Date: Sun, 21 Jun 2026 08:02:01 -0400
On Sunday, 21 June 2026 01:29:09 Eastern Daylight Time Simon Schröder wrote:
> > On Jun 20, 2026, at 10:46 PM, Thiago Macieira via Std-Proposals
> > <std-proposals_at_[hidden]> wrote:>
> Well, we already have special rules for member functions that are defined
> inline. We could see PEMs the same way.
I agree so long as they are the same rules: functions whose implementations
are inside of the class body. If the implementation is outside, then it
shouldn't be inline and should not therefore have internal linkage.
Why do we want different rules? What's the gain? What are we trying to solve?
And at what cost, given that it increases complexity if they are different from
the main class?
> > On Jun 20, 2026, at 10:46 PM, Thiago Macieira via Std-Proposals
> > <std-proposals_at_[hidden]> wrote:>
> Well, we already have special rules for member functions that are defined
> inline. We could see PEMs the same way.
I agree so long as they are the same rules: functions whose implementations
are inside of the class body. If the implementation is outside, then it
shouldn't be inline and should not therefore have internal linkage.
Why do we want different rules? What's the gain? What are we trying to solve?
And at what cost, given that it increases complexity if they are different from
the main class?
-- Thiago Macieira - thiago (AT) macieira.info - thiago (AT) kde.org Principal Engineer - Intel Data Center - Platform & Sys. Eng.
Received on 2026-06-21 12:02:17
