projects / platform / upstream / coreclr.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: e210c1c) | patch
JIT: ensure float folding is done using float precision (#20578)
authorAndy Ayers <andya@microsoft.com>
Thu, 25 Oct 2018 15:17:49 +0000 (08:17 -0700)
committerGitHub <noreply@github.com>
Thu, 25 Oct 2018 15:17:49 +0000 (08:17 -0700)
commit1370fb92bc9c1ae9475e92f8638c827128c84527
treec38e5fdd8c21ba48b3e3d4c64825843292b2519ftree | snapshot
parente210c1cef74a6c245f91b8215772895e57a4813ecommit | diff
JIT: ensure float folding is done using float precision (#20578)

Cast float folded values back to float before assigning to the double
the jit uses for storing FP literal constants.

Fixes #20561.

No diffs on Core; exposing the original bug requires building RyuJit with
an older x86 C++ compiler that uses x87 floating point.
src/jit/gentree.cpp diff | blob | history
Domain: Dotnet / Core;