[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#164135: gcc bug?: sin(a) * sin(b) gives wrong result



tags 164135 + fixed
retitle 164135 [fixed in gcc-3.2] sin(a) * sin(b) gives wrong result 
thanks

Khalid Shukri writes:
> Package: gcc-3.0
> Version: 1:3.0.4-7
> Kernel: self-compiled 2.4.18
> libc6: 2.2.5-11.2
> libc6-dev: 2.2.5-11.2
> System: woody on a Pentium 4 with 2 GHz and 1 GB RAM
> It looks to me as if gcc can't parse the following line correctly:
> double c=sin(a) * sin(b);
> Excuse me if this is just a stupid programming error of my part, I'm not
> an experienced C programmer.
> I compiled the following little programm:
> 
> #include <stdlib.h>
> #include <tgmath.h>
> 
> int main () 
> {
> 	const double a=-0.591084;
> 	const double b=0.618719;
>         double res=sin(a) * sin(b);
>         printf ("one step: %f\n", res);
> 	res=(sin(a)) * (sin(b));
> 	printf ("one step: %f\n", res);
> 	double sina=sin(a);
> 	double sinb=sin(b);
> 	res=sina * sinb;
> 	printf ("two steps: %f\n", res);
> 	exit(0);
> } 
> 
> with gcc-3.0 -lm -o parsbug parsbug
> and it gives me the following output:
> 
> one step: 0.336391
> one step: 0.336391
> two steps: -0.323207
> 
> Only the last line is correct.
> 
> By the way, gcc-2.95 always gives me a lot of parse errors and doesn't compile this at all.
> 
> Cheers
> Khalid
> -- 
> ###############################################################################
> #                                                                       
> # 
> # La traditionelle lucidité des dépressifs, souvent décrite comme un    
> #
> # désinvestissement radical à l'égard des préoccupations humaines, se   
> #
> # manifeste en tout premier lieu par un manque d'intérêt pour les
> questions   #
> # effectivement peu intéressantes. Ainsi peut-on, à la rigueur, imaginer
> un   #
> # dépressif amoureux, tandis qu'un dépressif patriote paraît franchement
> #
> # inconcevable.                                                         
> #
> # (M. Houellebecq: Les particules élémentaires)                         
> #
> #                                                                       
> #
> ###############################################################################
> 
> 
> 
> -- 
> To UNSUBSCRIBE, email to debian-gcc-request@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org



Reply to: