--- Additional Comments From pinskia at gcc dot gnu dot org 2005-04-27
07:13 ---
The first example is invalid is a dup of bug 19404. Now the second example
looks to be a bug in
libstdc++.
--
What|Removed |Added
-
--- Additional Comments From pcarlini at suse dot de 2005-04-27 08:54
---
The second example seems to me also a duplicate of 19404. This is a reduced
testcase:
class Foo { };
template void operator/(const Foo&, T);
enum { _S_word_bit = 1 };
class vector_bool
{ void _M_allocate() { (_S_w
--- Additional Comments From dominik dot strasser at infineon dot com
2005-04-27 09:07 ---
(In reply to comment #4)
> The second example seems to me also a duplicate of 19404. This is a reduced
> testcase:
>
> class Foo { };
> template void operator/(const Foo&, T);
> enum { _S_word_bit
--- Additional Comments From dominik dot strasser at infineon dot com
2005-04-27 09:09 ---
Created an attachment (id=8750)
--> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8750&action=view)
Use const static member instead of enum
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=212
--
What|Removed |Added
Component|c++ |libstdc++
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21244
--- Additional Comments From dominik dot strasser at infineon dot com
2005-04-27 09:28 ---
(In reply to comment #8)
> On second thought, maybe we can safely change the enum to not be anonymous...
I think so, too as I can't see any user error in the second example.
--
http://gcc.gnu.
--- Additional Comments From pcarlini at suse dot de 2005-04-27 09:37
---
Yes, you are right, but I don't want to fiddle too much with that constant, in
particular risking to change its size (the standard doesn't guarantee that the
underlying type of that anonymous enum is int), seems sa