[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20120907211738.GT27715@brightrain.aerifal.cx>
Date: Fri, 7 Sep 2012 17:17:38 -0400
From: Rich Felker <dalias@...ifal.cx>
To: musl@...ts.openwall.com
Subject: Re: C11 api coverage
On Fri, Sep 07, 2012 at 11:47:44AM -0700, Isaac Dunham wrote:
> On Wed, 5 Sep 2012 08:40:10 +0200
> Szabolcs Nagy <nsz@...t70.net> wrote:
>
>
> > ok i removed annex k symbols
> >
> > (and updated the c11 coverage table)
> Basically, CMPLX*, noreturn, a few types in uchar.h/time.h, and several
> optional features.
Aside from stdnoreturn.h, noreturn stuff is done.
uchar.h and time.h have some nontrivial functions we need to add, I
think...
> Setting __STDC_NO_{ATOMICS,THREADS}__ should make musl almost fully
> conformant...
It's impossible for musl to set these; that's up to the compiler, and
it would have to be done in the gcc specfile or builtin gcc specs.
Of course we can just publish that the command to invoke the C11
compiler is:
gcc -std=c11 -D__STDC_NO_ATOMICS__ -D__STDC_NO_THREADS__
However, I'd actually like to get atomics supported soon, at least on
GNUC compilers that have the right __builtin or __sync builtins. (Or
adapting atomic.h to be a public header, but I think the stuff in
stdatomic.h has to be somewhat type-generic, which makes it hard...)
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.
