Date: Wed, 12 Aug 2020 22:25:54 +0300
On Wed, 12 Aug 2020 at 22:20, Bjarne Stroustrup via Liaison
<liaison_at_[hidden]> wrote:
>
> Grammar factoring should be driven by needs of semantic and use cases, not merely ease of expressing syntax.
Here's the thing though: this change makes the language more regular
and simpler. It makes it possible to put labels
anywhere in a function scope, and just jump to them (jumping into some
scopes is still not allowed, no change there).
No magic rules like "oh but it can't be the very last thing, you
need to insert this no-op after it, not because we can't parse it and
analyze it, but because we just.. ..say so".
Again, it's probably not covering use cases that we think a large
volume of C++ programmers run into. But
it has some rather attractive qualities.
<liaison_at_[hidden]> wrote:
>
> Grammar factoring should be driven by needs of semantic and use cases, not merely ease of expressing syntax.
Here's the thing though: this change makes the language more regular
and simpler. It makes it possible to put labels
anywhere in a function scope, and just jump to them (jumping into some
scopes is still not allowed, no change there).
No magic rules like "oh but it can't be the very last thing, you
need to insert this no-op after it, not because we can't parse it and
analyze it, but because we just.. ..say so".
Again, it's probably not covering use cases that we think a large
volume of C++ programmers run into. But
it has some rather attractive qualities.
Received on 2020-08-12 14:29:37