LOOKUP to self segV's on 64 bit server.

[Jerry M. Rains]

On Monday 21 March 2005 11:32, John Esak wrote:
> (I found this stuck in my Outbox... don't know if it made here once
> already... apologies if it did.)
>
> > -----Original Message-----
> > From: filepro-list-bounces at lists.celestial.com
> > [mailto:filepro-list-bounces at lists.celestial.com]On Behalf Of Jeff
> > Harrison
> > Sent: Friday, March 18, 2005 12:08 PM
> > To: filepro-list at seaslug.org
> > Subject: RE: LOOKUP to self segV's on 64 bit server.
> >
> > --- John Esak <john at valar.com> wrote:
> > > Jerry,
> > > I'm going to just assume that you are doing this
> > > lookup from an @entsel
> > > processing stub??  Otherwise, it is quite obvious
> > > that you are creating a
> > > deadlock situation. In other words looking up to the
> > > record you are locking
> > > yourself is _never_ going to happen. Regardless, it
> > > probably shouldn't
> > > segv... but assuming you are not in @entsel, it
> > > should just sit and wait...
> > > forever...
> > >
> > > So, I'm guessing you left some info out, or this
> > > type of lookup has _nver_
> > > worked anywhere before... one or the other or both.
> > >
> > > John Esak
> >
> > Well, a protected lookup would not work, but the type
> > of lookup Jerry described was not protected.  You CAN
> > do read only lookups (which is of course the default)
> > to locked records.
> >
> > Jeff Harrison
>
> You're right, of course. But that is what I meant by "leaving _something_
> out". If there is no -p on the lookup line, which I should have checked
> more clearly... there is nothing problematic with looking up to a record
> you already have locked... but to what purpose then?  You already have the
> fields at your beck and call. They are not going to change becuase you
> lookup to them.  In any case, I complicated the matter unnecessarily.
> Sorry.

Jeff is right.  What this processing does is allow me to keep an audit trail
of any field that changes during an update.  Do a non-protected lookup.
Throw an array on both the lookup and the main file and loop through all
fields looking for changes.

Anyway, I figured out that the problem was 5.0.14D4.  I tried a 5.0.13D4
dclerk and the problem is gone.  So I put copied a fp_v5.0.13D4 directory to
the machine until Ken can fix .14.

Jerry

> JE
>
> _______________________________________________
> Filepro-list mailing list
> Filepro-list at lists.celestial.com
> http://mailman.celestial.com/mailman/listinfo/filepro-list

-------------------------------------------------------