<string_view> for libc++

[Yaron Keren]

Regarding the __min (and __max) macros, other than being defined they are
not used in MinGW, so would it make sense to enhance __undef_min_max to
undefine __min and __max (if defined) even without a warning in addition to
undefining min and max?

Yaron



2014/1/7 G M <gmisocpp at gmail.com>

> Hi Marshall
>
> Some early feed back for your patch:
>
> 1. string_view implements it's own __min function, but this is
> unfortunately a macro in mingw's stdlib.h
>
> and therefore compilation yields these type of error's with clang++ and
> g++ using mingw:
> ----
> /libcxx/include\string_view:181:18: error: expected ')'
>     constexpr _T __min ( const _T& __t1, const _T& __t2 )
>
>                  ^
> c:\mingw\bin\..\lib\clang\3.5\../../../x86_64-w64-mingw32/include\stdlib.h:113:26:
> note: expanded from macro '__min'
> #define __min(a,b) (((a) < (b)) ? (a) : (b))
> ----
>
> Changing __min to ___min which allowed it to compile for me.
>                          ^
>
> 2. With the above change in place, I find I c++1y is required. Makes
> sense, but using string_view without a c++1y compiler might be a worthwhile
> consideration if it's easy.
>
> /libcxx/include\string_view:241:13: warning: use of this statement in a
> constexpr function is a C++1y extension
>       [-Wc++1y-extensions]
>             if (__pos >= size ())
>             ^
> /libcxx/include\string_view:292:23: warning: variable declaration in a
> constexpr function is a C++1y extension
>       [-Wc++1y-extensions]
>             size_type rlen_ = ____min ( __n, size() - __pos );
>                       ^
> /libcxx/include\string_view:290:13: warning: use of this statement in a
> constexpr function is a C++1y extension
>       [-Wc++1y-extensions]
>             if (__pos > size())
>
> 3. g++ reports this:
> --
> In file included from sv.cpp:1:0:
> /libcxx/include/string_view:199:53: error: declaration of 'typedef
> std::experimental::__library_fundamentals::basic_strin
> _view<_CharT, _Traits>::const_reverse_iterator
> std::experimental::__library_fundamentals::basic_string_view<_CharT, _Trai
> s>::reverse_iterator' [-fpermissive]
>          typedef const_reverse_iterator              reverse_iterator;
>                                                      ^
> In file included from /libcxx/include/memory:600:0,
>                  from /libcxx/include/algorithm:627,
>                  from /libcxx/include/string:439,
>                  from /libcxx/include/string_view:173,
>                  from sv.cpp:1:
> /libcxx/include/iterator:529:29: error: changes meaning of
> 'reverse_iterator' from 'class std::__1::reverse_iterator<cons
>  _CharT*>' [-fpermissive]
>  class _LIBCPP_TYPE_VIS_ONLY reverse_iterator
> --
> I commented out line 199 and it compiled. i.e.:
>         //typedef const_reverse_iterator              reverse_iterator;
>
> That's it for now.
>
> I couldn't reply directly to your original post on this subject because I
> couldn't find your post in my inbox, only on the archives list. I don't
> know how to get around that problem when it happens.
>
> Hope this helps.
>
> _______________________________________________
> cfe-commits mailing list
> cfe-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20140107/f1477322/attachment.html>