[Bug target/62251] New: [5.0 Regression] FAIL: gfortran.dg/quad_2.f90 execution test

[Bug target/62251] New: [5.0 Regression] FAIL: gfortran.dg/quad_2.f90 execution test

[danglin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62251

            Bug ID: 62251
           Summary: [5.0 Regression] FAIL: gfortran.dg/quad_2.f90
                    execution test
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: danglin at gcc dot gnu.org
              Host: hppa2.0w-hp-hpux11.11
            Target: hppa2.0w-hp-hpux11.11
             Build: hppa2.0w-hp-hpux11.11

Executing on host: /test/gnu/gcc/objdir/gcc/testsuite/gfortran/../../gfortran
-B/test/gnu/gcc/objdir/gcc/testsuite/gfortran/../../
-B/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libgfortran/
/test/gnu/gcc/gcc/gcc/testsuite/gfortran.dg/quad_2.f90 
-fno-diagnostics-show-caret -fdiagnostics-color=never    -O0   -pedantic-errors
 -B/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libgfortran/.libs -L/test
/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libgfortran/.libs
-L/test/gnu/gcc/objdir
/hppa2.0w-hp-hpux11.11/./libgfortran/.libs
-B/test/gnu/gcc/objdir/hppa2.0w-hp-hp
ux11.11/./libquadmath/.libs
-L/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libqu
admath/.libs -L/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libquadmath/.libs 
-
lm    -o ./quad_2.exe    (timeout = 300)
spawn /test/gnu/gcc/objdir/gcc/testsuite/gfortran/../../gfortran
-B/test/gnu/gcc/objdir/gcc/testsuite/gfortran/../../
-B/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libgfortran/
/test/gnu/gcc/gcc/gcc/testsuite/gfortran.dg/quad_2.f90
-fno-diagnostics-show-caret -fdiagnostics-color=never -O0 -pedantic-errors
-B/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libgfortran/.libs
-L/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libgfortran/.libs
-L/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libgfortran/.libs
-B/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libquadmath/.libs
-L/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libquadmath/.libs
-L/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libquadmath/.libs -lm -o
./quad_2.exePASS: gfortran.dg/quad_2.f90   -O0  (test for excess errors)
Setting LD_LIBRARY_PATH to
.:/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libgfortran/.libs:/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libgfortran/.libs:/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libquadmath/.libs:/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libquadmath/.libs:/test/gnu/gcc/objdir/gcc:.:/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libgfortran/.libs:/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libgfortran/.libs:/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libquadmath/.libs:/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libquadmath/.libs:/test/gnu/gcc/objdir/gcc
spawn [open ...]

Program aborted. Backtrace:
#0  0xC1B3A79F
FAIL: gfortran.dg/quad_2.f90   -O0  execution test

Revision 214122 was OK.

[Bug target/62251] [5.0 Regression] FAIL: gfortran.dg/quad_2.f90 execution test

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62251

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |5.0

[Bug target/62251] [5 Regression] FAIL: gfortran.dg/quad_2.f90 execution test

[danglin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62251

--- Comment #1 from John David Anglin <danglin at gcc dot gnu.org> ---
Test fails at line 58:
         if (str3 /= "   1.41421356237309504880168872420969798") call abort()

_gfortrani_compare_string (len1=len1@entry=80, 
    s1=0x800003fffdff0b90 "   1.41421356237309504880168872420969800", ' '
<repeats 40 times>, "1.41421356237309504880168872420969800", ' ' <repeats 43
times>, s1@entry=<error reading variable: can't compute CFA for this frame>, 
    len2=len2@entry=40, 
    s2=s2@entry=0x4000000000001a70 "  
1.414213562373095048801688724209697981.41421356237309504880168872420969798")
    at ../../../gcc/libgfortran/intrinsics/string_intrinsics_inc.c:85

(gdb) bt
#0  0x800003fffdfd7544 in _gfortrani_compare_string (len1=len1@entry=80, 
    s1=0x800003fffdff0b90 "   1.41421356237309504880168872420969800", ' '
<repeats 40 times>, "1.41421356237309504880168872420969800", ' ' <repeats 43
times>, s1@entry=<error reading variable: can't compute CFA for this frame>, 
    len2=len2@entry=40, 
    s2=s2@entry=0x4000000000001a70 "  
1.414213562373095048801688724209697981.41421356237309504880168872420969798")
    at ../../../gcc/libgfortran/intrinsics/string_intrinsics_inc.c:85
#1  0x40000000000036c0 in test_qp ()
    at /test/gnu/gcc/gcc/gcc/testsuite/gfortran.dg/quad_2.f90:58
#2  0x4000000000003840 in main (argc=<optimized out>, argv=<optimized out>)
    at /test/gnu/gcc/gcc/gcc/testsuite/gfortran.dg/quad_2.f90:78

With debug comments removed:
# ./quad_2.exe
>   1.00000000000000000000000000000000000<
>1.00000000000000000000000000000000000<
>   1.41421356237309504880168872420969800<
>1.41421356237309504880168872420969800<

Program aborted. Backtrace:
#0  0xC00000000033D86B
ABORT instruction (core dumped)

[Bug target/62251] [5 Regression] FAIL: gfortran.dg/quad_2.f90 execution test

[burnus at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62251

Tobias Burnus <burnus at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |burnus at gcc dot gnu.org

--- Comment #2 from Tobias Burnus <burnus at gcc dot gnu.org> ---
(In reply to John David Anglin from comment #1)
> Test fails at line 58:
>          if (str3 /= "   1.41421356237309504880168872420969798") call abort()
                        > >1.41421356237309504880168872420969800<

It looks as if there is a difference of 1 ULP.

Thus, one option would be to permit both results ("str /= ... and str /= ...").

[Bug target/62251] [5 Regression] FAIL: gfortran.dg/quad_2.f90 execution test

[burnus at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62251

--- Comment #3 from Tobias Burnus <burnus at gcc dot gnu.org> ---
Or, alternatively and possible better: to understand why there is now a ULP
difference, which didn't exist in GCC 4.9.

[Bug target/62251] [5 Regression] FAIL: gfortran.dg/quad_2.f90 execution test

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62251

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING
   Last reconfirmed|                            |2015-02-17
                 CC|                            |jakub at gcc dot gnu.org
     Ever confirmed|0                           |1

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Yes, can you please debug this?
Try e.g. GCC 5 compiled testcase against 4.9 libgfortran or vice versa, if the
difference is in the binaries, do side-by-side debugging of the 4.9 and 5
version in the debugger and watch where things differ, if it is on the library
side, identify which call differs?

[Bug target/62251] [5 Regression] FAIL: gfortran.dg/quad_2.f90 execution test

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62251

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P4

[Bug target/62251] FAIL: gfortran.dg/quad_2.f90 execution test

[danglin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62251

John David Anglin <danglin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[5 Regression] FAIL:        |FAIL:
                   |gfortran.dg/quad_2.f90      |gfortran.dg/quad_2.f90
                   |execution test              |execution test

--- Comment #5 from John David Anglin <danglin at gcc dot gnu.org> ---
This is not a regression.  In 4.9 and previous to r214187 on trunk, the test
was unresolved:

spawn /test/gnu/gcc/objdir/gcc/testsuite/gfortran/../../gfortran
-B/test/gnu/gcc
/objdir/gcc/testsuite/gfortran/../../
-B/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.
11/./libgfortran/ fortran_largest_fp_has_sqrt19380.f90
-fno-diagnostics-show-car
et -fdiagnostics-color=never
-B/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libg
fortran/.libs -L/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libgfortran/.libs
-L/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libgfortran/.libs
-B/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libquadmath/.libs
-L/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libquadmath/.libs
-L/test/gnu/gcc/objdir/hppa2.0w-hp-hpux11.11/./libquadmath/.libs -lm -o
fortran_largest_fp_has_sqrt19380.exe
/usr/ccs/bin/ld: Unsatisfied symbols:
   sqrtl (first referenced in /var/tmp//cc8YZhsr.o) (code)
collect2: error: ld returned 1 exit status
compiler exited with status 1
output is:
/usr/ccs/bin/ld: Unsatisfied symbols:
   sqrtl (first referenced in /var/tmp//cc8YZhsr.o) (code)
collect2: error: ld returned 1 exit status

UNSUPPORTED: gfortran.dg/quad_2.f90   -O0

sqrtl is not available on hpux.  It seems at some point before r214253
the test started running, probably using sqrtq from libquadmath.  I'm not
sure why this happened.

[Bug target/62251] FAIL: gfortran.dg/quad_2.f90 execution test

[dave.anglin at bell dot net]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62251

--- Comment #6 from dave.anglin at bell dot net ---
On 2015-02-20, at 1:49 PM, danglin at gcc dot gnu.org wrote:

> sqrtl is not available on hpux.  It seems at some point before r214253
> the test started running, probably using sqrtq from libquadmath.  I'm not
> sure why this happened.

The current code now uses the libcall, _U_Qfsqrt, in libm.

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

[Bug target/62251] FAIL: gfortran.dg/quad_2.f90 execution test

[dave.anglin at bell dot net]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62251

--- Comment #7 from dave.anglin at bell dot net ---
On 2015-02-21 1:56 PM, dave.anglin at bell dot net wrote:
> The current code now uses the libcall, _U_Qfsqrt, in libm.
The change from sqrtl to _U_Qfsqrt occurred in r214211:

2014-08-20  Joost VandeVondele <Joost.VandeVondele@mat.ethz.ch>

     * options.c (gfc_init_options_struct): assert that the frontend sets
         flag_errno_math and flag_associative_math.

[Bug target/62251] FAIL: gfortran.dg/quad_2.f90 execution test

[danglin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62251

John David Anglin <danglin at gcc dot gnu.org> changed:

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

--- Comment #9 from John David Anglin <danglin at gcc dot gnu.org> ---
Resolved.