Re: C++ parsing, was intermediate representation.

mefrill@yandex.ru (Vladimir) writes:


| henry@spsystems.net (Henry Spencer) wrote in message
| > In his "The Design and Evolution of C++", Stroustrup says that he let
| > Aho and Johnson talk him out of writing his own recursive-descent
| > parser for C++... and he now thinks that was a big mistake.
|
| Stroustrup means using YACC or Bison I think, the programms that
| generate LALR(1) parser and, consequently, demand input grammar must
| be LALR(1). It is known C++ grammar is not LALR(k) for any k. So,
| Stroustrup spent much time trying to write LALR grammar for C++. For
| the moment there is GLR parser as the next generation LALR parsers,
| which makes writing C++ frot-end fast and simple time spending.


In April 2004, I attended the Compiler Construction (part of ETAPS)
conference where a parser for C++ (Elkhound), based on GLR technology,
was presented. My understanding is that, in practice, it was far from
handling C++, notably most uses of templates. I don't know whether
any substantial progress has been made since then.


--
                                                                                                              Gabriel Dos Reis
                                                                                      gdr@integrable-solutions.net

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