Re: merge function called AFTER user actions

[Joel E. Denny]

On Mon, 26 Sep 2005, Derek M Jones wrote:

> > My point here is that the current functionality seems natural and necessary 
> > for all GLR parsers.  Do you disagree?
>
> Either before or after is necessary.  Both would be helpful.
> I'm not sure the term 'natural' can be applied to a human
> invention.

I see now that I wasn't being very careful with my language in that email. 
Sorry for the confusion.

By `seems natural', I meant `seems natural to humans who work with GLR 
parsers'.  I wasn't actually referring to Nature, of course.

In place of `for all GLR parsers', I was really thinking `for all GLR 
parsers that need to merge semantic values'.  Clearly, there are other 
ways to resolve ambiguities: syntax errors and %dprec.

I suppose the functionality you suggest could be used in conjunction with 
any of these approaches to ambiguity resolution.

Joel