C++ Logo


Advanced search

Re: [std-proposals] Every variable is volatile, everything is laundered, no optimisation

From: Jonathan Wakely <cxx_at_[hidden]>
Date: Mon, 28 Aug 2023 10:03:16 +0100
On Sun, 27 Aug 2023, 23:45 Gašper Ažman via Std-Proposals, <
std-proposals_at_[hidden]> wrote:

> Jason is completely correct, but I want to state it more succinctly.
> You are asking for help from the wrong people.
> The standard does *not* specify translation to assembly. The standard
> specifies the behaviour of the C++ abstract machine, which, if you get
> down to it, is a segmented addressing model architecture with very
> loose concurrency forward progress model and allowances for lowering
> to a linear addressing model, but even worse, *actaually*, it's a
> parametrized abstract machine where implementations have to fill in
> all the blanks where it says "implementation-defined".
> The standard therefore literally doesn't deal with translation to any
> given assembly, and you'd have to write thousands of pages of text to
> specify what you want for x86, and then it wouldn't get passed.
> Remember, the standard doesn't even say that vtables exist!
> It's *the implementations* that define lowering into assembly, given
> what the standard says the syntax means. GCC can ignore what the
> standard says if its users don't mind. The standard doesn't even have
> words to do so.
> Talk to the GCC people and there you might find a helpful ear. Some
> committee people might even just change hats and be gcc people for
> you. I do have to repeat that on *this* reflector, you're just wasting
> your time.

And everybody else's time.

As usual.

What you want is not expressible *in the standard* without
> decades of work nobody wants to do and nobody wants to implement
> afterwards.
> Abandon hope, ye who enter here.
> --
> Std-Proposals mailing list
> Std-Proposals_at_[hidden]
> https://lists.isocpp.org/mailman/listinfo.cgi/std-proposals

Received on 2023-08-28 09:03:32