As stated xenofarm reports 2 errors. IMO this one is a bit harder.
Pike version=7.7.13
The problem fails in running test_resolve.pike. Pike segfaults.
After logging I found the program to fail with one file. If I remove the file with which this occurs (FIPS10_4.pmod) it doesn't segfault anymore.
When using hilfe, the same error occurs:
master()->resolv(Standards.FIPS10_4.pmod);
Segmentation fault
This *only* occurs on my sparc64 (debian) and not on a i386.
I've created an test-program which inherits Standards.FIPS10_4 it also segfaults.
I've attached an strace on this process which I've included as attachment (hope it helps).
I can send more debugging info if necessary, but I have no clue where to go next.
Best regards,
Marc
Sorry to be replying to myself,
It seems somehow it segfaults because this static array "static array(array(string)) divisions"
found in the troublesome file contains to much data which makes the stack overflow (or something like it). When I reduce the amount of data in it, it doesn't segfault anymore.
But normally speaking one would be allowed to have lots of data in an array wouldn't one?
/marc
It seems somehow it segfaults because this static array "static array(array(string)) divisions"
found in the troublesome file contains to much data which makes the stack overflow (or something like it). When I reduce the amount of data in it, it doesn't segfault anymore.
Thanks, I've found where the problem originates.
But normally speaking one would be allowed to have lots of data in an array wouldn't one?
True. Unfortunately fixing the problem means changing the code generator.
On Mon, Feb 28, 2005 at 02:15:00PM +0100, Henrik Grubbström (Lysator) @ Pike (-) developers forum wrote:
True. Unfortunately fixing the problem means changing the code generator.
Does this mean it isn't going to be fixed?
No, it just means that it's nontrivial to fix.
It sounds like something way above my hat. However is there something I can do that could help to solve it?
First, thanks for your effort. But sadly it didn't solve it. Running test_resolve.pike (from xeno testsuite) by hand still ends with a segfault.
"/usr/local/pike/7.7.13/lib/modules/Standards.pmod/IDNA.pmod" Resolving "Standards.IDNA"... "/usr/local/pike/7.7.13/lib/modules/Standards.pmod/FIPS10_4.pmod" Resolving "Standards.FIPS10_4"... Segmentation fault
(Removing Standards.FIPS10_4 still helps resolve the segfault)
Kind regards,
Marc
On Wed, Mar 16, 2005 at 01:20:00PM +0100, Henrik Grubbström (Lysator) @ Pike (-) developers forum wrote:
Potential fix applied to Pike 7.7.
pike-devel@lists.lysator.liu.se
-
Henrik Grubbstr�m (Lysator) ＠ Pike (-) developers forum -
Marc Dirix