Re: pro*c seg fault

d_at_xevion.net (d_at_xevion.net) wrote in message news:<93c5b0ca.0402200425.3fe5880_at_posting.google.com>...
> Hello,
> Basically I was looking to confirm something, this is the output of a
> backtrace on a core dump. As far as I can tell it was an oracle
> library which called the memcpy (I have none in my code, either before
> or after the precompiler is run).
>
> I was wondering if there was any workaround for this, as I thought
> oracle would just do a sizeof() on the destination (which I don't
> control).
>
> The backtrace looks like,
> (gdb) bt
> #0 0xff3806f4 in memcpy () from
> /usr/platform/SUNW,Sun-Fire-280R/lib/libc_psr.so.1
> #1 0xfe8f70a4 in kpusnchr () from
> /opt/oracle/product/9.2.0.1.0/lib32/libclntsh.so.9.0
> #2 0xfe90dd48 in sqlcas () from
> /opt/oracle/product/9.2.0.1.0/lib32/libclntsh.so.9.0
> #3 0xfe902170 in sqlexp () from
> /opt/oracle/product/9.2.0.1.0/lib32/libclntsh.so.9.0
> #4 0xfe8fd898 in sqlcmex () from
> /opt/oracle/product/9.2.0.1.0/lib32/libclntsh.so.9.0
> #5 0xfe8fdd84 in sqlpex () from
> /opt/oracle/product/9.2.0.1.0/lib32/libclntsh.so.9.0
>
> This is running oracle 9.1.0.2 on solaris (sparc, using 32 bit library
> files).
>
> Thanks

without source, there's little to say.

How are your variables declared? Are they in a DECLARE section? are they really defined? IOW is the pointer initialized? are you passing a pointer instead of an array? If a pointer, did you MALLOC space for it to point to? or are you using a VARCHAR? Did you set the size of the VARCHAR correctly?

There are so may ways for you to pass a bad value to ORACLE.

Try to reduce the error to a small example. Post the source code, related tables definitions and maybe even a couple rows of data. I wouldn't blame ORACLE until I'd eliminated all cases of error on my side.

Ed Received on Fri Feb 20 2004 - 12:16:14 CST