[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[lmi] MinGW-w64 gcc-6.3.0 std::rint() anomaly
|
From: |
Greg Chicares |
|
Subject: |
[lmi] MinGW-w64 gcc-6.3.0 std::rint() anomaly |
|
Date: |
Sat, 19 Aug 2017 16:36:30 +0000 |
|
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.2.1 |
Vadim--Do you have any idea why 'fenv_lmi_test' fails with MinGW-w64
gcc-6.3.0, whereas it succeeded with earlier versions? The error
messages below are from cygwin, but I saw the same outcome with a
debian cross build. On debian at least, the test succeeds with
'build_type=safestdlib', perhaps because that build type turns off
optimization. The failures are due to std::rint() not respecting
the current rounding direction.
Running fenv_lmi_test:
???? test failed: '-3' == '-2'
[file /lmi/src/lmi/fenv_lmi_test.cpp, line 197]
???? test failed: '1' == '2'
[file /lmi/src/lmi/fenv_lmi_test.cpp, line 199]
???? test failed: '-1' == '-2'
[file /lmi/src/lmi/fenv_lmi_test.cpp, line 205]
???? test failed: '3' == '2'
[file /lmi/src/lmi/fenv_lmi_test.cpp, line 207]
???? test failed: '-1' == '-2'
[file /lmi/src/lmi/fenv_lmi_test.cpp, line 212]
???? test failed: '1' == '2'
[file /lmi/src/lmi/fenv_lmi_test.cpp, line 213]
[...]
???? 6 test errors detected; 55 tests succeeded
- [lmi] MinGW-w64 gcc-6.3.0 std::rint() anomaly,
Greg Chicares <=