Re: [C++ coroutines 6/6] Testsuite.

[JunMa <JunMa at linux dot alibaba dot com>]

在 2019/11/20 下午9:11, JunMa 写道:
> 在 2019/11/20 下午7:22, Iain Sandoe 写道:
> > Hello JunMa,
> >
> > JunMa <JunMa@linux.alibaba.com> wrote:
> >
> > > 在 2019/11/17 下午6:28, Iain Sandoe 写道:
> > > I find that the patches donot support 'for co_await'. And it is
> > > quiet simple to implement range based 'for co_await' based on your
> > > patches, since it's just need few more works on range for source to
> > > source transform. Any reason for that?
> > If I understand your question correctly,
> >
> > for example TS n4775, there was:
> >
> > [stmt.stmt]
> >    ….
> >    for co_await ( for-range-declaration : for-range-initializer ) 
> > statement
> >
> > yes?
> >
> > This was removed by a later committee resolution, and was *not* merged
> > to the C++20 Working Draft (I am currently working to n4835).
> >
> > So, the reason it is not implemented in GCC at present, is that it is 
> > not clear
> > exactly what form it might take if introduced in some future proposal 
> > for
> > enhanced coroutines.
> >
> > hope that answers the question,
> > thanks
> > Iain
> Hi Iain,
>     Thanks, that make sense.
>
> Regards
> JunMa
Hi Iain

In current implementation of the allocating storage for coroutine,
it does not follow the rules in n4835 which look up in the scope of
the promise type first. Is there any reason? If not, would youimplement
this following the rules of n4835? Thanks.

Regards
JunMa