|
From: | Mark M. Ito |
Subject: | Re: [Bug-gsl] non-linear LS fit example in documentation: bug? |
Date: | Thu, 22 Jan 2009 16:07:34 -0500 |
User-agent: | Thunderbird 2.0.0.18 (X11/20081119) |
Brian,OK. That makes sense. Thanks for looking into that. Looks like using the "break" will save a few iterations.
I guess that having the solver return a "continue" and having that cause the program to exit the iteration loop was making me nervous.
-- Mark Brian Gough wrote:
Thanks for the example. In this case GSL_CONTINUE is an error code which indicates that the iterate function could not make any reasonable progress, although it didn't encounter any fatal errors (therefore calling the function again is possible). With the example program this occurs when it's essentially at the minimum. In hindsight, GSL_ENOPROG would have been a better choice--I'll log that as a bug and put a note in the manual.
[Prev in Thread] | Current Thread | [Next in Thread] |