ffc team mailing list archive

[Anders Logg <logg@xxxxxxxxx>]

Kristian, could you fix so that the values in the table of
coefficients in evaluate_basis are rounded to zero when smaller in
absolute value than FFC_EPSILON?

The coefficients will otherwise look different on different machines
and make it difficult to run regression tests.

(Your machine and mine seem to agree, but I know of at least one other
machine that doesn't...)

/Anders