[Bug math/16315] New: Bad overflowing pow results for directed rounding

["jsm28 at gcc dot gnu.org" <sourceware-bugzilla at sourceware dot org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=16315

            Bug ID: 16315
           Summary: Bad overflowing pow results for directed rounding
           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 results of pow on overflow should respect the rounding mode and the sign of
the result.  Instead, for dbl-64 on x86_64 at least, they are infinities
regardless of the rounding mode, e.g.:

Failure: Test: pow_towardzero (-0x2p+0, 0x1.fffffep+112)
Result:
 is:          inf   inf
 should be:   1.79769313486231570814e+308   0x1.fffffffffffff0000000p+1023

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