penneyj@servio.UUCP (D. Jason Penney) (09/01/89)
Has anyone else seen this problem? I have a program with a SIGINT trap installed using sigvec(). This program works fine on six other systems, including 5 other Unix environments. On my DECstation-3100 running Ultrix 3.x, this SIGINT handler works fine UNLESS I invoke it while the program is generating output. In this case, I get the message, longjmp botch Then, if I'm lucky, I get a core dump with a SIGILL (SEXC_CPU, coprocessor unusable) code. The call stack from the resulting core looks like, [stack frames deleted] vvvvv This is our general purpose fault handler vvvv 4 HostFaultHandler(sig = 4, code = 76, scp = 0x7fffe914) ["hostunix.c":1815, 0x410d78] 5 abort.abort(0x0, 0x0, 0x436b98, 0x0, 0x424ee0) ["setjmperr.c":26, 0x43db2c] 6 _filbuf(0x0, 0x0, 0x0, 0x0, 0x0) ["setjmp.s":113, 0x433c1c] 7 main(argc = 0, argv = (nil)) ["/beta/product/programs/gemstone55/fred.hc":2981, 0x424edc] As you can see, the error seems to be generated from within _filbuf, a C i/o routine. Any clues, guys? Thanks in advance... -- D. Jason Penney Ph: (503) 629-8383 Beaverton, OR 97006 uucp: ...uunet!servio!penneyj STANDARD DISCLAIMER: Should I or my opinions be caught or killed, the company will disavow any knowledge of my actions...