RE: strange behavior in 9.40.FC6

There were 2 defects fixed in the patch. Case #411757 opened with IBM Informix regarding a problem with stored procedure guphrase(pv_assn_num LIST(INTEGER NOT NULL). When executed the procedure is doing a sequential scan of the foassn table as well as scanning the entire index it should be using. Problem is with the use of the LIST data type. If a CHAR type is added to the arg list and included in the query where the LIST is used, the query does an indexed read of foassn. update 09/14/2004 - This problem is a software defect in IDS and has been referred to the IDS development group. The defect number is 169632. Problem with IDS 9.40.FC6 and processing LIST variables in a stored procedure. We are getting error "9602: Illegal attempt to convert a collection type into another type." when the procedure is executed. Case #425892 opened with IBM Informix. This problem is documented as defect #171816 and we are waiting for a patch to 9.40.FC6. The patch will include a fix for this defect as well as defect #169632 (see above). 171816 did have a work-around, using type casting when calling the procedure. The strange behavior problem is corrected by not setting PDQPRIORITY when updating stats on procedures. Bill > -----Original Message----- > From: owner-informix-list@iiug.org [SMTP:owner-informix-list@iiug.org] > On Behalf Of Neil Truby > Sent: Wednesday, May 04, 2005 3:54 AM > To: informix-list@iiug.org > Subject: Re: strange behavior in 9.40.FC6 > > > "Bill Dare" <dareb@jevic.com> wrote in message > news:1115154586.41877d638e6a36c3447a085ec262c671@teranews... > > > > I've actually got a patch loaded, 9.40.FC6X2. The server is > randomly > > allocating multiple scan and sometimes group threads to sessions for > no > > apparent reason. Onstat -g mgm doesn't report any PDQ queries > running. > > The session has PDQPRIORITY set to 0. Onstat -u doesn't show the > scan > > threads doing much work, if any. Usually just a few reads. > > > > Just wondering if anyone has seen this behavior in 9.40.FC6. I > think > > the patch is the likely cause and I plan to back it out this weekend > to > > verify that. > > > > Bill > > I have no insight to your problem but am interested to know what bug > in FC5 > or earlier you have the patch for. Could it be this one?: > 171489 SERVER CRASHES IN CMPRSTR/CMPRNAME IF DICTIONARY CACHE IS TOO > SMALL. > > I'd like to know for the benefit of one of my clients who is about to > apply > an FC6 patched version too. > > thanks > Neil > > sending to informix-list