Re: [PATCH] Include <memory> from system.h (PR bootstrap/82610)

[David Malcolm <dmalcolm at redhat dot com>]

On Mon, 2017-10-23 at 15:51 +0200, Richard Biener wrote:
> On Mon, Oct 23, 2017 at 2:58 PM, David Malcolm <dmalcolm@redhat.com>
> wrote:
> > On Sun, 2017-10-22 at 09:28 +0200, Gerald Pfeifer wrote:
> > > On Thu, 19 Oct 2017, David Malcolm wrote:
> > > > > In file included from /scratch/tmp/gerald/gcc-
> > > > > HEAD/gcc/unique-
> > > > > ptr-tests.cc:23:
> > > > > In file included from /scratch/tmp/gerald/gcc-
> > > > > HEAD/gcc/../include/unique-ptr.h:77:
> > > > > In file included from /usr/include/c++/v1/memory:629:
> > > > > /usr/include/c++/v1/typeinfo:199:2: error: no member named
> > > > > 'fancy_abort' in namespace 'std::__1'; did you mean simply
> > > > > 'fancy_abort'?
> > > > >         _VSTD::abort();
> > > > >         ^~~~~~~
> > > > > /usr/include/c++/v1/__config:390:15: note: expanded from
> > > > > macro
> > > > > '_VSTD'
> > > > > #define _VSTD std::_LIBCPP_NAMESPACE
> > > > 
> > > > There seem to have been similar problems on OS X:
> > > >   https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82610
> > > 
> > > Yes, I believe it's the same actually (unearthed by clang as
> > > system
> > > compiler).
> > > 
> > > > The proposed fix there is to include <memory> in system.h,
> > > > which
> > > > presumably would fix this also.
> > > 
> > > That appears to work around the bootstrap failure on my tester as
> > > well.
> > > 
> > > How can we go about fixing this in the tree?
> > > 
> > > Gerald
> > 
> > Here's the patch by fxcoudert from the PR (plus a ChangeLog entry)
> > 
> > Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu;
> > reported by fxcoudert as fixing the issue on darwin and by
> > Gerald as fixing the issue on "newer versions of FreeBSD that use
> > clang 4.0 as system compiler".
> > 
> > OK for trunk?
> 
> Not entirely happy as unique-ptr.h doesn't use <memory> but well.

I'm not sure I understand you here.

include/unique-ptr.h has:

  #if __cplusplus >= 201103

  /* In C++11 mode, all we need is import the standard
     std::unique_ptr.  */
  template<typename T> using unique_ptr = std::unique_ptr<T>;

  /* Pull in move as well.  */
  using std::move;

  #else /* C++11 */

  ...etc..., most of the file, the pre-C++11 implementation

So in C++11 and later it's using std::unique_ptr, for which, as I
understand it <memory> is the standard include, e.g.: 
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2012/n3337.pdf
says in (20.6.2 Header <memory> synopsis [memory.syn]):

"The header <memory> defines several types and function templates that
describe properties of pointers and pointer-like types, manage memory
for containers and other template types, [...] The header also defines
the templates unique_ptr, shared_ptr, weak_ptr, and various template
functions that operate on objects of these types (20.7)."

Would you prefer the includes of <memory> in gcc/system.h and
include/unique-ptr.h to be guarded by #if __cplusplus >= 201103 ? (not
sure if it works yet, but I can try it)

Thanks
Dave


> Ok to unbreak bootstrap.
> 
> Thanks,
> Richard.
> 
> > Sorry again about the breakage.
> > 
> > gcc/ChangeLog:
> >         PR bootstrap/82610
> >         * system.h [__cplusplus]: Include <memory>.
> > ---
> >  gcc/system.h | 1 +
> >  1 file changed, 1 insertion(+)
> > 
> > diff --git a/gcc/system.h b/gcc/system.h
> > index f0664e9..d6e1637 100644
> > --- a/gcc/system.h
> > +++ b/gcc/system.h
> > @@ -233,6 +233,7 @@ extern int errno;
> >  # include <vector>
> >  #endif
> >  # include <cstring>
> > +# include <memory>
> >  # include <new>
> >  # include <utility>
> >  #endif
> > --
> > 1.8.5.3
> >