diff options
| author | H. Peter Anvin <hpa@zytor.com> | 2017-04-05 21:47:20 -0700 |
|---|---|---|
| committer | H. Peter Anvin <hpa@zytor.com> | 2017-04-05 21:47:20 -0700 |
| commit | 3646e7dde050c06101d1b1bb4dd4f944354777de (patch) | |
| tree | ebfbaaa7dbaf8fbd9a840aa78f6d4c750cb35345 | |
| parent | 90cc4154ae1dd5152ba0c3a47ab8ae357fd0530a (diff) | |
| download | nasm-3646e7dde050c06101d1b1bb4dd4f944354777de.tar.gz | |
BR 3392399: float: get rid of bogus exponent calculation
It turns out that the calculation of "twopwr" in ieee_flconvert_bin()
was more complex than necessary, and wrong in the case of a pure
fraction.
Reported-by: Roel <roelsuidgeest@zonnet.nl>
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
| -rw-r--r-- | asm/float.c | 4 | ||||
| -rw-r--r-- | test/hexfp.asm | 25 |
2 files changed, 27 insertions, 2 deletions
diff --git a/asm/float.c b/asm/float.c index b35d7a5e..fd66ef38 100644 --- a/asm/float.c +++ b/asm/float.c @@ -1,6 +1,6 @@ /* ----------------------------------------------------------------------- * * - * Copyright 1996-2009 The NASM Authors - All Rights Reserved + * Copyright 1996-2017 The NASM Authors - All Rights Reserved * See the file AUTHORS included with the NASM distribution for * the specific copyright holders. * @@ -545,7 +545,7 @@ static bool ieee_flconvert_bin(const char *string, int bits, mp = mult; ms = (LIMB_BITS-1)-l; - twopwr = seendot ? twopwr-bits+l : l+1-bits; + twopwr += l+1-bits; } if (seendigit) { diff --git a/test/hexfp.asm b/test/hexfp.asm new file mode 100644 index 00000000..02c972f8 --- /dev/null +++ b/test/hexfp.asm @@ -0,0 +1,25 @@ +;; BR 3392399 + +; All of these should be the same value... +%macro fp 1 + %1 0.5 + %1 5e-1 + %1 0x1.0p-1 + %1 0x0.8p0 + %1 0x0.8 + %1 0x8p-4 + %1 0x.8 + %1 0x1p-1 + %1 0x0.1p3 + %1 0x0.01p7 + %1 0x0.01p7 + %1 0x0.001p11 + +%endmacro + + fp do + fp dt + fp dq + fp dd + fp dw + fp dd |