On Wed, 4 Aug 1999, Vadim Mikheev wrote:
> Anto Prijosoesilo wrote:
> >
> > #0 0x16f62 in _bt_isortcmpinit ()
> > #0 0x16f62 in _bt_isortcmpinit ()
> > #1 0x13c3c in btgettuple ()
>
> _bt_isortcmpinit couldn't be called from btgettuple!!!
> This's static func in nbtsort.c used for index creation,
> not for querying.
>
> > Has anyone seen this problem? I'd like to get a more detailed debugging
> > info if someone can show me how to compile it with debugging info.
>
> I use Makefile.custom (in .../pgsql/src) with CUSTOM_COPT= -ggdb3.
Sorry for the duplicate posting of the original question. My first posting
bounced because I wasn't subscribed to the list. I'm now subscribed to the
admin-digest list so I don't see the posting much more often than once a
day and I didn't realize that the first posting actually made it to the
list until just a few minutes ago.
I've enabled debugging with -g3 in Makefile.custom and recompiled. Size(1)
returned the same info for the original compile and the debug compile but
unfortunately I've deleted the old cores before checking it against the
debug compile.
The compile flags now include "-O -g3". The "-O" was there in the original
compile. Would that have caused the backtrace to show the wrong functions?
I tried a lengthy script that crashed the backend yesterday but for
some reason it didn't crash it now. A large query that crashed the backend
for some reason didn't give me a core.
I'll have to wait until the load increases, probably this afternoon or
tonight when most of the users log in.
Thanks,
Anto.
