Boost logo

Boost :

From: Peter Dimov (pdimov_at_[hidden])
Date: 2023-05-18 22:05:29


Glen Fernandes wrote:
> On Thu, May 18, 2023 at 5:31 PM Christian Mazakas via Boost <
> boost_at_[hidden]> wrote:
>
> > We could maybe solve the publicity issue by having the implementation
> > become a typedef when the header is available.

What "publicity issue"?

> I'm fine with this, but if we go this route it might be better that Compat
> depend on no other Boost libraries.
>
> e.g. They wouldn't be able to count on the features of Boost.Assert.

No, Boost libraries should never use `assert` or `throw` directly (unless
they duplicate the relevant functionality of Assert and ThrowException.)

Note that an implementation consisting of e.g. `using std::latch;` does not
use `assert` or `throw` so it doesn't violate the above rule.

Interestingly, said implementation would actually fail our current
latch_test.cpp which uses BOOST_ENABLE_ASSERT_HANDLER to check
precondition violations, but we'll fix that test if we ever switch to it.


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk