This is the mail archive of the cygwin@sources.redhat.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Net version 1.1.0: gcc 2.95.2-2 compiler bug


On Mon, Jul 10, 2000 at 01:42:35PM -1000, Jim Heasley wrote:
>In porting some astronomical data handling software from linux/unix to
>Cygwin (the latest nework release, 1.1.0 I believe), I discovered an
>error in how gcc 2.95.2-2 generates output of a floating or double
>variable set to 0.0 when formatted with a %E format.  I didn't get any
>hits on this problem in the FAQ or mail lists.  The following little
>test program illustrates the problem:

To investigate this problem, please check out the 'newlib' part of the
cygwin package.  That's where things like printf are implemented.

cgf

>main()
>{
>	float value = 0.0;
>	double dvalue = 0.0;
>	printf("%11.0f %15.6f %13.5E\n",value,value,value);
>	printf("%11.0f %15.6f %21.14E\n",dvalue,dvalue,dvalue);
>	dvalue = value = 3.0;
>	printf("%11.0f %15.6f %13.5E\n",value,value,value);
>	printf("%11.0f %15.6f %21.14E\n",dvalue,dvalue,dvalue);
>}
>	        
>Under linux/unix, this little program produces the following
>output:
>
>          0        0.000000   0.00000E+00
>          0        0.000000  0.00000000000000E+00
>          3        3.000000   3.00000E+00
>          3        3.000000  3.00000000000000E+00
>
>But the following is produced by the Cygwin gcc 2.95.2-2:
>
>	  0        0.000000   0.
>          0        0.000000  0.
>          3        3.000000   3.00000E+00
>          3        3.000000  3.00000000000000E+00
>
>When the output is sent to the terminal screen, the results for 0.0 with
>the E format come out as 0.  followed by some "strange" characters and
>then E+00.  Saving stdout to a file and then moving from Cygwin to linux,
>I find there are a bunch of control characters after the leading 0.  and
>preceeding the E+00. 
>
>Sorry, I don't know enough about compilers to even take a wack a solving
>this problem, but would very much appreciate a patch for the sources
>or a working binary with the patch incorporated if someone could kindly
>generate one. 
>
>Thanks,
>
>Jim Heasley
>Professor of Astronomy
>
>Institute for Astronomy			heasley@hoku.ifa.hawaii.edu
>University of Hawaii			phone: 808-956-6826
>2680 Woodlawn Drive			fax:   808-956-9580
>Honolulu, HI 96822
>
>--
>Want to unsubscribe from this list?
>Send a message to cygwin-unsubscribe@sourceware.cygnus.com

-- 
cgf@cygnus.com                        Cygnus Solutions, a Red Hat company
http://sourceware.cygnus.com/         http://www.redhat.com/

--
Want to unsubscribe from this list?
Send a message to cygwin-unsubscribe@sourceware.cygnus.com


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]