On Tue, Jul 16, 2002 at 04:51:19PM -0500, Philip Molter wrote:
> On Tue, Jul 16, 2002 at 04:40:05PM -0500, Philip Molter wrote:
> : Well, when I run the code with trace(9), it causes the script (it's
> : a web script running under mod_perl) to fail with not even a warning
> : to the error log, so I doubt this will be helpful, but here's the
> : output anyway.
> :
> : [REPLACED 1.28 WITH 1.29]
> : Tx::DBI::db=HASH(0xfd26a4) trace level set to 9 in DBI 1.29-nothread
> : Note: perl is running without the recommended perl -w option
> : -> selectall_arrayref for DBD::mysql::db (Tx::DBI::db=HASH(0x4d3b28)~0xfd26a
> : 4 'SELECT customer FROM customer WHERE customer IN ( ? )' undef '1')
> :
> : [CLEAN 1.29 BUILD]
> : Tx::DBI::db=HASH(0xfd2914) trace level set to 9 in DBI 1.29-nothread
> : Note: perl is running without the recommended perl -w option
> : -> selectall_arrayref in DBD::_::db for DBD::mysql::db (Tx::DBI::db=HASH(0xa
> : 8f524)~0xfd2914 'SELECT customer FROM customer WHERE customer IN ( ? )' undef '1
> : ')
>
> And I promise this is the last reply to myself. This gdb output
> may not be too helpful (no line numbers for one), but at least we
> can see that there's a problem somewhere in XS_DBI_dispatch. This
> would be the core dump produced by the clean 1.29 build.
>
> I will try to prune this down to a much more manageable test case.
Getting the line numbers on the stack trace would be a big help.
Build the DBI using
perl Makefile.PL -g
Tim.
