Re: A fast way to match regexp's and extract parts of them

In article 91-10-016 roberto@cernvax.cern.ch (roberto bagnara) writes:
  > Let's have a set of such `extended regular expressions' (ERE), each of
  > these being associated to an `action'. That `action' has as many parameters
  > as there are '{}'-enclosed sub-expression in the associated ERE.
  > For example:
  >
  > RE1) a{b*c}d{[0-9]*} action: foo($1, $2)
  > RE2) _{[A-Za-z_][A-Za-z0-9_]*} action: enter_priv_symbol($1)


This problem is addressed in a recent paper in the journal
`Software---Practice and Experience'. The paper describes how the
subexpressions can be matched during scanning with a DFA, making
re-scanning unnecessary. A second paper describes a tool based on
the techniques.


It was in either the September or October issue. I am sorry
that I do not have more precise details to hand.
--
Stephen Adams Email: S.R.Adams@ecs.soton.ac.uk
Electronics and Computer Science Tel: 0703 593649
University of Southampton Fax: 0703 593045
Southampton SO9 5NH, UK
--

Return to the comp.compilers page.
Search the comp.compilers archives again.