This discussion is archived
6 Replies Latest reply: May 9, 2012 9:00 AM by 936167 RSS

pro*c program causing core dump in 11g

708410 Newbie
Currently Being Moderated
hello every one,
I am trying to debug a pro*c program which is resulting in a core dump. It used to work fine in with Oracle 10g precompiler but is causing a core dump with 11g. When I run dbx here is what I get.

dbx wreg309
Type 'help' for help.
[using memory image in core]
reading symbolic information ...
Segmentation fault in u_fsetcodepage_3_8 at 0x9000000014f4f70 ($t1)
0x9000000014f4f70 (u_fsetcodepage_3_8+0x68) f87f0010 std r3,0x10(r31)


Any ideas what this means ? Thanks.
  • 1. Re: pro*c program causing core dump in 11g
    Mark Williams-Oracle Employee ACE
    Currently Being Moderated
    Hi,

    Doing some searching suggests that you might be using the "International Components for Unicode" (http://site.icu-project.org/home) either directly or indirectly with an IBM product.

    The particular function where the segmentation fault is occurring looks to be setting the codepage:

    http://icu-project.org/apiref/icu4c/ustdio_8h.html#7381b833b6835034008985da1fb9acc2

    A couple of search hits indicate that a #define fsetcodepage u_fsetcodepage_3_8 might be in a header file used on your system.

    Regards,

    Mark
  • 2. Re: pro*c program causing core dump in 11g
    708410 Newbie
    Currently Being Moderated
    Thanks! You are dead on. We are using Banner 8 libraries which uses UFT 8 character set and does use ustdio.h. Now I am trying to figure out how to fix this.

    Thanks.
  • 3. Re: pro*c program causing core dump in 11g
    Mark Williams-Oracle Employee ACE
    Currently Being Moderated
    Now I am trying to figure out how to fix this.

    Hmm, well that will be the fun part. :) If Pro*C was upgraded, the Oracle Client must have been too, perhaps other things as well... Does the segmentation fault always reproduce? Do compiler settings (i.e. optimization) have any impact?

    - Mark
  • 4. Re: pro*c program causing core dump in 11g
    708410 Newbie
    Currently Being Moderated
    Segmentation fault happens as soon as program is run and happens every time. One thing is I have noticed is that it only happens in programs using curses.h. I am wondering if curses is not compatible with ustdio.
  • 5. Re: pro*c program causing core dump in 11g
    712047 Newbie
    Currently Being Moderated
    We are having simular problems trying to run programs that use curses.h

    Has any solution been found yet?

    Thanks for any help.

    --danny                                                                                                                                                                                                                                                                                               
  • 6. Re: pro*c program causing core dump in 11g
    936167 Newbie
    Currently Being Moderated
    Hi,
    I came across your problem on the Oracle Discussions Forum from back on June of 2009.

    I am working with a pro*c program, in Banner 8. I getting the same message from a core dump,, that you got. I was hoping
    you might have written down what you did to resolve it.

    My pro*c program is key to running all the SQR code we have. So it's very important. The version of sqr that gets
    linked into it is 32-bit and our environment is 64-bit. Our contract with Oracle for SQR has lapsed (it's a long
    and expensive story and this is probably not the place). My whole migration to Oracle 11 is being held up by this.

    I realize it's been a while since you worked on it but if you could tell me how you resolved your problem, I might be
    able to do the same.

    Thanks,
    Tom Mayne
    North Shore Community College
    Danvers MA
    tmayne@northshore.edu
    cell 978 423 6867

Legend

  • Correct Answers - 10 points
  • Helpful Answers - 5 points