[Bug c++/19115] New: __builtin_frexpl and std::frexp(long double) broken

[Bug c++/19115] New: __builtin_frexpl and std::frexp(long double) broken

[jlquinn at gcc dot gnu dot org]

The attached test generates the following output:

i=-803365344
i=1

Both __builtin_frexpl and std::frexp(long double) give the above incorrect output.

However, ::frexpl works as expected.

-- 
           Summary: __builtin_frexpl and std::frexp(long double) broken
           Product: gcc
           Version: 4.0.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: jlquinn at gcc dot gnu dot org
                CC: gcc-bugs at gcc dot gnu dot org
  GCC host triplet: powerpc-ibm-aix5.2.0.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19115

[Bug c++/19115] __builtin_frexpl and std::frexp(long double) broken

[jlquinn at gcc dot gnu dot org]

------- Additional Comments From jlquinn at gcc dot gnu dot org  2004-12-21 19:37 -------
Created an attachment (id=7799)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=7799&action=view)
frexp testcase


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19115

[Bug target/19115] __builtin_frexpl and std::frexp(long double) broken

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-12-21 19:40 -------
I should note that long double support for aix needs a little tweak to work.  (at least that is what I 
remembered hearing from David).

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|c++                         |target
   GCC host triplet|powerpc-ibm-aix5.2.0.0      |
 GCC target triplet|                            |powerpc-ibm-aix5.2.0.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19115

[Bug target/19115] __builtin_frexpl and std::frexp(long double) broken

[dje at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dje at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19115

[Bug target/19115] __builtin_frexpl and std::frexp(long double) broken

[dje at gcc dot gnu dot org]

------- Additional Comments From dje at gcc dot gnu dot org  2004-12-25 20:14 -------
Is this compiled with -mlong-double-128 or using the default, which is
-mlong-double-64 on AIX?  GCC currently does not build a long-double-128
multilib on AIX, which would be necessary for that mode to work completely
correctly.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19115

[Bug target/19115] __builtin_frexpl and std::frexp(long double) broken

[jlquinn at gcc dot gnu dot org]

------- Additional Comments From jlquinn at gcc dot gnu dot org  2004-12-27 22:28 -------
I did the stock configuration, with the exception of specifying an install dir
and --with-cpu=power4.  So the floating point is whatever comes out of the box.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19115

[Bug target/19115] __builtin_frexpl and std::frexp(long double) broken

[dje at gcc dot gnu dot org]

------- Additional Comments From dje at gcc dot gnu dot org  2004-12-28 00:10 -------
libstdc++ is calling frexpl in libc.a, which expects an AIX format long double
value, not a double.  Things work if one uses -mlong-double-128.  At some point
we need to default to long-double-128 for AIX for things to be self-consistent.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19115

[Bug target/19115] long double should be changed over to 128bit by default

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-12-28 20:51 -------
Confirmed, changing the summary to refect reality.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2004-12-28 20:51:42
               date|                            |
            Summary|__builtin_frexpl and        |long double should be
                   |std::frexp(long double)     |changed over to 128bit by
                   |broken                      |default


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19115

[Bug target/19115] long double should be changed over to 128bit by default

[jlquinn at gcc dot gnu dot org]

------- Additional Comments From jlquinn at gcc dot gnu dot org  2005-02-18 02:59 -------
(In reply to comment #5)
> libstdc++ is calling frexpl in libc.a, which expects an AIX format long double
> value, not a double.  Things work if one uses -mlong-double-128.  At some point
> we need to default to long-double-128 for AIX for things to be self-consistent.

What are the concerns with making this switch?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19115

[Bug target/19115] long double should be changed over to 128bit by default

[dje at gcc dot gnu dot org]

------- Comment #8 from dje at gcc dot gnu dot org  2006-07-11 16:59 -------
Switching actually causes problems on AIX due to AIX math.h header enabling
additional macros with types that conflict with GCC.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19115

[Bug target/19115] long double should be changed over to 128bit by default

[dje at gcc dot gnu dot org]

------- Comment #9 from dje at gcc dot gnu dot org  2008-08-31 14:36 -------
This default was changed for GCC targeted at AIX 6.1.


-- 

dje at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=19115