[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Reductions during Bison error handling

From: Hans Aberg
Subject: Re: Reductions during Bison error handling
Date: Tue, 14 May 2002 23:09:14 +0200

At 13:41 -0600 2002/05/14, Richard Stallman wrote:
>    If one should be specific, then perhaps 1b should be called the
>    reduce/reduce conflict default reduction, and 1a the LALR(1) algorithm
>    default reduction.
>That is reasonable.

Paul pointed out that it appears to not be depending on the LALR algorithm,
but on the implementation, so my suggestion for 1b is wrong.

>  Could you look thru the Bison manual for places
>where being specific in those ways could help prevent confusion?

And following Paul's remark, there should be no mentioning of how 1b is
defined  in the Bison manual, as that is only an implementation issue
(essentially, implementors favorite choice of "most common reduction").
(And one does not want to put unnecessary restraints on implementations.)

Actually, a Bison manual search for the word "default" (case ignored)
yielded 48 occurrences, and none of them in connection with the word
reduction, so there appears to be no problem here. Also, the Bison manual
does mention that Bison resolves a reduce/reduce conflict by choosing to
use the rule that appears first in the grammar, just as POSIX requires.

  Hans Aberg
                  * Email: Hans Aberg <mailto:address@hidden>
                  * Home Page: <http://www.matematik.su.se/~haberg/>
                  * AMS member listing: <http://www.ams.org/cml/>

reply via email to

[Prev in Thread] Current Thread [Next in Thread]