[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 13 Oct 2014 14:06:22 -0400
From: Rich Felker <dalias@...c.org>
To: musl@...ts.openwall.com
Subject: Re: [PATCH] Convert some is* macros to inline functions
On Mon, Oct 13, 2014 at 09:00:56PM +0300, Sergey Dmitrouk wrote:
> On Mon, Oct 13, 2014 at 07:35:14AM -0700, Jens Gustedt wrote:
> > Am Montag, den 13.10.2014, 17:20 +0300 schrieb Sergey Dmitrouk:
> > In section 7.1.4 the C standard explicitly says:
> >
> > > Any function declared in a header may be additionally implemented
> > > as a function-like macro defined in the header, ...
>
> Didn't look there, you're right. I was checking description of headers
> instead.
>
> > > Please find the attached patch that proposes replacing these macros
> > > with inline functions.
> >
> > I don't think that this is necessary.
> >
> > They only advantage of inline functions, here would be that the
> > conversion of the arguments would be done with implicit conversions
> > instead of casts. (For the macros this could be achieved by using
> > compound literals instead of casts, but well...)
>
> It's not necessary for C (as I know now), but it's required by ะก++
> standard in 17.6.1.2:
>
> > Names that are defined as functions in C shall be defined as functions
> > in the C++ standard library. 175)
>
> > 175) This disallows the practice, allowed in C, of providing a masking
> > macro in addition to the function prototype. The only way to achieve
> > equivalent inline behavior in C++ is to provide a definition as an
> > extern inline function.
>
> Current headers do not conform to C++ when included as <header.h>, and
> that's what I'm trying to fix.
>
> Would you consider a version that uses inline functions only when
> __cplusplus is defined? There is already 'extern "C"', so I guess it
> makes sense.
Suppressing the macros for C++ would be acceptable if this is required
(it's not clear to me; my understanding was that it's only required
for the <c_____> headers, not the <______.h> ones, and that the
wrappers for the former already take care of removing the macros). The
external inline functions are probably not acceptable; there are all
sorts of ugly issues with extern inline function support that I don't
want to deal with.
Rich
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
