On Tue, Jun 29, 2004 at 08:05:26AM +0200, Michael Peppler wrote:
> On Tue, 2004-06-29 at 00:31, Tim Bunce wrote:
> > A release candidate for DBI 1.43 is available for testing at:
> >
> > http://homepage.eircom.net/~timbunce/DBI-1.43-rc1-20040628.tar.gz
> >
> > =head1 CHANGES in DBI 1.43 (svn rev 373), 28th June 2004
> >
>
> > Fixed last_insert_id(...) usage check thanks to Rudy Lippan.
>
> Hmmm.
>
> I've got last_insert_id() in DBD::Sybase::db, but it doesn't get called
> from the dispatcher.
>
> DBI 1.43-ithread default trace level set to 0x0/10 (pid 2880)
> Note: perl is running without the recommended perl -w option
> >> last_insert_id DISPATCH (DBI::db=HASH(0xa0c926c) rc1/1 @5 g0
> ima2101 pid#2880) at t/main.t line 166
> not ok 17
>
> The code:
>
> if($DBI::VERSION > 1.42) {
> # This will only work w/ DBI >= 1.43
> $dbh->do("create table #idtest(id numeric(9,0) identity, c
> varchar(20))");
> $dbh->do("insert #idtest (c) values ('123456')");
> DBI->trace(10);
> my $value = $dbh->last_insert_id(undef,undef,undef,undef);
> if($value > 0) {
> print "ok 17\n";
> } else {
> print "not ok 17\n";
> }
> DBI->trace(0);
> }
>
> This worked when I applied Rudy's patch (in particular the O=>0x2820 vs.
> O=>0x2100) to 1.42.
Ah. It's using
IMA_SHOW_ERR_STMT|IMA_STUB (0x2100)
instead of
IMA_SHOW_ERR_STMT|IMA_NOT_FOUND_OKAY (0x2800)
I probably applied Rudy's change by hand and missed that part.
Fixed now. Thanks.
Tim.
p.s. I'm not keen in adding 0x0020 (IMA_NO_TAINT_OUT). The only methods that
use it currently are quote() and quote_indentifier(). If TaintOut is
enabled (implying that you don't trust the database/server) then I
think it's reasonable for last_insert_id to be tainted.
