apache / commons-numbers / 782

Builds Branch Commit Type Ran Committer Via Coverage
782 master NUMBERS-185: Allow Precision.compareTo using a max ULP to be used for sorting. Updated tests to ensure compareTo is symmetric for the arguments. Updated the documentation to state expected behaviour with NaN arguments. push aherbert travis-ci-com pending completion   set done
781 master Replace deprecated calls. push Gilles Sadowski travis-ci-com pending completion   set done
780 master NUMBERS-181: Track changes push aherbert travis-ci-com pending completion   set done
779 master NUMBERS-183: Track changes push aherbert travis-ci-com pending completion   set done
778 master NUMBERS-182: Avoid overflow in LogBeta for tiny arguments push aherbert travis-ci-com pending completion   set done
777 master NUMBERS-180: Track changes push aherbert travis-ci-com pending completion   set done
776 master NUMBERS-180: Gamma ratio and delta ratio Added support for the ratio of two gamma functions. Copied the test data from the source Boost implementation. push aherbert travis-ci-com pending completion   set done
775 master Remove redundant variable assignment Use e directly and update to d in the condition that accepts the interpolation. The condition that rejects interpolation updates e anyway. push aherbert travis-ci-com pending completion   set done
774 master Avoid underflow computing the initial value This is an extreme edge case for the solver called with no bracketing. It ensures the bracket can be used to return a value if within the current function accuracy tolerance. push aherbert travis-ci-com pending completion   set done
773 master Avoid overflow computing the initial value push aherbert travis-ci-com pending completion   set done
772 master NUMBERS-179: Update threshold in regularisedGammaPrefix Modifies the threshold from a > 150 (Boost version) to a > 128. push aherbert travis-ci-com pending completion   set done
771 master Lower threshold for a ~ z This threshold requires optimisation. The threshold was too high to pass existing tests in Commons Statistics (see STATISTICS-39). push aherbert travis-ci-com pending completion   set done
770 master Move the direct computation before checks for base near 1 When the direct computation can be used then there is no requirement for the Lanczos code that uses pow(z / (z+g-0.5), a). Thus there is no need to avoid a base near 1 in the power functio... push aherbert travis-ci-com pending completion   set done
769 master Add another log1pmx test case to hit code coverage push aherbert travis-ci-com pending completion   set done
767 master NUMBERS-178: Add Factorial.doubleValue(int) Deprecate the FactorialDouble class. push aherbert travis-ci-com pending completion   set done