Related articles |
---|
regular expressions (bug-report) megatest!plethorax!djones@uu4.psi.com (1993-03-14) |
Re: regular expressions (bug-report) Paul.Klint@cwi.nl (1993-03-15) |
Re: regular expressions (bug-report) mab@wdl39.wdl.loral.com (1993-03-15) |
regular expressions (bug-report) jos@thinx.convex.com (Jos Horsmeier) (1993-03-16) |
Re: regular expressions (bug-report) pardo@cs.washington.edu (1993-03-19) |
Re: regular expressions (bug-report) henry@zoo.toronto.edu (1993-03-25) |
Re: regular expressions (bug-report) kanze@us-es.sel.de (1993-03-26) |
Newsgroups: | comp.compilers |
From: | Paul.Klint@cwi.nl (Paul Klint) |
Keywords: | lex, bibliography |
Organization: | CWI, Amsterdam |
References: | 93-03-046 |
Date: | Mon, 15 Mar 1993 10:54:20 GMT |
Dave Jones writes:
|> However it is not immediately clear to me how to handle character range
|> matches. For example, most reg-exp packages allow you to indicate a range
|> of matches with a construct such as this:
|> [a-zA-Z]
|> ...
|> The simplest way would be to split the construct [a-zA-Z] into 52
|> single-character transitions. ...
You may want to have a look at the following paper:
J. Heering, P. Klint, J. Rekers,
Incremental generation of lexical scanners,
ACM TOPLAS, Vol. 14, Number 4 (October 1992),
pp. 490-520.
In Section 6.2. (Character classes) on pages 507-510 the above problem
is addressed.
Paul Klint
--
Return to the
comp.compilers page.
Search the
comp.compilers archives again.