MIME-version: 1.0
Content-type: text/plain; charset=ISO-8859-1; format=flowed
Content-transfer-encoding: 7bit
User-Agent: Mozilla/5.0 (Windows NT 6.0; rv:17.0) Gecko/20130801
            Thunderbird/17.0.8
Message-ID:  <524E7349.4030306@clear.net.nz>
Date:         Fri, 4 Oct 2013 20:50:33 +1300
Reply-To: Mailing list for the LaTeX3 project
              <LATEX-L@LISTSERV.UNI-HEIDELBERG.DE>
Sender: Mailing list for the LaTeX3 project <LATEX-L@LISTSERV.UNI-HEIDELBERG.DE>
From: Andrew Parsloe <aparsloe@CLEAR.NET.NZ>
Subject: Bug in l3fp for sines near 2pi
To: LATEX-L@LISTSERV.UNI-HEIDELBERG.DE
Precedence: list
Envelope-To: <rainer.schoepf@GMX.NET>
Status: R

I recently stumbled on the following bug in the l3fp package:

\fp_eval:n { sin( 2pi - .0001 ) } ==> 9999.999000000002

which seems a little excessive for a function bounded by +1, -1.

\fp_eval:n { sin( 2pi - .001 ) } seems to evaluate correctly. The value 
between .001 and .0001 where the bug arises is between .001853071795, 
which is too small and brings on the error, and .001853071796, which 
doesn't.

\fp_eval:n { sin( 2pi + .0001 ) } seems to evaluate correctly. However, 
the buggy behaviour repeats at 4pi and 6pi and presumably at any 
positive even multiple of pi, but not at 0 (= 0pi). For negative even 
multiples,

\fp_eval:n { sin( -2pi - .0001 ) } seems to evaluate correctly but

\fp_eval:n { sin(-2pi + .0001) } ==> -9999.999000000002

This aside, I greatly appreciate the facilities that the l3fp package 
provides.

Andrew

P.S. Is this a *very* quiet list, or have I not configured things 
correctly so that I'm missing postings to it?