projects / external / glibc.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: af6b179) | patch
Correct IBM long double nextafterl.
authorAlan Modra <amodra@gmail.com>
Wed, 2 Apr 2014 03:16:19 +0000 (13:46 +1030)
committerAlan Modra <amodra@gmail.com>
Wed, 2 Apr 2014 03:16:19 +0000 (13:46 +1030)
commitb0abbc21034f0e5edc49023d8fda0616173faf17
tree0fed607adb9e4140a33d412fe9d6ec187fd403actree | snapshot
parentaf6b17973cbc07ac06cfb40eeab5cc2391fb489acommit | diff
Correct IBM long double nextafterl.

Fix for values near a power of two, and some tidies.

[BZ #16739]
* sysdeps/ieee754/ldbl-128ibm/s_nextafterl.c (__nextafterl): Correct
output when value is near a power of two.  Use int64_t for lx and
remove casts.  Use decimal rather than hex exponent constants.
Don't use long double multiplication when double will suffice.
* math/libm-test.inc (nextafter_test_data): Add tests.
* NEWS: Add 16739 and 16786 to bug list.
ChangeLog diff | blob | history
NEWS diff | blob | history
math/libm-test.inc diff | blob | history
sysdeps/ieee754/ldbl-128ibm/s_nextafterl.c diff | blob | history
Domain: System / Toolchain;