Sure, g++ 3.2 all over the place. And with -O0 -fno-inline everything works fine. ---- Lev Assinovsky Aelita Software Corporation (now is a part of Quest Software) O&S InTrust Framework Division, Team Leader ICQ# 165072909 > -----Original Message----- > From: lrtaylor@xxxxxxxxxx [mailto:lrtaylor@xxxxxxxxxx] > Sent: Tuesday, May 18, 2004 8:16 PM > To: Lev Assinovsky; gcc-help@xxxxxxxxxxx > Subject: RE: Optimized exception problem! > > > This looks kind of like you're throwing an exception that isn't being > caught for some reason. We've had problems like this, and > have usually > tracked it down to problems throwing exceptions across library > boundaries and/or not linking using g++. Are you linking everything > C++ish using g++? All C++ libraries and programs (or libraries and > program containing any C++) should be linked using g++. This can also > happen if you simply don't have a catch block that catches the type of > exception being thrown. That will cause your program to > abort as well. > > Cheers, > Lyle > > -----Original Message----- > From: gcc-help-owner@xxxxxxxxxxx > [mailto:gcc-help-owner@xxxxxxxxxxx] On > Behalf Of Lev Assinovsky > Sent: Tuesday, May 18, 2004 7:00 AM > To: Gcc-Help (E-mail) > Subject: Optimized exception problem! > > Hi folks! > I have a big application and sometimes (not every build) > I get the following crash in the dynamically loaded shared object): > #0 0xdd8cefb7 in _lwp_kill () from /usr/lib/libc.so.1 > #1 0xdd80a3ad in thr_kill () from /usr/lib/libthread.so.1 > #2 0xdd8e50fd in raise () from /usr/lib/libc.so.1 > #3 0xdd8cff5e in abort () from /usr/lib/libc.so.1 > #4 0xdd9ec4bc in _ZN10__cxxabiv111__terminateEPFvvE () from > /usr/ADCAgent/libstdc++.so.5 > #5 0xdd9ec3a3 in __gxx_personality_v0 () from > /usr/ADCAgent/libstdc++.so.5 > #6 0xdd963d21 in _Unwind_GetTextRelBase () from > /usr/ADCAgent/libgcc_s.so.1 > #7 0xdd963e52 in _Unwind_RaiseException () from > /usr/ADCAgent/libgcc_s.so.1 > #8 0xdd9ec640 in __cxa_throw () from /usr/ADCAgent/libstdc++.so.5 > #9 0xdb526552 in > ... > ... > > It happens with gcc 3.2 in Solaris 8 Intel only (not in > sparc) and with > optimization (-O2, -O3). > Unfortunately I am unable to reproduce this bug in a small test case > that's why > I can't show the source code here. > Could anybody suggest the way of how to catch the bug? > Or at least tell me please what the purpose of __gxx_personality_v0 > function is? > I didn't find it in gcc source. > Any help will be greatly appreciated! > > ---- > Lev Assinovsky > Quest Software > O&S InTrust Framework Division, Team Leader > >
