[Bug libfortran/58015] New: FAIL: gfortran.dg/round_4.f90: Unsatisfied symbol "nextafterl"

[Bug libfortran/58015] New: FAIL: gfortran.dg/round_4.f90: Unsatisfied symbol "nextafterl"

[danglin at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58015

            Bug ID: 58015
           Summary: FAIL: gfortran.dg/round_4.f90: Unsatisfied symbol
                    "nextafterl"
           Product: gcc
           Version: 4.9.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libfortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: danglin at gcc dot gnu.org
              Host: hppa64-hp-hpux11.11
            Target: hppa64-hp-hpux11.11
             Build: hppa64-hp-hpux11.11

spawn /test/gnu/gcc/objdir/gcc/testsuite/gfortran/../../gfortran
-B/test/gnu/gcc
/objdir/gcc/testsuite/gfortran/../../
-B/test/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/
/test/gnu/gcc/gcc/gcc/testsuite/gfortran.dg/round_4.f90
-fno-diagnostics-show-caret -fdiagnostics-color=never -O0 -pedantic-errors
-B/test/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.libs
-L/test/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.libs
-L/test/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libgfortran/.libs
-B/test/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libquadmath/.libs
-L/test/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libquadmath/.libs
-L/test/gnu/gcc/objdir/hppa64-hp-hpux11.11/./libquadmath/.libs -lm -o
./round_4.exe
ld: Unsatisfied symbol "nextafterl" in file /var/tmp//ccSPc0rP.o
1 errors.collect2: error: ld returned 1 exit statuscompiler exited with status
1output is:ld: Unsatisfied symbol "nextafterl" in file /var/tmp//ccSPc0rP.o1
errors.collect2: error: ld returned 1 exit status
FAIL: gfortran.dg/round_4.f90  -O0  (test for excess errors)

[Bug libfortran/58015] FAIL: gfortran.dg/round_4.f90: Unsatisfied symbol "nextafterl"

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58015

--- Comment #2 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
Is this PR different from pr58113 beside the missing nextafterl on
hppa64-hp-hpux11.11?

[Bug libfortran/58015] FAIL: gfortran.dg/round_4.f90: Unsatisfied symbol "nextafterl"

[dave.anglin at bell dot net]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58015

--- Comment #3 from dave.anglin at bell dot net ---
On 21-Sep-13, at 11:13 AM, dominiq at lps dot ens.fr wrote:

> Is this PR different from pr58113 beside the missing nextafterl on
> hppa64-hp-hpux11.11?

Don't know.  It looks like libquadmath has nextafter, so it may be  
possible
to fix this.

Dave
--
John David Anglin    dave.anglin@bell.net

[Bug libfortran/58015] FAIL: gfortran.dg/round_4.f90: Unsatisfied symbol "nextafterl"

[dave.anglin at bell dot net]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58015

--- Comment #4 from dave.anglin at bell dot net ---
On 9/21/2013 11:13 AM, dominiq at lps dot ens.fr wrote:
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58015
>
> --- Comment #2 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
> Is this PR different from pr58113 beside the missing nextafterl on
> hppa64-hp-hpux11.11?
I hacked c99_functions.c to provide nextafterl using nextafterq from 
libquadmath.  With
this, I see the bug in pr58113.

Regarding nextafterl, I'm thinking about an include hack to math.h for 
hppa*-*-hpux11*.
On all HP-UX systems, the "l" and "q" long double and quad math 
functions are equivalent.

Dave

[Bug libfortran/58015] FAIL: gfortran.dg/round_4.f90: Unsatisfied symbol "nextafterl"

[ebotcazou at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58015

--- Comment #6 from Eric Botcazou <ebotcazou at gcc dot gnu.org> ---
Why not just adapt the existing implementation of nextafterf to long double?

[Bug libfortran/58015] FAIL: gfortran.dg/round_4.f90: Unsatisfied symbol "nextafterl"

[dave.anglin at bell dot net]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58015

--- Comment #7 from dave.anglin at bell dot net ---
On 13-Dec-13, at 5:08 PM, ebotcazou at gcc dot gnu.org wrote:

> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58015
>
> --- Comment #6 from Eric Botcazou <ebotcazou at gcc dot gnu.org> ---
> Why not just adapt the existing implementation of nextafterf to long  
> double?


This requires frexpl and scalbnl which also don't currently exist.   
They have further dependencies.

What did work was a simple routine to call nextafterq in libquadmath.   
Quad and long double
are equivalent on hppa.

Then I thought I should create a long double port of libquadmath and  
integrate it into libgcc on hppa.
However, I haven't had time to complete this approach but it would  
provide a complete set of c99 long
double functions.

Dave
--
John David Anglin    dave.anglin@bell.net

[Bug libfortran/58015] FAIL: gfortran.dg/round_4.f90: Unsatisfied symbol "nextafterl"

[ro at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58015

--- Comment #8 from Rainer Orth <ro at gcc dot gnu.org> ---
Author: ro
Date: Tue Feb 25 09:27:25 2014
New Revision: 208114

URL: http://gcc.gnu.org/viewcvs?rev=208114&root=gcc&view=rev
Log:
Skip gfortran.dg/erf_3.F90 and gfortran.dg/round_4.f90 on Solaris 9 (PR
libfortran/59313, libfortran/58015)

    PR libfortran/59313
    * gfortran.dg/erf_3.F90: Skip on sparc*-*-solaris2.9*.

    PR libfortran/58015
    * gfortran.dg/round_4.f90: Skip on *-*-solaris2.9*.

Modified:
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/gfortran.dg/erf_3.F90
    trunk/gcc/testsuite/gfortran.dg/round_4.f90

[Bug libfortran/58015] FAIL: gfortran.dg/round_4.f90: Unsatisfied symbol "nextafterl"

[danglin at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58015

--- Comment #9 from John David Anglin <danglin at gcc dot gnu.org> ---
Author: danglin
Date: Mon Mar 24 00:32:41 2014
New Revision: 208781

URL: http://gcc.gnu.org/viewcvs?rev=208781&root=gcc&view=rev
Log:
    PR libfortran/59313
    PR libfortran/58015
    * gfortran.dg/erf_3.F90: Skip on hppa*-*-hpux*.
    * gfortran.dg/round_4.f90: Likewise.


Modified:
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/gfortran.dg/erf_3.F90
    trunk/gcc/testsuite/gfortran.dg/round_4.f90

[Bug libfortran/58015] FAIL: gfortran.dg/round_4.f90: Unsatisfied symbol "nextafterl"

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58015

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.9.0                       |4.9.1

--- Comment #10 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
GCC 4.9.0 has been released

[Bug libfortran/58015] FAIL: gfortran.dg/round_4.f90: Unsatisfied symbol "nextafterl"

[fxcoudert at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58015

Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |fxcoudert at gcc dot gnu.org
         Resolution|---                         |WONTFIX

--- Comment #11 from Francois-Xavier Coudert <fxcoudert at gcc dot gnu.org> ---
I don't think we want to provide a full long double fallback libc as part of
gfortran. So, I'll close this as WONTFIX: the testcases have been XFAIL'ed, and
the proper course now is to for those targets to provide a full libc, or for
users not to rely on long double types in the meantime.