Re: Tip of current tree: Seg fault in query

"Kevin O'Gorman" <kogorman@pacbell.net> writes:
> Program received signal SIGSEGV, Segmentation fault.
> attnameAttNum (rd=0x10000, a=0x82172a0 "product_level") at parse_relation.c:967
> 967        for (i = 0; i < rd->rd_rel->relnatts; i++)
> (gdb) bt
> #0  attnameAttNum (rd=0x10000, a=0x82172a0 "product_level") at parse_relation.c:967
> #1  0x809a818 in checkInsertTargets (pstate=0x8218350, cols=0x82172d0, attrnos=0xbffff058)
>     at parse_target.c:374

It looks to me like pstate->p_target_relation contains garbage (it seems
unlikely that 0x10000 is a valid pointer).  I changed the contents of
struct ParseState recently, so I'm wondering if you haven't just got
a problem with inconsistent object files.  Does a make clean and rebuild
fix it?
        regards, tom lane