Re: if-then-else ambiguity

Umm wrote:
>
> Would it be a correct way, of solving a shift reduce problem in
> yacc/bison of the if-then-else by defining a %left precedence (the
> biggest one) to the else?? that way, we are sure (we were already but
> at least there will not be a warning) that in the presence of the
> else, it will be immediately reduced and not shifted, thus making it
> belong to the closest if.
>
> Is this correct??


I think the resolution must be to SHIFT. I believe this is the default
bahavior.


SHIFT leaves IF-THEN-ELSE on the stack which is then reduced.


REDUCE reduces the IF-THEN on the stack, leaving the ELSE as input to
match an Outer IF-THEN.


-Duncan
--

Return to the comp.compilers page.
Search the comp.compilers archives again.