Re: [TCLCORE] TIP 282 status

[Peter S <pet...@gm...>]

On 19/02/2017, avl <av...@lo...> wrote:
>
> Florent's expr is an (imho bizarre) reconciliation of expr towards plain
> Tcl. As such, it is of course strictly incompatible with Tcl's expr,
> and has zero chances of ever getting accepted as replacement for expr,
> but it is some unambiguous language on its own.

The same unambiguity could be achieved without turning the syntax into
a bizarre ordeal... Example:

  expr { foo@( ... ) = ... }  calculate expression, and use it as array index

Which happens to be identical to his foo@(...) expression.

Then writing

  expr { foo@("entier,$y") = ... }

has the same effect as his foo@"entier,$y", and

  expr { foo@({entier,$y}) = ... }

has the same effect as his foo@{entier,$y}.

No need to change the syntax completely, for a magic char before the
paren to work... Which is IMO identical to your suggestion to tag
array with a special char, his choice of char being the '@' mark,
written before the open paren. Yes, that makes it unambiguous, but
even without changing the function call syntax :) That's pointless and
even irrelevant.