Date: Thu, 23 Oct 2025 16:03:10 -0700
Roger:
> > > > But that was closed: “SG21 has consensus against pursuing this
> > > > paper any further.”
Ville:
> > > We are in CD ballot, and we have NB comments the request of which is
> > > almost literally to reopen that decision.
Herb:
> > Yes, and even besides that as soon as contracts went to EWG (in Tokyo I
> > think) I've been regularly insisting to the SG21 and EWG chairs that the
> > opinion that matters is in EWG, which as a larger group can have a different
> > opinion from SG21. An SG stops owning (and being a gatekeeper of) a paper
> > when it sends the paper onward to EWG.
> >
> > So since Tokyo, for any change proposal to P2900 contracts (as opposed to
> > future extensions), I've been regularly asking that the discussion must come
> > to EWG, regardless of whether or not the chairs might choose additionally
> > schedule any SG21 discussion to generate opinions to offer as advisory input
> > to EWG.
Ville:
> So, you're saying that the closing of
> https://github.com/cplusplus/papers/issues/2277 after an SG23 discussion in Sofia
I think you mean SG21?
> was procedurally incorrect?
I wasn't aware of this at the time (can't keep up with all papers), but some thoughts -- I could be overlooking something: As an extension (it's proposing an additional feature, not just a change) it was already procedurally late for C++26's new-feature deadline. To the extent it's a change proposal to the current design that's now in EWG, it would ideally have been going (also) to EWG with the usual caveat of "schedule permitting" in case EWG doesn't have bandwidth to get to all papers. But I could be overlooking something; this is a quick reply to your question.
Regardless of history, though, my understanding is that this paper's suggestion will be discussed in Kona EWG -- in the context of not shipping contracts in C++26 to see if we want to try this alternative design or some other alternative design instead -- because (I think?) it's suggested by your new paper and/or a NB comment which will be discussed.
Chairs, please correct anything I got wrong!
Herb
> > > > But that was closed: “SG21 has consensus against pursuing this
> > > > paper any further.”
Ville:
> > > We are in CD ballot, and we have NB comments the request of which is
> > > almost literally to reopen that decision.
Herb:
> > Yes, and even besides that as soon as contracts went to EWG (in Tokyo I
> > think) I've been regularly insisting to the SG21 and EWG chairs that the
> > opinion that matters is in EWG, which as a larger group can have a different
> > opinion from SG21. An SG stops owning (and being a gatekeeper of) a paper
> > when it sends the paper onward to EWG.
> >
> > So since Tokyo, for any change proposal to P2900 contracts (as opposed to
> > future extensions), I've been regularly asking that the discussion must come
> > to EWG, regardless of whether or not the chairs might choose additionally
> > schedule any SG21 discussion to generate opinions to offer as advisory input
> > to EWG.
Ville:
> So, you're saying that the closing of
> https://github.com/cplusplus/papers/issues/2277 after an SG23 discussion in Sofia
I think you mean SG21?
> was procedurally incorrect?
I wasn't aware of this at the time (can't keep up with all papers), but some thoughts -- I could be overlooking something: As an extension (it's proposing an additional feature, not just a change) it was already procedurally late for C++26's new-feature deadline. To the extent it's a change proposal to the current design that's now in EWG, it would ideally have been going (also) to EWG with the usual caveat of "schedule permitting" in case EWG doesn't have bandwidth to get to all papers. But I could be overlooking something; this is a quick reply to your question.
Regardless of history, though, my understanding is that this paper's suggestion will be discussed in Kona EWG -- in the context of not shipping contracts in C++26 to see if we want to try this alternative design or some other alternative design instead -- because (I think?) it's suggested by your new paper and/or a NB comment which will be discussed.
Chairs, please correct anything I got wrong!
Herb
Received on 2025-10-23 23:03:12