Re: Bison & flex on start conditions

[Hans Aberg]

As you point out, the Bison generated parser uses a lookahead token  
that may or may not be required in each specific parse, embedded into  
the LALR(1) algorithm used to create it, and there appears to be no  
easy way around that. One can think of inserting some user code at  
the place where the Bison parser asks for a new token, though. But I  
am not sure it will give the effect you are asking for.

On 13 Jan 2006, at 14:10, Steve Murphy wrote:

> Hello--
>
> Been playing with bison/flex for years; think they're great.
>
> There are few interoperability issues between them, however.
>
> One that would be nice to smooth out would be the start-condition
> capability in flex... and how to trigger it from the bison grammar end
> of things.
>
> Bison can call a function in the flex world, that will run the
>  BEGIN(XXX) and set the start condition for flex. But usually, it's
> already too late... bison has it's lookahead token. Flex already
> evaluated the token under the old regime, and returned a token.
>
> Has anybody got a solution how to trigger flex start conditions from
> bison, so everything works out nice?
>
> murf
>
> --
> Steve Murphy <address@hidden>
> Electronic Tools Company
>
>
>
> _______________________________________________
> address@hidden http://lists.gnu.org/mailman/listinfo/help-bison