Branch: refs/heads/master
  Home:   https://github.com/D-Programming-Language/phobos
  Commit: a3035d7d77e8ec5e807a9f2fa16db2c1f74c369c
      https://github.com/D-Programming-Language/phobos/commit/a3035d7d77e8ec5e807a9f2fa16db2c1f74c369c
  Author: Martin <noone@nowhere.com>
  Date:   2015-11-08 (Sun, 08 Nov 2015)

  Changed paths:
    M std/internal/math/gammafunction.d

  Log Message:
  -----------
  Slightly relax std.internal.math.gammafunction unittest for 64-bit reals

Using LDC master on Win64, I get a relative diff of 1.068e-15 for the second tuple (input = 1/8), exceeding the allowed upper bound by not even 7%. I chose to be generous and doubled the limit. ;)

More details for the 2nd tuple:
actual:   0x1.0a333fd8b6880p+2
expected: 0x1.0a333fd8b687bp+2
rel diff: 1.067682e-15

The other 4 tuples are perfect matches.


  Commit: 3db5a3bfe1db260dc494d64840a02de4c171843b
      https://github.com/D-Programming-Language/phobos/commit/3db5a3bfe1db260dc494d64840a02de4c171843b
  Author: David Nadlinger <code@klickverbot.at>
  Date:   2015-11-09 (Mon, 09 Nov 2015)

  Changed paths:
    M std/internal/math/gammafunction.d

  Log Message:
  -----------
  Merge pull request #3799 from kinke/gammafunction

Slightly relax std.internal.math.gammafunction unittest for 64-bit reals


Compare: https://github.com/D-Programming-Language/phobos/compare/2812f9a67656...3db5a3bfe1db