Am Donnerstag, den 29.04.2021, 20:51 -0400 schrieb Bjarne Stroustrup:
> Treating built-in arrays as "legacy only" (whatever that means" is IMO a
> non-starter, but 40+ years after the C manual was cloned to start the
> C++ manual, incompatibilities or confusing formulations could have crept
> into either or both standards documents.

With treating them as "legacy only" I mean something like this:

p2128r3

"However C arrays are not widely used by C++,
spending time on them might therefore not be useful."

Where the authors propose that C arrays should not be further
evolved, i.e. not get the same syntax extension for multi-dim.
arrays in the future.

Compare this to

P1997

which proposel mostly simple changes which would make them fit
in better with the rest of C++'s types.


To me these two paper represent two directions C++ could go:

1.) Relegating C arrays to niche uses which I will
- in the long run - cause more problems for C and C++
compatiblity.

2.) Or better integrating them into the language.



Best,
Martin









> On 4/28/2021 3:52 AM, Uecker, Martin via Liaison wrote:
> > This was not primarly a discussion about the advantages of each
> > programming language but about native arrays which are part of both
> > languages. It is also highly relevant to this study group
> > which approach C++ takes here. If C++ decides to treat
> > these as a legacy concept only it will be difficult to
> > maintain compatibility in this area.