avx2, two-table: don't segfault for NaN input
authorEll <ell_se@yahoo.com>
Mon, 19 Aug 2019 14:33:29 +0000 (17:33 +0300)
committerEll <ell_se@yahoo.com>
Mon, 19 Aug 2019 14:37:15 +0000 (17:37 +0300)
commit8f58a1bac7a2aff41a3b13d00d5888e13ac46f5d
tree567c10a2ab5d0c96bd5928c8d69334e89be8f5c6
parentee4a8ba6399eb9864f5bccabef02f6a7b3ea6974
avx2, two-table: don't segfault for NaN input

In the AVX2 and two-table linear-float => gamma-int8 conversions,
tweak the input bounds-check to handle NaN values.  NaN would
previously lead to an out-of-bounds table lookup, and a segfault
(see issue #43).
extensions/avx2-int8.c
extensions/two-table.c