Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!decwrl!decvax!ima!compilers-sender
From: oakhill!epsilon!steve@cs.utexas.edu
Newsgroups: comp.compilers
Subject: Re: I need answer to a simple YACC question.
Summary: Here is the code you need.
Message-ID: <3359@ima.ima.isc.com>
Date: 16 Feb 89 23:09:32 GMT
References: <3281@ima.ima.isc.com> <3328@ima.ima.isc.com>
Sender: compilers-sender@ima.ima.isc.com
Reply-To: oakhill!epsilon!steve@cs.utexas.edu
Organization: Motorola Inc. Austin, Tx
Lines: 64
Approved: compilers@ima.UUCP

In article <3328@ima.ima.isc.com>, djones@megatest.uucp (Dave Jones) writes:
> [yacc error states are hard to analyze, particularly if you want to see
> the follow sets in a particular context.]

In fact here is the code that does this error dumping:

257,275c256,257
< 			if ((yy_n = yypact[yystate]) > YYFLAG && yy_n < YYLAST)
< 			  {
< 			  short index;
< 			  extern char *p_error();
< 			  fprintf(stderr,"expecting: ");
< 			  for (index = yy_n>0 ? yy_n : 0;
< 			       index < YYLAST;
< 			       ++index)
< 			    {
< 			    if (yychk[yyact[index]] == index - yy_n
< 			        && index - yy_n != YYERRCODE)
< 			    fprintf(stderr,"%s, ",p_error(index-yy_n));
< 			    }
< 			  fprintf(stderr,"found:");
< 			  fprintf(stderr,"%s\n",p_error(yychar));
< 			  }
< 			else
< 			  yyerror( "syntax error" );
< 			  goto skip_init;
---
> 				yyerror( "syntax error" );
> 				goto skip_init;

For those of you who cannot tell this is a diff of yaccpar that does that
does this error dumping.  This yaccpar is from the standard SYSTEM V.  You
do need to supply the function char *p_error(int).  This function takes the
token number passed in, and changes it into a string name for that token.
Errors come out in the fashion:

expecting: ')', ',', '+', found '['.

Not totally coherent, but better than 'syntax error'.  This message gets
very long in some cases.

Three last notes:
  1) Credit for this change does not belong to me but to one of my partners
     at Remora, Inc.  I think he got this trick from a book (which my copy
     is unfortunately at home) on building a compiler in UNIX (reference
     supplied if requested).  At least I can claim I understand what is
     happening here.
  2) Do not comment on the programing style too much.  This code was hacked
     today from our (Remora's) version of yaccpar into the UNIX version
     just for this post.  The code as posted works (is tested).
  3) Excuse the cryptic nature of this message, I am typing it in on my
     own lunch time when I should be debuging Motorola's code.

                   enough from this mooncalf - Steven
--
Steven R Weintraub                        cs.utexas.edu!oakhill!devsys!steve
Motorola Inc.  Austin, Texas 
(512) 440-3023 (office) (512) 453-6953 (home)

--
Send compilers articles to ima!compilers or, in a pinch, to Levine@YALE.EDU
Plausible paths are { decvax | harvard | yale | bbn}!ima
Please send responses to the originator of the message -- I cannot forward
mail accidentally sent back to compilers.  Meta-mail to ima!compilers-request