This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug c++/35401] libstdc++ linked to system /usr/lib/libgcc_s.1.dylib not new gcc4.3 libgcc_s.1.dylib
- From: "agn at noc dot soton dot ac dot uk" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 29 Feb 2008 22:30:56 -0000
- Subject: [Bug c++/35401] libstdc++ linked to system /usr/lib/libgcc_s.1.dylib not new gcc4.3 libgcc_s.1.dylib
- References: <bug-35401-15848@http.gcc.gnu.org/bugzilla/>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Comment #16 from agn at noc dot soton dot ac dot uk 2008-02-29 22:30 -------
Many thanks to Jack Howarth for putting in this work to sort out the bug.
Given the comments on the gcc list, I'll reduce the severity of the bug from
blocker to critical.
What originally brought this bug to my notice was that a python extension
linked to gcc43/libstdc++.6.dylib and gcc43/libgcc_s.1.dylib gave a
segmentation fault when called.
I suspected that the use of conflicting c libraries gcc43/libgcc_s.1.dylib and
(implicitly, through gcc43/libstdc++.6.dylib) /usr/lib/libgcc_s.1.dylib might
be causing the segmentation fault.
On the other hand... a similar extension linked to gcc43/libgfortran.3.dylib
and gcc43/libgcc_s.1.dylib was ok, even though gcc43/libgfortran.3.dylib is
itself linked to /usr/lib/libgcc_s.1.dylib
Perhaps the mislinking has more effect on the libstdc++ than on libgfortran?
Apologies if (as is likely) my contribution here is based on ignorance.
George Nurser.
--
agn at noc dot soton dot ac dot uk changed:
What |Removed |Added
----------------------------------------------------------------------------
Severity|blocker |critical
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35401