Am Mittwoch, dem 22.01.2025 um 15:53 +0100 schrieb Michael Matz:
> Hello,
>
> On Tue, 21 Jan 2025, Martin Uecker wrote:
>
> > > > Coudn't you use the rule that .len refers to the closest enclosing
> > > > structure
> > > > even without __self__ ? This would then also disambiguate between
> > > > designators
> > > > and other uses.
> > >
> > > Right now, an expression cannot start with '.', which provides the
> > > disambiguation between designators and expressions as initializers.
> >
> > You could disambiguate directly after parsing the identifier, which
> > does not seem overly problematic.
>
> Which way? When you allow ". identifier" as primary expression, then in
>
> struct S s = { .x = 42 };
>
> the initializer can be parsed as designated initializer (with error
> when 'x' is not a member of S) or as assignment expression like in
>
> struct T t = { foo = 42 };
>
> You need to decide which is which after seeing the ".". I'm guessing what
> you mean is that on seeing ".ident" as first two tokens inside in
> initializer-list you go the designator route, and not the
> initializer/assignment-expression route, even though the latter can now
> also start with ".ident".
What I mean is that after parsing the dot followed by an identifier x,
if x is the name of a member of the structure S which is being initialized,
it is a designator, otherwise it is an expression that uses .x to refer
to some member of an enclosing definition.
So you do not need to look further. But maybe I am missing something
else.
Martin
> But then, what about:
>
> struct U u = { .y };
>
> It's certainly not a designation anymore, but you only know after not
> seeing the '='. So here it's actually an assignment-expression. And as
> you allowed ".ident" as primary expression this might rightfully refer to
> some in-scope 'y' member of some outer struct (or give an error).
>
> Note further that you may have '{ .y[1][3].z }', which is still not a
> designation, but an expression under your proposal, whereas
> '{ .y[1][3].z = 1 }' would remain a designation. This shows that you
> now need arbitrary look-ahead to disambiguate the two. A Very Bad Idea.
>
>
> Ciao,
> Michael.
--
Univ.-Prof. Dr. rer. nat. Martin Uecker
Graz University of Technology
Institute of Biomedical Imaging
