Re: EDIF2 Lex/YACC parser needed

bevan@cs.man.ac.uk (Stephen J Bevan)
Thu, 4 Aug 1994 08:25:45 GMT

          From comp.compilers

Related articles
EDIF2 Lex/YACC parser needed dpappas@ichips.intel.com (1994-08-03)
Re: EDIF2 Lex/YACC parser needed wjw@iaehv.iaehv.nl (1994-08-03)
Re: EDIF2 Lex/YACC parser needed gdonl@gv.ssi1.com (1994-08-04)
Re: EDIF2 Lex/YACC parser needed bevan@cs.man.ac.uk (1994-08-04)
| List of all articles for this month |

Newsgroups: comp.compilers
From: bevan@cs.man.ac.uk (Stephen J Bevan)
Keywords: parse
Organization: Department of Computer Science; University of Manchester
References: 94-08-042
Date: Thu, 4 Aug 1994 08:25:45 GMT
Status: RO

In article 94-08-042 dpappas@ichips.intel.com (Derek Pappas) writes:
      I need an EDIF2 Lex/YACC parser. Please send me a pointer to an ftp
      site and the ip address if you know it. ...


      [This question has come up occasionally since 1987 without anyone ever
      offering one. Perhaps the EDIF users can get together and whip one up.
      -John]


Producing an EDIF version 2 0 0 parser using Lex/YACC is a bit awkward
given that the language is defined in three levels each of which has a
grammar which is a superset of the next level down and all of which
can appear at some point in a single file. If you punt on dealing
with the higher levels and concentrate on level 0 so you only have one
grammar, you still have the problem that "standard" YACCs usually blow
up because there are too many productions* -- BISON can cope ok,
though. You then have the fun of trying to put some decent error
recovery into the grammar. All in all, a pretty tedious job -- at
least I found it so.


If you aren't tied to using YACC, then there are some alternatives.
There is the Berkeley EDIF version 2 0 0 toolkit, see the comp.lsi.cad
FAQ for more information. If you aren't averse to commercial software
then contact edif-support@cs.man.ac.uk and ask about companies that
provide parsers/toolkits for EDIF version 2 0 0 and/or the recently
released version 3 0 0.


------------


* just under 300 in the official EBNF, probably at least twice that
    when you expand out any use of { ... } and probably into the
    thousands if you try and expand out any use of < ... > into a format
    YACC can accept.
--


Post a followup to this message

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