Re: Bison's Semantic Parser

[Akim Demaille]

>>>>> "Hans" == Hans Aberg <address@hidden> writes:

Hans> I got the email below from Robert Corbett: He says he picked the
Hans> semantic parser Berkeley Yacc about 12 years ago, and does not
Hans> mind taking out of Bison as well.

>> X-From_: address@hidden Thu Jul 12 23:27:31 2001
>> Date: Thu, 12 Jul 2001 14:27:25 -0700 (PDT) From: Robert Corbett
>> <address@hidden> Subject: Re: Bison's Semantic Parser
>> To: address@hidden Reply-to: Robert Corbett
>> <address@hidden> MIME-version: 1.0
>> 
>>> Bison still has a %semantic_parser feature, which uses the hairy
>>> skeleton file. But the so generated parser appears to be broken.
>>> 
>>> What do feel about taking out this %semantic_parser feature and
>>> what is connected to it?
>> I have no problem with it.  I deleted all that code from Berkeley
>> Yacc about 12 years ago.  As I said in my previous e-mail, the only
>> use for that code is for the static semantics directed error
>> recovery scheme I developed during my doctoral research.
>> 
>> Sincerely, Bob Corbett

Hans> So in view of that, the comments by Paul Hilfinger, and the fact
Hans> that no-one seems to have known how to use it for the last
Hans> decade at least, I think it is best to simply take it out of
Hans> Bison.

Hans> Then one may add some other features to Bison (in the form of %
Hans> options), say based on later Berkeley research, but that's gonna
Hans> be a whole another story.

Myabe it's solely for my own interest, but I would like not to bury
this too soon.  I'm fine with hiding the feature (which is actually
mostly already the case), and take some time to think about it.

Right now, we focus onto releasing 1.29.