About concern 10 (aka ES-047), please note that saying we discussed this before does not solve the problem.

We have a problem. It is a problem for language teachability. It leads to UB. Mitigations are just workarounds. There is no solution to that problem at this point.

That might be acceptable for a white-paper or a Technical Specification.

In my humble opinion, this is not acceptable for an International Standard.

On Fri, Oct 17, 2025 at 1:07 PM Timur Doumler via SG21 <sg21@lists.isocpp.org> wrote:

On 17 Oct 2025, at 13:26, Daniela Engert via SG15 <sg15@lists.isocpp.org> wrote:
It is. Otherwise you'd get UB with 'observe' contract evaluation semantics.

See P3846R0, Concern 10 "Observing consecutive contract assertions is dangerous".

With regards to why we allow pre and post to be evaluated multiple times, this is explained in quite some detail in P2899R1 Section 3.5.7. Also, Ville wrote a great motivation paper for this design decision a while ago that I recommend you check out: P3264R1.

Hope this helps.

Cheers,
Timur
_______________________________________________
SG21 mailing list
SG21@lists.isocpp.org
Subscription: https://lists.isocpp.org/mailman/listinfo.cgi/sg21
Link to this post: http://lists.isocpp.org/sg21/2025/10/11326.php