diff options
| author | Herbert Valerio Riedel <hvr@gnu.org> | 2014-11-19 11:09:33 +0100 |
|---|---|---|
| committer | Herbert Valerio Riedel <hvr@gnu.org> | 2014-11-19 11:37:07 +0100 |
| commit | e2af452cd533778c5447719c59429d72bb1fe00d (patch) | |
| tree | 5b8961b0a7ad233f861d160f2830067c399619cd /testsuite/tests/numeric | |
| parent | 42244668af6d8c1dd6a2d64af90ed57d8ecd8d88 (diff) | |
| download | haskell-e2af452cd533778c5447719c59429d72bb1fe00d.tar.gz | |
Restore exact old semantics of `decodeFloat`
`integer-gmp2` uses the new 64bit-based IEEE deconstructing primop
introduced in b62bd5ecf3be421778e4835010b6b334e95c5a56.
However, the returned values differ for exceptional IEEE values:
Previous (expected) semantics:
> decodeFloat (-1/0)
(-4503599627370496,972)
> decodeFloat (1/0)
(4503599627370496,972)
> decodeFloat (0/0)
(-6755399441055744,972)
Currently (broken) semantics:
> decodeFloat (-1/0 :: Double)
(-9223372036854775808,-53)
> decodeFloat (1/0 :: Double)
(-9223372036854775808,-53)
> decodeFloat (0/0 :: Double)
(-9223372036854775808,-53)
This patch reverts to the old expected semantics.
I plan to revisit the implementation during GHC 7.11 development.
This should address #9810
Reviewed By: austin, ekmett, luite
Differential Revision: https://phabricator.haskell.org/D486
Diffstat (limited to 'testsuite/tests/numeric')
| -rw-r--r-- | testsuite/tests/numeric/should_run/T9810.hs | 25 | ||||
| -rw-r--r-- | testsuite/tests/numeric/should_run/T9810.stdout | 14 | ||||
| -rw-r--r-- | testsuite/tests/numeric/should_run/all.T | 1 |
3 files changed, 40 insertions, 0 deletions
diff --git a/testsuite/tests/numeric/should_run/T9810.hs b/testsuite/tests/numeric/should_run/T9810.hs new file mode 100644 index 0000000000..b8ce1ba83e --- /dev/null +++ b/testsuite/tests/numeric/should_run/T9810.hs @@ -0,0 +1,25 @@ +main = do + -- NOTE: the `abs` is to compensate for WAY=optllvm + -- having a positive sign for 0/0 + + putStrLn "## Double ##" + print $ idRational ( 1/0 :: Double) + print $ idRational (-1/0 :: Double) + print $ abs $ idRational ( 0/0 :: Double) + print $ idReencode ( 1/0 :: Double) + print $ idReencode (-1/0 :: Double) + print $ abs $ idReencode ( 0/0 :: Double) + + putStrLn "## Float ##" + print $ idRational ( 1/0 :: Float) + print $ idRational (-1/0 :: Float) + print $ abs $ idRational ( 0/0 :: Float) + print $ idReencode ( 1/0 :: Float) + print $ idReencode (-1/0 :: Float) + print $ abs $ idReencode ( 0/0 :: Float) + where + idRational :: (Real a, Fractional a) => a -> a + idRational = fromRational . toRational + + idReencode :: (RealFloat a) => a -> a + idReencode = uncurry encodeFloat . decodeFloat diff --git a/testsuite/tests/numeric/should_run/T9810.stdout b/testsuite/tests/numeric/should_run/T9810.stdout new file mode 100644 index 0000000000..52a7e8f4eb --- /dev/null +++ b/testsuite/tests/numeric/should_run/T9810.stdout @@ -0,0 +1,14 @@ +## Double ## +Infinity +-Infinity +Infinity +Infinity +-Infinity +Infinity +## Float ## +Infinity +-Infinity +Infinity +Infinity +-Infinity +Infinity diff --git a/testsuite/tests/numeric/should_run/all.T b/testsuite/tests/numeric/should_run/all.T index 76181a2115..62622799b8 100644 --- a/testsuite/tests/numeric/should_run/all.T +++ b/testsuite/tests/numeric/should_run/all.T @@ -63,3 +63,4 @@ test('T7233', normal, compile_and_run, ['']) test('NumDecimals', normal, compile_and_run, ['']) test('T8726', normal, compile_and_run, ['']) test('CarryOverflow', omit_ways(['ghci']), compile_and_run, ['']) +test('T9810', normal, compile_and_run, ['']) |