From mboxrd@z Thu Jan 1 00:00:00 1970 From: Robert Lipe To: law@cygnus.com Cc: "H.J. Lu" , acs@acm.org, egcs@cygnus.com, jason@cygnus.com Subject: Re: 1008 segfaults in genattr Date: Fri, 21 Nov 1997 23:32:00 -0000 Message-id: <19971122013206.43180@dgii.com> References: <19971121234907.48822@dgii.com> <18163.880182906@cygnus.com> X-SW-Source: 1997-11/msg00710.html > I'm kinda curious why your 2nd call to exit is aborting; any chance you > could assembly level debug the 2nd call? Could it just be some paranoid > libc guy aborting if exit is called recursively?!? That's exactly what it's doing. I don't have to disassemble it. I had one of the DS guys at SCO look at the source. I thought I'd mentioned that, but given the way this thread has wandered, I can't blame anyone for not following every detail. :-) > I'd also be interested in whether or not a C++ hello world program will > call __do_global_dtors_aux twice on sco5. No, it does not. (gdb) break __do_global_dtors_aux Breakpoint 2 at 0x8048590 (gdb) run Starting program: /home/play/junk/gcc/a.out Breakpoint 1 at 0x8003ded9 hello Breakpoint 2, 0x8048590 in __do_global_dtors_aux () (gdb) cont Continuing. Program exited normally.