Actually, I already do use unique names for statement handles. No affect
on the crashes.
Since this problem is so transient, and so dependant on what code is being
executed (and presumably where the code happens to get loaded into memory),
I'd guess that chaning the variable names just made the symptoms go away,
but didn't actually get to the root of the problem. I've have cases where
putting in an extra comment or taking a comment out will affect the
crashes.
Simon Oliver
<simon.oliver@um To: Paul Van Deursen
<[EMAIL PROTECTED]>
ist.ac.uk> cc: [EMAIL PROTECTED],
[EMAIL PROTECTED]
Subject: Re: DBD::ODBC crashes on exit
05/16/02 11:37
AM
Paul Van Deursen wrote:
>
> I had the same problem on a NT 4 machine.
> I say had because all my problems disappeared after I made all the
> statement handle names unique.
>
> I used to create $sth and call finish afterwards and create a new $sth
> which gave me that famous error.
> After I renamed them to $sth1, $sth2 etc. (in the same subroutine) this
problem disappeared.
> It seems that even after a finish() call the statement handles stay
active
> but I don't have a clue why.
There should be no need to call $sth->finish unless you didn't retrieve
all records from handle. In any case, to destroy a statement handle use
`undef $sth` - there should be no need for new variable names.
Ryan, let us know if it fixes the problem.
--
Simon Oliver
