Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!snorkelwacker.mit.edu!thunder.mcrcim.mcgill.edu!poly-vlsi!big
From: big@vlsi.polymtl.ca (Patrick Drolet)
Newsgroups: comp.os.msdos.programmer
Subject: Re: scanf bug in TC
Message-ID: <1991Mar8.194255.17091@vlsi.polymtl.ca>
Date: 8 Mar 91 19:42:55 GMT
References: <1991Mar4.190946.26628@novell.com>
Sender: news@vlsi.polymtl.ca (USENET News System)
Organization: Ecole Polytechnique de Montreal
Lines: 41

In article <1991Mar4.190946.26628@novell.com> tporczyk@na.excelan.com (Tony Porczyk) writes:
>The News Manager)
>Nntp-Posting-Host: na
>Reply-To: tporczyk@na.excelan.com (Tony Porczyk)
>Organization: Standard Disclaimer
>References: <1991Mar4.152035.13078@ux1.cso.uiuc.edu>
>Distribution: usa
>Date: Mon, 4 Mar 1991 19:09:46 GMT
>
>In article <1991Mar4.152035.13078@ux1.cso.uiuc.edu> vernon@blueeyes.kines.uiuc.edu (vernon) writes:
>>scanf: floating point formats not linked
>>abnormal program termination
>>I have been informed that this is a bug in TC, but that  there is kludge
>
>Bug? Kludge? Before we get into that - did you #include<proper.header)?
>
>Tony
><as in "Organization" - see above>
>
Ok, a friend of mine had a similar bug using atof().  He fixed it using a
dumb procedure (that he even never called!) so the linker would not forget
to link everything needed on floats...

procedure is the following:

	float linkfloat( float *f1 )
	{
		float f2 = *f1;
		linkfloat(*f2);
		return 0.0;
	}

He told me that he got this information from Borland tech support...

MAKE SURE SUCH PROCEDUREIS IN YOUR CODE BUT DO NOT CALL IT!!!!
Using a float address, it tells the linker to link floating stuff.


					Patrick Drolet
					big@info.polymtl.ca
>