Two fold. The build system should be preventing this sort of thing, and secondly because some compilers have a fallback where they just try to do the best they can and treat these sequences they don’t understand as “unknown”. I’m trying not to break existing behavior, basically :v

Izzy

On Mon, Sep 9, 2019 at 07:01, Botond Ballo <botond_ballo@yahoo.ca> wrote:
> When a compiler is given a target tuplet that it is unable to rectify or target, the invocation
> of the compiler is considered ill-formed, no diagnostic required.

Why not require a diagnostic for this?

Thanks,
Botond

On Monday, September 9, 2019, 8:24:02 AM EDT, Isabella Muerte via SG15 <sg15@lists.isocpp.org> wrote:


Hello all,

As today is the cut off for papers for CppCon, I've written one small brief one I'd like to have discussed regarding "target tuplets". I know we've been on a modules kick recently, but this is tooling related, and I would argue deserves some time. I've attached a copy and (following discussion) a P version will be mailed for Belfast.



Izzy
_______________________________________________
SG15 mailing list
SG15@lists.isocpp.org
https://lists.isocpp.org/mailman/listinfo.cgi/sg15