On 12/16/2014 10:10 PM, Luc Maisonobe wrote: > Hi all, > > Can we consider MATH-1173 is solved, and postpone MATH-1146, MATH-1153 > (I was not able to integrate the patch without breaking other tests), > MATH-1158 (I don' tunderstand from the discussion if the patch > completely breaks compatibility or if part of it can be applied), > MATH-1172 (which has both a target at 3.4 and 4.0), and MATH-1176 (until > deeper analysis is performed about this being a real problem or being > numerical noise)?
from my POV, it is safe to postpone MATH-1176. The main issue is the default choice of singularity threshold which we should definitely adjust for 4.0, and the small differences with other implementations most likely numerical noise (it is in the range of 1e-16). > I would like to start release candidate tomorrow. ok great. Thomas --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@commons.apache.org For additional commands, e-mail: dev-h...@commons.apache.org
