Hi, On 10/01/2014 07:53 PM, tluthe@physik.uni-bielefeld.de wrote:

I have come across a (very rare) error when calling the cl_N(const char *) constructor: For a particular argument with ~10000 digits (see attached file) the constructor seems to insert an extra 0 after the first digit. The input is 3.40805... while the finished cl_N reads 30.40805...

The problem seems to be related to the input length, as adding or removing a digit at the end leads to the correct result.

Thank you very much for being so attentive and reporting this bug! I've committed a patch. It will be picked up with the next release. In the meantime, you're welcome to test it by cloning the git repository. -richy. -- Richard B. Kreckel <http://in.terlu.de/~kreckel/>