RE: C++ lexer (GCC 3.1.1) requires knowledge of other C dialects

["Gary Funck" <gary@Intrepid.Com>]

> -----Original Message-----
> From: Joe Buck [mailto:Joe.Buck@synopsys.com]
> Sent: Thursday, August 01, 2002 2:42 PM
> To: Gary Funck
> Cc: gcc@gcc.gnu.org
> Subject: Re: C++ lexer (GCC 3.1.1) requires knowledge of other C
> dialects
>
>
>
> > Although this is a bit of philosophical debate (regarding the merits of
> > handbuilt vs. program-generated parsers), I'll point out that the ANTLR
> > parser generator (http://www.antlr.org) generates LL(k)
> recursive descent
> > parsers with additional features which include: [ many features ]
>
> However, the ANTLR FAQ explains that you can't write a C++ parser in ANTLR
> because of the complexities of the language.  They recommend hand-written.
>
> See http://www.jguru.com/faq/view.jsp?EID=531848
>

Yes, though further down that same page, they mention the following:
"I should mention that John Lilley produced a fairly decent C++ parser for
ANTLR's predecessor, PCCTS, but it required a modification to PCCTS and is
not compatible with latest ANTLR."

PCCTS is described here:
http://www.antlr.org/pccts133.html
which in turn refers to this implementation of a C++ grammar:
http://www.empathy.com/pccts/
and
http://www.empathy.com/pccts/download.html
and this interesting resource page
http://www.empathy.com/pccts/roskind.html
which contains Jim Roskind's implementation of a C++ parser written for an
LALR parser generator (yacc/bison), along with an analysis of the
difficulties encountered.

While we're on the topic, there is this a backtracking implementation of
yacc, known as btyacc, which according to the authors has been used to build
a C++ parser:
http://www.siber.org/btyacc/

It turns out that ANTLR might not be suitable for building a parser for GCC
because ANTLR is written in Java, and neither ANTLR nor PCCTS are GPL'd
(though they are explicitly public domain with a BSD-like copyright notice).
I wasn't trying to imply that these would be the tool of choice, but simply
noting that these automatic parser generators meet many of the technical
objections mentioned by advocates of handbuilt recursive descent parsers. In
the sense that bison is a GPL'd implementation of YACC, there would nothing
to prevent re-implemtnating (and extending) any of the tools mentioned above
in a form suitable for inclusion into GCC.