Re: Top-Down Parser Construction Conjectures

sja@vinkku.hut.fi (Sakari Jalovaara)
Mon, 18 Jan 1993 10:50:05 GMT

          From comp.compilers

Related articles
Re: Compiler Construction in Ada crigler@osceola.cs.ucf.edu (1993-01-08)
Top-Down Parser Construction Conjectures bart@cs.uoregon.edu (1993-01-18)
Re: Top-Down Parser Construction Conjectures W.Purvis@daresbury.ac.uk (1993-01-18)
Re: Top-Down Parser Construction Conjectures sja@vinkku.hut.fi (1993-01-18)
Re: Top-Down Parser Construction Conjectures parrt@ecn.purdue.edu (1993-01-19)
Re: Top-Down Parser Construction Conjectures pardo@cs.washington.edu (1993-01-20)
| List of all articles for this month |

Newsgroups: comp.compilers
From: sja@vinkku.hut.fi (Sakari Jalovaara)
Organization: Helsinki University of Technology, Finland
Date: Mon, 18 Jan 1993 10:50:05 GMT
Keywords: parse, performance
References: 93-01-048 93-01-122

>> hand-coded recursive descent, and very impressive performance stats have
>> been reported in this group. So this technique is of more than simple
>> academic interest.


Oh, recursive descent is easily faster than yacc. I wrote a C parser in
yacc and later replaced it with a mostly recursive descent hand- coded
parser ("mostly": expressions are not RD; labels require a kludge.) The
hand-coded parser is easily several times faster than the yacc version.


Transforming a yacc parser into RD is easy-ish since yacc directly gives a
list of lookahead tokens for any given parse state (e.g. you can easily
see which tokens can start a declarator.)
++sja
--


Post a followup to this message

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