[Bug libstdc++/56922] New: set: the default constructor should be explicit

[Bug libstdc++/56922] New: set: the default constructor should be explicit

[akim.demaille at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56922

             Bug #: 56922
           Summary: set: the default constructor should be explicit
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libstdc++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: akim.demaille@gmail.com


Hi all,

According to http://llvm.org/bugs/show_bug.cgi?id=15724, there is no default
constructor for set, only a constructor whose arguments are optional, but it
turns out that this constructor is explicit.  So, at least according to C++98
(I don't know about C++11), the following piece of code is expected to fail.


$ cat /tmp/set.cc
#include <set>
int
main()
{
  std::set<int> s = {};
}
$ g++-mp-4.8 -std=c++11 -Wall /tmp/set.cc

Clang rejects it.

$ clang++-mp-3.3 -std=c++11 -Wall /tmp/set.cc -stdlib=libc++
/tmp/set.cc:5:17: error: chosen constructor is explicit in copy-initialization
  std::set<int> s = {};
                ^   ~~
/opt/local/libexec/llvm-3.3/bin/../lib/c++/v1/set:378:14: note: constructor
declared here
    explicit set(const value_compare& __comp = value_compare())
             ^
1 error generated.

The actual case at hand is a function with an optional argument which is a set,
and I used "= {}", which is rejected by clang++.

Cheers!

[Bug libstdc++/56922] set: the default constructor should be explicit

[akim.demaille at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56922

--- Comment #1 from Akim Demaille <akim.demaille at gmail dot com> 2013-04-11 16:08:07 UTC ---
FWIW: http://cplusplus.github.io/LWG/lwg-active.html#2193

[Bug libstdc++/56922] set: the default constructor should be explicit

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56922

--- Comment #2 from Paolo Carlini <paolo.carlini at oracle dot com> 2013-04-11 16:16:07 UTC ---
I was about to reply that the issue isn't just about std::set and I was trying
to remember why we decided many years ago to *not* have the explicit in the
singled out default constructors ;) Anyway, cool, I read the active DR as
implying that we are already Ok vs the proposed resolution. Therefore we should
likely suspend this in the meanwhile.

[Bug libstdc++/56922] set: the default constructor should be explicit

[akim.demaille at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56922

--- Comment #3 from Akim Demaille <akim.demaille at gmail dot com> 2013-04-11 16:23:57 UTC ---
Agreed.  Sorry for the noise, I was not aware of this page.

[Bug libstdc++/56922] set: the default constructor should be explicit

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56922

--- Comment #4 from Jonathan Wakely <redi at gcc dot gnu.org> 2013-04-11 16:33:38 UTC ---
Yes, libstdc++ is already correct, there's nothing for us to do here

[Bug libstdc++/56922] [lwg/2193] set: the default constructor should be explicit

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56922

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |SUSPENDED
   Last reconfirmed|                            |2014-01-29
            Summary|set: the default            |[lwg/2193] set: the default
                   |constructor should be       |constructor should be
                   |explicit                    |explicit
     Ever confirmed|0                           |1

--- Comment #5 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Suspending until 2193 is resolved either way.

[Bug libstdc++/56922] [lwg/2193] set: the default constructor should be explicit

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56922

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|SUSPENDED                   |RESOLVED
         Resolution|---                         |WORKSFORME

--- Comment #6 from Jonathan Wakely <redi at gcc dot gnu.org> ---
http://cplusplus.github.io/LWG/lwg-defects.html#2193