Re: Is this a new idea?

drw@euclid.mit.edu (Dale R. Worley)
Mon, 16 Nov 1992 22:24:28 GMT

          From comp.compilers

Related articles
[11 earlier articles]
Re: Is this a new idea? clyde@hitech.com.au (1992-11-07)
Re: Is this a new idea? dlarsson%abbaut@Sweden.EU.net (1992-11-11)
Re: Is this a new idea? macrakis@osf.org (1992-11-11)
Re: Is this a new idea? pardo@cs.washington.edu (1992-11-12)
Re: Is this a new idea? thinkage!dat@math.uwaterloo.ca (1992-11-11)
Re: Is this a new idea? andrewb@lynx.cs.washington.edu (1992-11-16)
Re: Is this a new idea? drw@euclid.mit.edu (1992-11-16)
Re: Is this a new idea? firth@sei.cmu.edu (1992-11-17)
Re: Is this a new idea? clyde@hitech.com.au (1992-11-18)
Re: Is this a new idea? macrakis@osf.org (1992-11-20)
| List of all articles for this month |

Newsgroups: comp.compilers
From: drw@euclid.mit.edu (Dale R. Worley)
Organization: MIT Dept. of Tetrapilotomy, Cambridge, MA, USA
Date: Mon, 16 Nov 1992 22:24:28 GMT
References: 92-10-113 92-11-047
Keywords: C, design

clyde@hitech.com.au (Clyde Smith-Stubbs) writes:
      ...you can't parse code that contains references to such things as
      typedefs that occur earlier in the code, if the typedef is missing
      or in the middle of being edited....


macrakis@osf.org (Stavros Macrakis) writes:
      This is a C-specific problem: a bug in the design of C's syntax. Most
      other languages don't commit such foolishness, except of course those
      with extensible syntax.


Any language has this problem if it has tokens whose syntactic
category can't be determined solely by the form of the token. In C's
case, "identifier" tokens can be either identifiers or type-names. In
Algol 68, "boldface words" can be either type names or operators.
Almost any language with an extensible set of operators is going to
run into this problem.


Dale


Dale Worley Dept. of Math., MIT drw@math.mit.edu
--


Post a followup to this message

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