Re: RE-ASSOC: a question about the associativity of RE concatenation

[Paul Eggert <yyyyyy@xxxxxxxxxxx>]

> Date: Wed, 10 Apr 2002 15:00:05 +0900
> From: Isamu Hasegawa <yyyyy@xxxxxxxxxxxxxx>
> 
> "Implementations may use other programming languages or algorithms,
> as long as the syntax supported is the same as that represented by
> the grammar".
> 
>(<http://www.opengroup.org/onlinepubs/007904975/utilities/xcu_chap01.html#tag_01_10>)

> If RE implementations must use Left-associative grammar, the standard
> must say "as long as the grammar is the same", I think.

But the syntax is determined by the grammar in combination with other
rules (e.g., lexical rules).  So the standard is saying that you can't
implement only the grammar; you must also implement the other
syntactic rules.

The intent of that quote is that one need not use C-language actions
in a Yacc-generated LALR(1) parser to implement regular expressions
and the like.  I wouldn't read too much into this freedom, though.
Yes, you can use Bison or ML or Ruby if you want to, but you must
still implement behavior that conforms to all the syntactic and
semantic requirements imposed by the standard.