Re: C threads, v. 6.2

[Szabolcs Nagy <nsz@port70.net>]

* Jens Gustedt <jens.gustedt@inria.fr> [2014-08-28 11:40:50 +0200]:
> Am Mittwoch, den 27.08.2014, 19:46 -0400 schrieb Rich Felker:
> > On Thu, Aug 28, 2014 at 12:11:45AM +0200, Jens Gustedt wrote:
> > > +#define _Nonnull(...) __attribute__((__nonnull__(__VA_ARGS__)))
> > > +#else
> > > +#define _Nonnull(...)
> > > +#endif
> > 
> > If we want to use attributes like this, it's a separate change from
> > adding threads. But I'm fairly much against doing it anyway since the
> > compiler already knows how to make these warnings for standard
> > functions, without help from the headers.
> 
> There will be a while, until compilers know these things for C11
> thread functions, I think.
> 
> BTw, I was tempted to do it the "C" way by using things like
> `[static 1]` instead of `*`, but then I thought you wouldn't like
> that.
> 

hm i think [static 1] would be better than a gcc specific attribute

but gcc <=4.9 and clang <=3.1 don't seem to warn about 0 argument then
(we could still use it though)

probably should be reported to gcc devs

> > > +int thrd_sleep(const struct timespec *req, struct timespec *rem)
> > > +{
> > > +  int ret = __syscall(SYS_nanosleep, req, rem);
> > > +  // compile time comparison, constant propagated
> > > +  if (EINTR != 1) {
> > > +    switch (ret) {
> > > +    case 0: return 0;
> > > +      /* error described by POSIX:                                    */

pls use tabs for those of us who are used to 8 width indent