[Bug libgomp/47758] New: 729 unexpected failures in the libgomp test suite on powerpc-apple-darwin8

[Bug libgomp/47758] New: 729 unexpected failures in the libgomp test suite on powerpc-apple-darwin8

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47758

           Summary: 729 unexpected failures in the libgomp test suite on
                    powerpc-apple-darwin8
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: libgomp
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: dominiq@lps.ens.fr
                CC: jakub@redhat.com, burnus@net-b.de, iains@gcc.gnu.org
              Host: powerpc-apple-darwin8
            Target: powerpc-apple-darwin8
             Build: powerpc-apple-darwin8


I ventured in building gcc 4.6.0 (trunk revision 170169) on
powerpc-apple-darwin8 configured with:

../work/configure --prefix=/opt/gcc/gcc4.6w --target=powerpc-apple-darwin8
--host=powerpc-apple-darwin8 --build=powerpc-apple-darwin8
--enable-languages=c,c++,fortran,lto --with-gmp=/sw --with-libiconv-prefix=/sw
--with-system-zlib --enable-checking=release --with-dwarf2 --disable-multilib
--disable-libquadmath --disable-libquadmath-support

and I got a high number of failures in libgomp.fortran:

        === libgomp Summary ===

# of expected passes        1793
# of unexpected failures    729
# of unsupported tests        2

FAIL: libgomp.fortran/allocatable1.f90  -O0  (test for excess errors)
FAIL: libgomp.fortran/allocatable1.f90  -O1  (test for excess errors)
FAIL: libgomp.fortran/allocatable1.f90  -O2  (test for excess errors)
FAIL: libgomp.fortran/allocatable1.f90  -O3 -fomit-frame-pointer  (test for
excess errors)
FAIL: libgomp.fortran/allocatable1.f90  -O3 -fomit-frame-pointer -funroll-loops
 (test for excess errors)
FAIL: libgomp.fortran/allocatable1.f90  -O3 -fomit-frame-pointer
-funroll-all-loops -finline-functions  (test for excess errors)
FAIL: libgomp.fortran/allocatable1.f90  -O3 -g  (test for excess errors)
FAIL: libgomp.fortran/allocatable1.f90  -Os  (test for excess errors)
...

Most (if not all) are of the following kind:

/libgomp/testsuite/.. -shared-libgcc -fmessage-length=0 -fopenmp
-B/opt/gcc/build/powerpc-apple-darwin8/./libgomp/../libquadmath/.libs/  -O0  
-B/opt/gcc/build/powerpc-apple-darwin8/./libgomp/../libgfortran/.libs  
-L/opt/gcc/build/powerpc-apple-darwin8/./libgomp/.libs -lgomp
-L/opt/gcc/build/powerpc-apple-darwin8/./libgomp/../libquadmath/.libs/
-L/opt/gcc/build/powerpc-apple-darwin8/./libgomp/../libgfortran/.libs
-lgfortran -lm   -o ./allocatable1.exe    (timeout = 300)
/usr/libexec/gcc/powerpc-apple-darwin8/4.0.1/ld: warning -L: directory name
(/opt/gcc/build/powerpc-apple-darwin8/./libgomp/../libquadmath/.libs/) does not
exist

I have tested different variants around the options  --disable-libquadmath
--disable-libquadmath-support without change. Note that with these two options
no libquadmath is build while I don't need them on powerpc-apple-darwin9 (so I
see powerpc-apple-darwin9/libquadmath, but no
powerpc-apple-darwin9/libquadmath/.libs and the libgomp tests pass without
failures).

[Bug libgomp/47758] 729 unexpected failures in the libgomp test suite on powerpc-apple-darwin8

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47758

--- Comment #1 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-02-15 22:21:28 UTC ---
Created attachment 23356
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23356
untested patch

Do you know when it last worked? My candidate would be the patch - assuming
that it is not a new issue.

http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=167127
2010-11-24  Iain Sandoe  <iains@gcc.gnu.org>

        * testsuite/libgomp.fortran/fortran.exp: Add paths for libquadmath.

Untested patch attached.

[Bug libgomp/47758] 729 unexpected failures in the libgomp test suite on powerpc-apple-darwin8

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47758

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

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

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-02-15 22:38:36 UTC ---
This is preapproved with a proper ChangeLog entry if it works.

[Bug libgomp/47758] 729 unexpected failures in the libgomp test suite on powerpc-apple-darwin8

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47758

--- Comment #3 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2011-02-16 12:51:44 UTC ---
> Do you know when it last worked? 

Well it was my first attempt to build 4.6 on my poor G4 (almost ten year old,
so slow with a crowded disk that I don't even have 4.5 installed in it;-). This
build was motivated by pr45381 and I stumbled on this PR only by chance.

> My candidate would be the patch - assuming that it is not a new issue.

The patch in comment #1 fixes the failures and has been tested also on
powerpc-apple-darwin9 and x86_64-apple-darwin10 without problem.

Thanks for the quick fix.

[Bug libgomp/47758] 729 unexpected failures in the libgomp test suite on powerpc-apple-darwin8

[iains at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47758

--- Comment #4 from Iain Sandoe <iains at gcc dot gnu.org> 2011-02-16 16:14:02 UTC ---
thanks, IIRC, when the patch was originally applied it was not possible to test
on Darwin 8 (libquadmath was broken for other reasons).

---

This is not an objection to the proposed solution - but - for the record:

I remain concerned that the libgomp test-suite is driven by xgcc (which does
not apply the fortran spec) and, thus, is different from the notional idea of
equivalence with the installed gfortran.

[if the fortran spec is used, I think you will find that the fails are also
resolved]

e.g.

Index: libgomp/testsuite/libgomp.fortran/fortran.exp
===================================================================
--- libgomp/testsuite/libgomp.fortran/fortran.exp       (revision 170210)
+++ libgomp/testsuite/libgomp.fortran/fortran.exp       (working copy)
@@ -8,6 +8,7 @@ set lang_library_path   "../libgfortran/.libs"
 set lang_link_flags    "-lgfortran"
 set lang_test_file_found 0
 set quadmath_library_path "../libquadmath/.libs"
+set lang_spec_path "../libgfortran"


 # Initialize dg.
@@ -44,6 +45,8 @@ if { $lang_test_file_found } {
        lappend ALWAYS_CFLAGS "ldflags=-L${blddir}/${quadmath_library_path}/"
        # Allow for spec subsitution.
        lappend ALWAYS_CFLAGS
"additional_flags=-B${blddir}/${quadmath_library_path}/"
+       # use the libgfortran spec.
+       lappend ALWAYS_CFLAGS
"additional_flags=-specs=${blddir}/${lang_spec_path}/libgfortran.spec"
        set ld_library_path
"$always_ld_library_path:${blddir}/${lang_library_path}:${blddir}/${quadmath_library_path}"
     } else {
         set ld_library_path "$always_ld_library_path"

[Bug libgomp/47758] 729 unexpected failures in the libgomp test suite on powerpc-apple-darwin8

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47758

--- Comment #5 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-02-16 17:44:48 UTC ---
Author: burnus
Date: Wed Feb 16 17:44:45 2011
New Revision: 170216

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=170216
Log:
2011-02-16  Tobias Burnus  <burnus@net-b.de>

        PR libgomp/47758
        * testsuite/libgomp.fortran/fortran.exp: Check for the existence
        of libquadmath.a before adding its libpath to ldflags.


Modified:
    trunk/libgomp/ChangeLog
    trunk/libgomp/testsuite/libgomp.fortran/fortran.exp

[Bug libgomp/47758] 729 unexpected failures in the libgomp test suite on powerpc-apple-darwin8

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47758

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

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

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-02-16 19:45:27 UTC ---
Fixed.