This is the mail archive of the glibc-bugs@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug math/15680] New: IBM long double inaccuracy


http://sourceware.org/bugzilla/show_bug.cgi?id=15680

            Bug ID: 15680
           Summary: IBM long double inaccuracy
           Product: glibc
           Version: 2.18
            Status: NEW
          Severity: normal
          Priority: P2
         Component: math
          Assignee: unassigned at sourceware dot org
          Reporter: amodra at gmail dot com

IBM long double values are printed inaccurately in base 10 floating point
formats when

a) the exponent of the long double approaches the lower end of its range, ie.
the low double is denormal, or,

b) the difference in exponents between the high and low double is greater than
53, or,

c) the difference in exponents is 53, the high and low doubles differ in sign,
and the high double is 0x1.0pnnn.

These user visible errors are all the result of bugs in
__mpn_extract_long_double, and similar bugs exist in related functions.

(c) is PR5268.
(b) is due to shift count exceeds width of type undefined behaviour.

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


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]