On 22 Jun 2026 16:27, Yexuan Xiao via Std-Proposals wrote:
> What I actually care about is code that intentionally serves C++ (with
> UCRT and Clang SIMD headers being typical examples), and as I have
> emphasized, some code that is entirely written in C++ also misuses
> static inline. I will not insist that libraries which do not like C++
> should support C++; in that case, C++ users will need to do some extra
> work anyway.

Your proposed solution affects all C++ users, both interested in modules
and C libraries. What I'm saying is that it is not suitable for the
latter category of C++ users.

And BTW, the amount of work needed to do to use C libraries without
`extern "C"` is much lower than that would be needed to get rid of the
`static` keywords.

> ------------------------------------------------------------------------
> *From:* Std-Proposals <std-proposals-bounces_at_[hidden]> on behalf
> of Andrey Semashev via Std-Proposals <std-proposals_at_[hidden]>
> *Sent:* Monday, June 22, 2026 21:14
> *To:* std-proposals_at_[hidden] <std-proposals_at_[hidden]>
> *Cc:* Andrey Semashev <andrey.semashev_at_[hidden]>
> *Subject:* Re: [std-proposals] Proposal: Deprecate namespace-scope
> declarations that are declared both static and inline
>
> On 22 Jun 2026 15:04, Yexuan Xiao via Std-Proposals wrote:
>> Your statement doesn't solve any problem. For users of modules, they
>> either have to abandon this library or seek improvements from it.
>> Basically, these C libraries only need to stop using static when the
>> __cplusplus macro is defined, and that would solve the issue. There's no
>> need to abandon C++ users, nor do C++ users need to abandon them. Why do
>> you assume that these C libraries will never be improved?
>
> I think that the argument from C library developers will be that the C++
> modules problem is a C++ problem, not a C problem or their library
> problem. And I would agree with their argument. There are libraries that
> don't even bother adding `extern "C"` to their headers, I don't expect
> them to make an extra effort to add a conditional compilation macro to
> wrap `static`.
>
> Bottom line is C compatibility is a very important use case for C++, so
> C++ should remain as compatible with C as possible. If any C++ features
> come into conflict with this compatibility, these features should adapt.