home *** CD-ROM | disk | FTP | other *** search
- Path: news.ioffe.rssi.ru!usenet
- From: "Eugene I Levin" <Levin@lh.ioffe.rssi.ru>
- Newsgroups: comp.lang.c++
- Subject: Re: Seriuos bug in BC4.52 !!!
- Date: Mon, 5 Feb 1996 17:22:13 +0300 (MSK)
- Organization: Ioffe Institute
- Distribution: world
- Message-ID: <ABLAX5nO2B@lh.ioffe.rssi.ru>
- NNTP-Posting-Host: relay.ioffe.rssi.ru
- X-Mailer: dMail [Demos Mail for DOS v1.23]
-
- CLVN11A@prodigy.com (Richard Croteau) wrote:
-
- > I just installed Borland C++ V4.52 on my 75MHz
- > pentium running Windows 95 (with MS Plus). The problem that
- > I have is that, in any simple programs involving floating
- > point numbers, I get a "floating point: Overflow" run time
- > error whenever I enter a number with a decimal point. This
- > error seems to be invalid. Running the same program on the
- > university computer does not result in this error and the
- > programs execute properly. Could some sort of installation
- > error have occurred?
-
- I think it might be interesting for you to read the following
- messages from comp.databases.paradox:
-
- Jamie Brown <Jamie_Brown%Legis@state.ak.us> wrote:
-
- > Try this:
- > Create a table with a single numeric field.
- > Enter 2 records with the values -25.2 and 132.2
- > Run a calc sum query.
- >
- > The correct answer of course is 107. I get 1061 on two different
- > machines both running Win95 (486 & pentium). DOS PDox 4.5 and PWF 5.0
- > under Win3.1 work correctly.
- >
- > If I add a third record with the value 1 the query answer is 1071!?
- >
- > I called Borland developer tech help. They couldn't reproduce the
- > problem. Can anyone else? Or am I going crazy?
-
- Then a lot of people reproduced the bug. Then there was a very hot
- discussion when different folks blamed Borland, Pentium processor,
- the Lord above etc. etc. Everything ended up with the following
- solution:
-
- >From: Jamie Brown <Jamie_Brown@legis.state.ak.us>
- >Subject: [NEWS] Floating Point Error
- >Date: Thu, 28 Dec 1995 13:20:03 -0900
- >
- >The answer to this problem seems to be that the SYSTEM AGENT in Win95
- >PLUS! causes floating point math errors. I removed it and the calc sum
- >worked correctly.
-
- >From: Jamie Brown <Jamie_Brown@legis.state.ak.us>
- >Subject: [NEWS] Re: Win95 & PFW5.0, 2+2=5 ???
- >Date: Tue, 02 Jan 1996 18:05:51 -0900
- >
- >Apparently in Win95 there is a known floating point math error that has
- >to do with using the System Agent. I removed it and my error vanished.
- >There is also supposed to be a patch available from MS BBS.
-
- Hope this helps,
- --
- Eugene I Levin, St.Petersburg, Russia (Levin@lh.ioffe.rssi.ru).
-
-