[Bug math/16337] New: ldbl-128 lgammal overflows for small negative arguments

[jsm28 at gcc dot gnu.org]

https://sourceware.org/bugzilla/show_bug.cgi?id=16337

            Bug ID: 16337
           Summary: ldbl-128 lgammal overflows for small negative
                    arguments
           Product: glibc
           Version: 2.19
            Status: NEW
          Severity: normal
          Priority: P2
         Component: math
          Assignee: unassigned at sourceware dot org
          Reporter: jsm28 at gcc dot gnu.org

The ldbl-128 implementation of lgammal overflows for small negative arguments. 
For example:

Failure: lgamma (-0x1p-10000): errno set to 34, expected 0 (unchanged)
Failure: Test: lgamma (-0x1p-10000)
Result:
 is:          inf   inf
 should be:   6.93147180559945309417e+03   0x1.b1378c84073b880c67e5p+12

-- 
You are receiving this mail because:
You are on the CC list for the bug.