Date: Thu, 29 Aug 2019 19:15:50 +0200
Nathan Sidwell <nathan_at_[hidden]> writes:
> Which one of those claims:
>
> 1) They are C++ source files,
> 2) we already have a convention for suffixing C++ source files,
> 3) that is the default position
>
> do you disagree with?
That depends on what exactly you mean by "C++ source file". Is a
header a C++ source file?
Perhaps we should rather use the more precisely defined term
"translation unit". Before modules we had one kind of translation
unit with a bunch of commonly used extensions (.cpp, .cxx, .cc, etc).
Now we have several different kinds. One can argue that they are
common enough to all use the same extension. Or one can argue that
they are different enough to warrant a distinction.
In a modular world an interface unit[*] replaces a header and, IMO,
similar to headers, it will be useful to both tools and humans
if they are easily distinguishable from other translation units.
Or do you think having distinct extension for headers was a
mistake?
> Which one of those claims:
>
> 1) They are C++ source files,
> 2) we already have a convention for suffixing C++ source files,
> 3) that is the default position
>
> do you disagree with?
That depends on what exactly you mean by "C++ source file". Is a
header a C++ source file?
Perhaps we should rather use the more precisely defined term
"translation unit". Before modules we had one kind of translation
unit with a bunch of commonly used extensions (.cpp, .cxx, .cc, etc).
Now we have several different kinds. One can argue that they are
common enough to all use the same extension. Or one can argue that
they are different enough to warrant a distinction.
In a modular world an interface unit[*] replaces a header and, IMO,
similar to headers, it will be useful to both tools and humans
if they are easily distinguishable from other translation units.
Or do you think having distinct extension for headers was a
mistake?
Received on 2019-08-29 12:17:58