On 28/10/2023 10.18, Corentin wrote:
> Alisdair has a paper aiming to change the terminology https://www.open-std.org/jtc1/sc22/wg21/docs/papers/2023/p2843r0.pdf
(it does not resolve any issue afaict, just remove the term "undefined behavior")

I'd much rather remove the unspecified-ness here entirely.
That requires a review of existing practice for each item,
though, to determine the amount of implementation breakage
we're going to cause.

> I'm not sure how we could do that but I'm hoping we could reach an agreement (maybe a poll in SG12?) that phases 1-6 of translations never cause undefined behavior (as understood by
> C++) and therefore not under the purview of SG12.

That's a plausible avenue.

> (in the general case, preprocessing and lexing related papers are very likely to require consulting either SG22, SG16, or both)

Yes.

Jens