
Jason> would be exceedingly simple.  Breaking it apart like this would
Jason> make it easier for people to experiment with different parsing/
Jason> lexing styles; it would even make it possible to write
Jason> prototype parsers/ lexers in perl or some language that's
Jason> easier to write.

Jason> Thoughts?

Well, at this point, I'm cool with the way things are, so let me be
the first to scream for a front-end script that maintains
compatibility with current behavior.  I.e. ifile would now be a shell
script that defaults to

   ifile-lexer [lexer flags] | ifile-classifier [classifier flags]

This would be useful for folks using ifile today who pick up a new
version but don't want to rewrite their software to understand this
new scheme.

Also, if we do this, we should write a document clearly defining what
is expected of a lexer and what is expected of a classifier, so BOTH
of these programs can be pluggable.  So if someone wants to write a
different lexer, great.  If they want to write a different classifier,
also great.
-- 
Dave Marquardt
Round Rock, TX