Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uwm.edu!zaphod.mps.ohio-state.edu!think.com!mintaka!spdcc!iecc!compilers-sender
From: boyland@aspen.berkeley.edu (John Boyland)
Newsgroups: comp.compilers
Subject: Re: Technical question about yacc and LALR parsers
Keywords: parse, design
Message-ID: <9104232230.AA18767@aspen.Berkeley.EDU>
Date: 23 Apr 91 22:30:11 GMT
References: <9104222013.AA09131@florida.HQ.Ileaf.COM>
Sender: compilers-sender@iecc.cambridge.ma.us
Reply-To: boyland@aspen.berkeley.edu (John Boyland)
Organization: Compilers Central
Lines: 34
Approved: compilers@iecc.cambridge.ma.us

In article <9104222013.AA09131@florida.HQ.Ileaf.COM>, jar@florida.HQ.Ileaf.COM (Jim Roskind x5570) writes:
|> [...] A second
|> example would would be (again verifiable by the parser generator) when a
|> "grammar preprocessor" back-substituted some rules and in doing so,
|> replicated the code actions.
|> [...]
|> 
|> Jim Roskind-   Author of a YACCable C++ grammar
|> jar@hq.ileaf.com or ...!uunet!leafusa!jar

Unfortunately, grammar preprocessors (such as one I've developed), often
have subtly different actions:

For example:

stmtlist : stmtlist ';' ? ';;' stmtlist
	{ $$ = sequence($stmtlist'1,$stmtlist'2); }

gets expanded into:

stmtlist : stmtlist ';' DOUBLESEMI stmtlist
	{ $$ = sequence($1,$4); }
         | stmtlist DOUBLESEMI stmtlist
	{ $$ = sequence($1,$3); }

(Duplicating rather than having a special optional-semicolon rule more often results in an LALR(1) grammar).

Notice how the actions differ in where arguments are found.

John Boyland
boyland@sequoia.Berkeley.EDU
-- 
Send compilers articles to compilers@iecc.cambridge.ma.us or
{ima | spdcc | world}!iecc!compilers.  Meta-mail to compilers-request.