[Mono-bugs] [Bug 45043][Wis] Changed - unwinding segfault

bugzilla-daemon@bugzilla.ximian.com bugzilla-daemon@bugzilla.ximian.com
Thu, 11 Sep 2003 19:01:24 -0400 (EDT)

Please do not reply to this email- if you want to comment on the bug, go to the
URL shown below and enter your comments there.

Changed by rwahl@gmx.de.


--- shadow/45043	2003-09-11 18:06:56.000000000 -0400
+++ shadow/45043.tmp.24839	2003-09-11 19:01:24.000000000 -0400
@@ -93,6 +93,25 @@
 Oops, sorry - the patch will not work. My test setup was modified. The
 reason might be that fallback_frame_state_for is a symbol local to
 glibc and cannot referenced externally. I try to find a solution for
 - ron
+------- Additional Comments From rwahl@gmx.de  2003-09-11 19:01 -------
+Ok, I think it will not be as easy as I thought. I even fear that the
+fallback implementation of glibc might be removed from it in the
+future - but since its not even exported it will not be of much help.
+I think the best solution (regards to the current and future versions
+of glibc and gcc) would be to find a way to use __frame_state_for of
+libgcc. To do this mono somehow has to register the .eh_frame
+information via __register_frame_info_bases() and then this should
+work. My knowlegde in this area is very limited so I hope there is one
+who can try fixing this issue.
+BTW, mono seems to work if compiled under RedHat. The reason is that
+RedHat has a slightly hacked compiler that seems to add some
+compatiblity code. But these patches are not official I think and
+cannot be relied on. So I would prefer a "real" solution here.
+To reproduce the problem you may even use the reflection.cs from the
+tests directory. It crashes with an unhandled exception.