Re: [LONG&BORING] ULR : Ultimate LR parser =)

Oliver Zeigermann <oliver@zeigermann.de>
31 Oct 2003 23:05:39 -0500

          From comp.compilers

Related articles
[LONG&BORING] ULR : Ultimate LR parser =) deadcow@free.fr (Nicolas Repiquet) (2003-10-27)
Re: [LONG&BORING] ULR : Ultimate LR parser =) oliver@zeigermann.de (Oliver Zeigermann) (2003-10-31)
| List of all articles for this month |

From: Oliver Zeigermann <oliver@zeigermann.de>
Newsgroups: comp.compilers
Date: 31 Oct 2003 23:05:39 -0500
Organization: T-Online
References: 03-10-112
Keywords: parse, LALR
Posted-Date: 31 Oct 2003 23:05:39 EST

Nicolas Repiquet wrote:
> - a precedence level
> - an associative indication


This is exactly what YACC offers you to resolve ambiguities. You may
write for example


%left '+' '-'
%left '*' '/'


to make '+', '-', '*' and '/' left associative and give '*' and '/'
precedence over '+' and '-'. Seems you reinvented the wheel :)


Note, however, this does not augment the set of parsable languages, but
only helps to eliminate ambiguities in the grammar. This means, you
could rewrite the grammar to parse the same language and not use
operator precedence - although it will certainly look less obvious then.


Oliver


Post a followup to this message

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