Linking problem in 7.10G

[Deskin, Bob]

I must have missed the message - happens sometimes. As for upward
compatibility, we considered what it was doing as incorrect (I grant you
that you may have a different opinion). When we see the product as being in
error, we usually (but not always) correct the problem as best we can. Our
feeling is that if a product error was being undetected, most customers
would prefer to know about a problem. Unfortunately that sometimes means
that things don't work the same way. But we can't have it both ways so we
make a design decision at the time.

Bob Deskin              
Senior Product Advisor  bob.deskin@cognos.com
Cognos Inc.             (613) 738-1338 ext 4205 FAX: (613) 228-3149
3755 Riverside Drive P.O. Box 9707 Stn. T, Ottawa ON K1G 4K9 CANADA


> ----------
> From: 	alex ling[SMTP:linga88@yahoo.com]
> Sent: 	Thursday, September 03, 1998 11:06 PM
> To: 	Bob.Deskin@Cognos.COM
> Cc: 	powerh-l@lists.swau.edu
> Subject: 	Re: Linking problem in 7.10G
> 
> 
> ---"Deskin, Bob" <Bob.Deskin@Cognos.COM> wrote:
> >
> > I just heard from Allison about the linking problem in 7.10G and I
> thought
> > everyone would be interested. The bug # is 245681 and has to do with
> getting
> > a data conversion error linking an InterBase long binary to an
> InterBase
> > short binary. Turns out that the error existed at least as far back as
> > 7.10E4 (and probably further back) but it was ignored rather than
> being
> > reported. During the development of the QTP Tracer (which is
> Allison's baby
> > and accomplishment - everyone say thank you to Allison), the code was
> > changed to report the error so it could be traced. Now the clincher.
> The
> > error only occurs when a value that is too large for the short
> binary is
> > used. In other words, this is a valid error message. This bug is being
> > rejected.
> > 
> 
> Yeah, I reported this problem to C/S sydney on 14 Aug and was informed
> that Ottawa (two, YES two staff) confirmed that it was a suspected bug
> since it was working perfectly in 710E4 (Bob, 710E4 is our current
> production version and it is not a problem on this version). Despite
> my cry for help and finally having to resort to hotsite status
> pursuit, no detailed investigation into this problem was being
> conducted by Cognos and no sound technical explaination was given on
> the nature of the problem that would help us and give us the
> confidence. I reached my own finding yesterday which was then pasted
> on to you last night.
> 
> Bug 245681 - Linking a LONG to a SHORT causing Data Conversion Error.
> The problem will occur when -
> 
> linking X (LONG) to Y (SHORT).
> where X is greater than 32767 (ie, 2^15 -1) since SHORT  in Interbase
> is 2-byte signed in PH.
> 
> The error did not occur in 710E4, 710F1.
> It surfaced in 710F2, 710F4, 710G (710F3 not tested).
>  
> I should not have to do all these work on my own in the dark. You guys
> have the C-code behind PH.  
> 
> Rejected? What about the issue of upward compatibility for released
> software???
> 
> Bob, answer to your first reply on this subject. It was the ordeal of
> being kept in the dark working on the problem on your own, along with
> the list of known bugs and who know what else that prevented our
> management from giving the green light to upgarde to 710G. Also
> because of OCC deadline for us to be Y2K compliant by end Oct 1998, we
> need Plan B which requires a known working version for us. We cannot
> rely on that 710G1 will be a working version for us after having
> relied on 710G. We are currently on 710E4.
> 
> Suggestion - C/S Asia Pacific needs more than just 1.5 staff!!
> 
> Regards,
> A Ling.
> 
> 
> _________________________________________________________
> DO YOU YAHOO!?
> Get your free @yahoo.com address at http://mail.yahoo.com
> 
= = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = =
Subscribe: "subscribe powerh-l" in message body to majordomo@lists.swau.edu
Unsubscribe: "unsubscribe powerh-l" in message to majordomo@lists.swau.edu
powerh-l@lists.swau.edu is gatewayed one-way to bit.listserv.powerh-l
This list is closed, thus to post to the list, you must be a subscriber.