[Bug c++/12491] [3.3/3.4 Regression] [eh] Destructor fails to compile when optimizations (inlining) are enabled

["wilson at specifixinc dot com" <gcc-bugzilla at gcc dot gnu dot org>]

------- Additional Comments From wilson at specifixinc dot com  2003-12-06 01:09 -------
Subject: Re:  [3.3/3.4 Regression] [eh] Destructor fails to
 compile when optimizations (inlining) are enabled

ebotcazou at gcc dot gnu dot org wrote:
> ------- Additional Comments From ebotcazou at gcc dot gnu dot org  2003-12-05 14:37 -------
> Can a C++ specialist tell if it is normal for the RTL code of Error::~Error() to
> be emitted twice (see .rtl at -O0 or -O -fno-inline)? If not, I think this is
> probably the source of the problem.

This came up on the gcc list recently.  Gaby answered that worst case we 
  need 3 of them, and it is common to need 2 of them.  Unfortunately, I 
can't find his answer at the moment.  If you unmangle the mangled names, 
you will see one described as "in-charge" and one described as 
"not-in-charge", though I don't know what that means.  Sometimes they 
are identical, but gcc doesn't know how to optimize out the redundant one.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=12491