Re: access to parser stack in SMIE

[Stephen Leake <stephen_leake@member.fsf.org>]

Stefan Monnier <monnier@iro.umontreal.ca> writes:

>> I'll still play with semantic; there are front-end tools like code
>> completion and function call template that rely on it. But it's less
>> urgent.
>
> Semantic support would be really neat, yes.
> [ Hopefully, at some point SMIE and Semantic will get closer, to the
>   point where a given language mode doesn't need to give 2 different
>   specifications of the language.  ]

You could write a translator from the bison BNF format to the current
SMIE BNF format, that should not be too hard. 

But I suspect that the subset of the language represented in the BNF
will be different for the two; the tools that semantic supports (_not_
indentation) require different things. Detailed lists of parameter names
(for function template/code completion), for one, which SMIE does not
need. I'll know more after I finish this current Ada mode project.

Neither set of tools wants a _complete_ BNF.

Automating a BNF filter to extract just the right subset for each tool
would be a really neat feat (and probably pass the Turing test :).

-- 
-- Stephe