Correct.  The module unit may have a “global module fragment” which is never imported itself.

 

-- Gaby

 

From: SG15 <sg15-bounces@lists.isocpp.org> On Behalf Of Daniela Engert via SG15
Sent: Friday, May 13, 2022 7:57 AM
To: sg15@lists.isocpp.org
Cc: dani <dani@ngrt.de>
Subject: Re: [SG15] Revisiting terminology, framing for meeting on 2022-05-13

 

Am 13.05.2022 um 16:37 schrieb Daniel Ruoso via SG15:

Em sex., 13 de mai. de 2022 às 10:31, Gabriel Dos Reis
<gdr@microsoft.com> escreveu:
Agreed, but what can this body (the study group) say more than "your
implementation needs to be unambiguous about what it considers
importable"?
 
I think having a clear terminology is a starting point. But we'll
definitely need to get into more details than that, including for
things like "when you distribute a library, how do you tell what are
the "importable" headers you ship.
 
Also, we don't import module units - we never do.
 
I don't think I understand what you mean. Would you mind clarifying?

I'm pretty sure he wants to say that you import the BMI rather than the module unit as the former is a compilation artefact of the latter.

Ciao
  Dani

 

--
PGP/GPG: 2CCB 3ECB 0954 5CD3 B0DB 6AA0 BA03 56A1 2C4638C5