On Wed, Apr 17, 2024 at 11:49 AM Rhidian De Wit wrote:
>
> I would agree with Brian and others who have said this, but this is getting a bit spammy.
> The constant updates flooded with small details about the internal working of the Microsoft
> ABI can be interesting, but I do not think this mailing list is the correct place for it. A blog
> post would be more suitable.
>
> As Brian said, I think the why is far more important than the how, especially because compiler
> vendors will most likely be able to implement all of this on their own.


There are two possible papers to submit:

    (1) Add functionality to std::any to get a base class

    (2) Add functionality to std::any to get a base class without
causing an ABI break

We're aiming for Option No. 2 and that is why it's necessary to find
out whether this is possible on the Microsoft compiler. Going for
Option No. 1 would just lead to a lot of time and effort wasted on
liaising with compiler vendors about whether it's possible without an
ABI break. So we're eliding that liaison by figuring it out by
ourselves and going for Option No. 2.

I remember 30 years ago, growing up in Dublin, my family had a coat
hanger hanging out the back of the television in the front room. I
always wondered how the television broadcaster could come out with
statements like "Lastnight 1.2 million people tuned in to see the
season finale of Friends" -- I mean how could they possibly know what
frequencies my television is focusing on from the signal coming in
from the coat hanger? Today in 2024 we've moved on to digital and we
can send emails with "request a read receipt", but I'm still wondering
how you and Brian know who is and who is not reading, and who is
taking interest and who is not taking interest, in posts on this
mailing list.