On Thu, Nov 25, 2010 at 6:52 AM, <luc.maison...@free.fr> wrote: > On a linux computer in an UTF8 environment and using Java 5, I get the > following messages in English and in French: > > Continued fraction diverged to NaN for value ∞ > Divergence de fraction continue à NaN pour la valeur ∞ > > I don't know if it will be visible through mail, but the last component of > both messages is the single character that represent infinities. It is the > same character you get using the ∞ xml escape sequence, or the ∞ > html escape sequence, or the \u221e java escape sequence. > > So I think the JVM is doing a good job by itself here. Could this be > related to limitations or configuration of your display method (terminal, > emacs, eclipse console ...) ? >
Thanks, Luc! My mistake. I was looking at the test output including the stacktrace using Mac OSX terminal app. I can now see the ∞ is there. In an ideal world, I would rather see "+INF" or "POSITIVE INFINITY" or somesuch (including the sign) and have it plain ascii; but I now see that this is coming from NumberFormat and don't see it as worth the trouble to change. > > Luc > > > ----- "Phil Steitz" <phil.ste...@gmail.com> a écrit : > > > Researching MATH-414, I am seeing this stacktrace: > > > > org.apache.commons.math.ConvergenceException: Continued fraction > > diverged to > > NaN for value ? > > at > > > org.apache.commons.math.util.ContinuedFraction.evaluate(ContinuedFraction.java:186) > > > > The "?" should represent positive infinity (the actual parameter > > passed to > > the exception constructor). Any ideas on how we can modify our > > framework to > > support formatting infinities and NaNs? > > > > Phil > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org > For additional commands, e-mail: dev-h...@commons.apache.org > >
