# See "What's new in v3.2.1" for this regression.

T=d f ld mf
V=x
R=-100, -1, .5
F=log(-x)
C=fp_log(-x)
