Skip floating-point edge-case tests when using an x87
authorBenjamin Barenblat <bbaren@google.com>
Mon, 7 Feb 2022 16:54:21 +0000 (16:54 +0000)
committerBenjamin Barenblat <bbaren@debian.org>
Mon, 7 Feb 2022 16:54:21 +0000 (16:54 +0000)
commit007c573c38672625c7bb41c8e26176ee30d9cbea
tree635a9671cd895d45c42442ce277a88b7be8aff19
parent32c854e9b7305e40c3f506164f80852b4c619a93
Skip floating-point edge-case tests when using an x87

Forwarded: yes
Applied-Upstream: https://github.com/abseil/abseil-cpp/commit/311bbd2e50ea35e921a08186840d3b6ca279e880

32-bit Intel CPUs use 80-bit floats for intermediate values, which can
change the results of floating point computations from what we normally
expect. Identify tests that are sensitive to the x87, and skip them when
we’re on 32-bit Intel.

The author works at Google. Upstream applied this patch as Piper
revision 378722613 and exported it to GitHub; the Applied-Upstream URL
above points to the exported commit.

Gbp-Pq: Name float-tests-disable-i386.diff
absl/random/beta_distribution_test.cc
absl/random/distributions_test.cc
absl/random/exponential_distribution_test.cc
absl/random/uniform_real_distribution_test.cc
absl/time/duration_test.cc