[Bug testsuite/63352] New: problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[Bug testsuite/63352] New: problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[richard at netbsd dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

            Bug ID: 63352
           Summary: problem with fmt_g0_1.f08  on i386-pc-solaris2.11
           Product: gcc
           Version: 4.9.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: testsuite
          Assignee: unassigned at gcc dot gnu.org
          Reporter: richard at netbsd dot org

Running the testsuite for gcc 4.9.1 I notice the following gfortran failures:
        === gfortran tests ===


Running target unix
FAIL: gfortran.dg/default_format_1.f90  -O0  execution test
FAIL: gfortran.dg/default_format_1.f90  -O1  execution test
FAIL: gfortran.dg/default_format_1.f90  -O2  execution test
FAIL: gfortran.dg/default_format_1.f90  -O3 -fomit-frame-pointer  execution
test
FAIL: gfortran.dg/default_format_1.f90  -O3 -fomit-frame-pointer -funroll-loops
 execution test
FAIL: gfortran.dg/default_format_1.f90  -O3 -fomit-frame-pointer
-funroll-all-loops -finline-functions  execution test
FAIL: gfortran.dg/default_format_1.f90  -O3 -g  execution test
FAIL: gfortran.dg/default_format_1.f90  -Os  execution test
FAIL: gfortran.dg/fmt_g0_1.f08  -O0  execution test
FAIL: gfortran.dg/fmt_g0_1.f08  -O1  execution test
FAIL: gfortran.dg/fmt_g0_1.f08  -O2  execution test
FAIL: gfortran.dg/fmt_g0_1.f08  -O3 -fomit-frame-pointer  execution test
FAIL: gfortran.dg/fmt_g0_1.f08  -O3 -fomit-frame-pointer -funroll-loops 
execution test
FAIL: gfortran.dg/fmt_g0_1.f08  -O3 -fomit-frame-pointer -funroll-all-loops
-finline-functions  execution test
FAIL: gfortran.dg/fmt_g0_1.f08  -O3 -g  execution test
FAIL: gfortran.dg/fmt_g0_1.f08  -Os  execution test


Concerning the fmt_g0_1.f08 cases, I notice in particular that while debugging
the following bits are the problem:
10          write(buffer, string) ':',1.0_8/3.0_8,':'                      │
11          if (buffer.ne.":.33333333333333331:") call abort               │

after 10, the value of buffer is:
(gdb) print buffer
$1 = ':.', '3' <repeats 16 times>, '2:', ' ' <repeats 30 times>

Testsuite problem or no?

This is running on SunOS 5.11 (illumos) pkgsrc trunk
Compiler version: 4.9.1 (GCC) 
Platform: i386-sun-solaris2.11
configure flags: --enable-languages='c obj-c++ objc go fortran c++'
--enable-shared --enable-long-long --with-local-prefix=/opt/local/gcc49
--enable-libssp --enable-threads=posix --with-boot-ldflags='-static-libstdc++
-static-libgcc -Wl,-R/opt/local/gcc47/lib/gcc/i386-sun-solaris2.11/4.7.3
-Wl,-R/opt/local/gcc47/lib -Wl,-R/opt/local/lib ' --with-system-zlib
--disable-nls --with-gmp=/opt/local --with-mpc=/opt/local
--with-mpfr=/opt/local --with-isl=/opt/local --disable-isl-version-check
--with-cloog=/opt/local --enable-__cxa_atexit
--with-gxx-include-dir=/opt/local/gcc49/include/c++/ --with-gnu-as
--with-as=/opt/local/bin/gas --without-gnu-ld --with-ld=/usr/bin/ld
--with-libiconv-prefix=/opt/local --prefix=/opt/local/gcc49
--build=i386-sun-solaris2.11 --host=i386-sun-solaris2.11
--infodir=/opt/local/gcc49/info --mandir=/opt/local/gcc49/man
>From gcc-bugs-return-462388-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Wed Sep 24 06:27:06 2014
Return-Path: <gcc-bugs-return-462388-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 4147 invoked by alias); 24 Sep 2014 06:27:05 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 4111 invoked by uid 48); 24 Sep 2014 06:27:02 -0000
From: "glisse at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/56253] fp-contract does not work with SSE and AVX FMAs (neither FMA4 nor FMA3)
Date: Wed, 24 Sep 2014 06:27:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: target
X-Bugzilla-Version: 4.7.2
X-Bugzilla-Keywords: missed-optimization
X-Bugzilla-Severity: normal
X-Bugzilla-Who: glisse at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-56253-4-F0G8hXZRpZ@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-56253-4@http.gcc.gnu.org/bugzilla/>
References: <bug-56253-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2014-09/txt/msg02222.txt.bz2
Content-length: 1031

https://gcc.gnu.org/bugzilla/show_bug.cgi?idV253

--- Comment #12 from Marc Glisse <glisse at gcc dot gnu.org> ---
(In reply to Agner Fog from comment #11)
> Thanks for the links Marc.
> You are right, the discussion in the gcc-patches mailing list ignores
> integer vectors. You need a solution that also allows optimizations on
> integer intrinsic functions (perhaps cast the vector type?).

If you follow the links, you can find:
https://gcc.gnu.org/ml/gcc-patches/2013-04/msg00374.html
where I handled some integer vector intrinsics as well (there were some bugs in
that patch, but the idea should be fine).

> The proposed solution of using vector extensions will not work on masked
> vector intrinsics in AVX512, so it wouldn't enable e.g. constant propagation
> through a masked intrinsic, but that is probably too much to ask for :)

I expect we can get most of the benefits from using vector extensions for very
little effort, while handling the esoteric intrinsics would be a lot more work
so it gets lower priority.

[Bug testsuite/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[richard at netbsd dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #1 from Richard PALO <richard at netbsd dot org> ---
This seems to be a bug in the write formatting for g0
Here is the compile with -fdump-parse-tree showing that the constant expression
is calculated exactly as '3.3333333333333331e-1_8'

Namespace: A-H: (REAL 4) I-N: (INTEGER 4) O-Z: (REAL 4)
procedure name = MAIN__
  symtree: 'MAIN__'      || symbol: 'MAIN__'       
    type spec : (UNKNOWN 0)
    attributes: (PROGRAM PUBLIC  SUBROUTINE)
  symtree: 'abort'       || symbol: 'abort'        
    type spec : (UNKNOWN 0)
    attributes: (PROCEDURE  SUBROUTINE)
  symtree: 'buffer'      || symbol: 'buffer'       
    type spec : (CHARACTER 50 1)
    attributes: (VARIABLE )
  symtree: 'string'      || symbol: 'string'       
    type spec : (CHARACTER 25 1)
    attributes: (VARIABLE IMPLICIT-SAVE)
    value: '(g0,g0,g0)               '

  code:
  WRITE UNIT=MAIN__:buffer FMT='(g0,g0,g0)'
  TRANSFER ':'
  TRANSFER 12340
  TRANSFER ':'
  DT_END
  IF (/= MAIN__:buffer ':12340:')
    CALL _gfortran_abort ()
  ENDIF
  WRITE UNIT=MAIN__:buffer FMT=MAIN__:string
  TRANSFER ':'
  TRANSFER 0
  TRANSFER ':'
  DT_END
  IF (/= MAIN__:buffer ':0:')
    CALL _gfortran_abort ()
  ENDIF
  WRITE UNIT=MAIN__:buffer FMT=MAIN__:string
  TRANSFER ':'
  TRANSFER 3.3333333333333331e-1_8
  TRANSFER ':'
  DT_END
  IF (/= MAIN__:buffer ':.33333333333333331:')
    CALL _gfortran_abort ()
  ENDIF
  WRITE UNIT=MAIN__:buffer FMT='(1x,a,g0,a)'
  TRANSFER ':'
  TRANSFER 3.3333333333333331e-1_8
  TRANSFER ':'
  DT_END
  IF (/= MAIN__:buffer ' :.33333333333333331:')
    CALL _gfortran_abort ()
  ENDIF
  WRITE UNIT=MAIN__:buffer FMT=MAIN__:string
  TRANSFER ':'
  TRANSFER 'hello'
  TRANSFER ':'
  DT_END
  IF (/= MAIN__:buffer ':hello:')
    CALL _gfortran_abort ()
  ENDIF
  WRITE UNIT=MAIN__:buffer FMT='(g0,g0,g0,g0)'
  TRANSFER ':'
  TRANSFER .true.
  TRANSFER .false.
  TRANSFER ':'
  DT_END
  IF (/= MAIN__:buffer ':TF:')
    CALL _gfortran_abort ()
  ENDIF
  WRITE UNIT=MAIN__:buffer FMT='(g0,g0,'','',g0,g0)'
  TRANSFER '('
  TRANSFER (complex 1.2344999999999999_8 2.4567000000000001_8)
  TRANSFER ')'
  DT_END
  IF (/= MAIN__:buffer '(1.2344999999999999,2.4567000000000001)')
    CALL _gfortran_abort ()
  ENDIF

[Bug fortran/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING
   Last reconfirmed|                            |2014-09-24
                 CC|                            |ro at CeBiTec dot Uni-Bielefeld.DE
     Ever confirmed|0                           |1

--- Comment #3 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
From

> (gdb) print buffer
> $1 = ':.', '3' <repeats 16 times>, '2:', ' ' <repeats 30 times>

Then I guess that the buffer content is 

:.33333333333333332:

and not

:.33333333333333331:

as expected by the test. Could you print the content of buffer after

    write(buffer, string) ':',1.0_8/3.0_8,':'

and

    write(buffer, '(1x,a,g0,a)') ':',1.0_8/3.0_8,':'

to check this guess? This look like a rounding problem on your platform (CCed
Rainer Orth). What is the result if you compile the test with --save-temps (see
PR323)?

[Bug fortran/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[richard at netbsd dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #4 from Richard PALO <richard at netbsd dot org> ---
Just checked with the native omnios gcc 4.8.1 and the problem already exists
there.

I commented as indicated to be able to get debugger output for 2nd part:
diff --git a/gcc/testsuite/gfortran.dg/fmt_g0_1.f08
b/gcc/testsuite/gfortran.dg/fmt_g0_1.f08
index ead6f81..3cceae6 100644
--- a/gcc/testsuite/gfortran.dg/fmt_g0_1.f08
+++ b/gcc/testsuite/gfortran.dg/fmt_g0_1.f08
@@ -8,7 +8,7 @@
     write(buffer, string) ':',0,':'
     if (buffer.ne.":0:") call abort
     write(buffer, string) ':',1.0_8/3.0_8,':'
-    if (buffer.ne.":.33333333333333331:") call abort
+!    if (buffer.ne.":.33333333333333331:") call abort
     write(buffer, '(1x,a,g0,a)') ':',1.0_8/3.0_8,':'
     if (buffer.ne." :.33333333333333331:") call abort
     write(buffer, string) ':',"hello",':'

with 4.9.1 buffer is as follows for the two writes with 1/3:
(gdb) p buffer
$1 = ':.', '3' <repeats 16 times>, '2:', ' ' <repeats 30 times>
(gdb) n
(gdb) p buffer
$2 = ' :.', '3' <repeats 16 times>, '2:', ' ' <repeats 29 times>

under 4.7.3:
(gdb) p buffer
$1 = ':.', '3' <repeats 16 times>, '1:', ' ' <repeats 30 times>
(gdb) n
(gdb) p buffer
$2 = ' :.', '3' <repeats 16 times>, '1:', ' ' <repeats 29 times>

[Bug fortran/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[richard at netbsd dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #5 from Richard PALO <richard at netbsd dot org> ---
Created attachment 33550
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33550&action=edit
output from --save-temps

[Bug fortran/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #6 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
Your are not doing what I asked for. Could you compile and run without/with
--save-temps the following test:

    character(25) :: string = "(g0,g0,g0)" 
    character(50) :: buffer
    write(buffer, string) ':',1.0_8/3.0_8,':'
    print *, "'", buffer, "'"
    print *, "'", " :.33333333333333331:", "'"
    write(buffer, '(1x,a,g0,a)') ':',1.0_8/3.0_8,':'
    print *, "'", buffer, "'"
    print *, "'", " :.33333333333333331:", "'"
end

and post the result?

[Bug fortran/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[richard at netbsd dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #7 from Richard PALO <richard at netbsd dot org> ---
$ cat test.f08
    character(25) :: string = "(g0,g0,g0)" 
    character(50) :: buffer
    write(buffer, string) ':',1.0_8/3.0_8,':'
    print *, "'", buffer, "'"
    print *, "'", " :.33333333333333331:", "'"
    write(buffer, '(1x,a,g0,a)') ':',1.0_8/3.0_8,':'
    print *, "'", buffer, "'"
    print *, "'", " :.33333333333333331:", "'"
end
$ gfortran  test.f08 -o test
$ ./test
 ':.33333333333333332:                              '
 ' :.33333333333333331:'
 ' :.33333333333333332:                             '
 ' :.33333333333333331:'

Output the same with/without
attached is the .s

[Bug fortran/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[richard at netbsd dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #8 from Richard PALO <richard at netbsd dot org> ---
Created attachment 33553
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33553&action=edit
test.s

[Bug fortran/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[richard at netbsd dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #10 from Richard PALO <richard at netbsd dot org> ---
Created attachment 33554
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33554&action=edit
test.s from 4.7.3

[Bug fortran/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[richard at netbsd dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #9 from Richard PALO <richard at netbsd dot org> ---
output from 4.7.3:
 ':.33333333333333331:                              '
 ' :.33333333333333331:'
 ' :.33333333333333331:                             '
 ' :.33333333333333331:'


and test.s.4.7.3 attached

[Bug target/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jvdelisle at gcc dot gnu.org
          Component|fortran                     |target

--- Comment #11 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
Comment 7 confirms my guess that there is a rounding problem on
i386-sun-solaris2.11 (the test is three-year old and succeeds on all the
revisions I have used on darwin, and all the tests I have looked at
https://gcc.gnu.org/ml/gcc-testresults/2014-09/).

This may be related to problems found with solaris during the fix for PR60128.

Note that there is no point to post assembly files: the problem is either in
libgfortran or (more probably) in the C library.

[Bug target/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[ro at CeBiTec dot Uni-Bielefeld.DE]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #12 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> ---
> --- Comment #11 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
> Comment 7 confirms my guess that there is a rounding problem on
> i386-sun-solaris2.11 (the test is three-year old and succeeds on all the
> revisions I have used on darwin, and all the tests I have looked at
> https://gcc.gnu.org/ml/gcc-testresults/2014-09/).
>
> This may be related to problems found with solaris during the fix for PR60128.
>
> Note that there is no point to post assembly files: the problem is either in
> libgfortran or (more probably) in the C library.

This seems to be an Illumos-specific problem to me: on Solaris 11.2/x86,
I get the same output for Dominique's testcase with gcc 4.9 and 4.8, and
both testcases Richard sees failing PASS for me.

    Rainer

[Bug target/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[richard at netbsd dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #13 from Richard PALO <richard at netbsd dot org> ---
Since this regresses on the same omnios (illumos) platform between gcc 4.7.3
and 4.8.1 are there some pointers on how to identify the issue in illumos?
(the fact the 4.8.1 tested is native omnios (R151012) eliminates any eventual
pkgsrc issue, for me at least).

that is, 'git diff local-4.7.3 local-4.8.1` seems like quite a bit of work.

otherwise, I'll have to port a more recent gdb because 7.6.1 balks under gcc
4.9.1
may be awhile before I have the time...

cheers

[Bug target/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[ro at CeBiTec dot Uni-Bielefeld.DE]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #14 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> ---
> --- Comment #13 from Richard PALO <richard at netbsd dot org> ---
[...]
> otherwise, I'll have to port a more recent gdb because 7.6.1 balks under gcc
> 4.9.1
> may be awhile before I have the time...

What porting is there to do?  For me, configure; make; make install just
works on Solaris.

    Rainer

[Bug target/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[richard at netbsd dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #15 from Richard PALO <richard at netbsd dot org> ---
(In reply to ro@CeBiTec.Uni-Bielefeld.DE from comment #14)
> > --- Comment #13 from Richard PALO <richard at netbsd dot org> ---
> [...]
> > otherwise, I'll have to port a more recent gdb because 7.6.1 balks under gcc
> > 4.9.1
> > may be awhile before I have the time...
> 
> What porting is there to do?  For me, configure; make; make install just
> works on Solaris.
> 
> 	Rainer

Well, I get this, for example:

In file included from opncls.c:26:0:
opncls.c: In function 'bfd_fopen':
bfd.h:529:65: error: right-hand operand of comma expression has no effect
[-Werror=unused-value]
 #define bfd_set_cacheable(abfd,bool) (((abfd)->cacheable = bool), TRUE)
                                                                 ^
opncls.c:263:5: note: in expansion of macro 'bfd_set_cacheable'
     bfd_set_cacheable (nbfd, TRUE);
     ^
>From gcc-bugs-return-463483-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Tue Oct 07 18:58:16 2014
Return-Path: <gcc-bugs-return-463483-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 31594 invoked by alias); 7 Oct 2014 18:58:15 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 31549 invoked by uid 55); 7 Oct 2014 18:58:12 -0000
From: "dave.anglin at bell dot net" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug ipa/63403] [5.0 Regression] ICE: in relative_time_benefit at ipa-inline.c:869
Date: Tue, 07 Oct 2014 18:58:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: ipa
X-Bugzilla-Version: 5.0
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: dave.anglin at bell dot net
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-63403-4-KeXGAJhufT@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-63403-4@http.gcc.gnu.org/bugzilla/>
References: <bug-63403-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2014-10/txt/msg00504.txt.bz2
Content-length: 362

https://gcc.gnu.org/bugzilla/show_bug.cgi?idc403

--- Comment #4 from dave.anglin at bell dot net ---
On 10/7/2014 2:48 PM, rsandifo at gcc dot gnu.org wrote:
> Can you try the patches I posted here:
>
> https://gcc.gnu.org/ml/gcc-patches/2014-09/msg02636.html
> https://gcc.gnu.org/ml/gcc-patches/2014-09/msg02637.html
Will do when I get home.

Thanks,
Dave

[Bug target/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[ro at CeBiTec dot Uni-Bielefeld.DE]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #16 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> ---
> Well, I get this, for example:
>
> In file included from opncls.c:26:0:
> opncls.c: In function 'bfd_fopen':
> bfd.h:529:65: error: right-hand operand of comma expression has no effect
> [-Werror=unused-value]
>  #define bfd_set_cacheable(abfd,bool) (((abfd)->cacheable = bool), TRUE)
>                                                                  ^
> opncls.c:263:5: note: in expansion of macro 'bfd_set_cacheable'
>      bfd_set_cacheable (nbfd, TRUE);
>      ^

Please configure with --disable-werror and file a gdb bug.  This is not
the place to discuss gdb build issues.

Thanks.
        Rainer

[Bug target/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[richard at netbsd dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #18 from Richard PALO <richard at netbsd dot org> ---
Created attachment 33676
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33676&action=edit
another test program

after coming across
http://stackoverflow.com/questions/16839658/printf-width-specificer-to-maintain-precision-of-floating-point-value
I modified the test program to the attached, and here is the output:
FLT_EVAL_METHOD = 2
DBL_MANT_DIG = 53
DECIMAL_DIG = 21
DBL_DIG = 15
e precision 10 => +3.3333333333e-01
f precision 10 => +0.3333333333
e precision 11 => +3.33333333333e-01
f precision 11 => +0.33333333333
e precision 12 => +3.333333333333e-01
f precision 12 => +0.333333333333
e precision 13 => +3.3333333333333e-01
f precision 13 => +0.3333333333333
e precision 14 => +3.33333333333333e-01
f precision 14 => +0.33333333333333
e precision 15 => +3.333333333333333e-01
f precision 15 => +0.333333333333333
e precision 16 => +3.3333333333333332e-01
f precision 16 => +0.3333333333333333
e precision 17 => +3.33333333333333312e-01
f precision 17 => +0.33333333333333332
e precision 18 => +3.333333333333333148e-01
f precision 18 => +0.333333333333333312
e precision 19 => +3.3333333333333331483e-01
f precision 19 => +0.3333333333333333148
e precision 20 => +3.33333333333333314830e-01
f precision 20 => +0.33333333333333331483
e precision 21 => +3.333333333333333148296e-01
f precision 21 => +0.333333333333333314830
e precision 22 => +3.3333333333333331482962e-01
f precision 22 => +0.3333333333333333148296
e precision 23 => +3.33333333333333314829616e-01
f precision 23 => +0.33333333333333331482962
e precision 24 => +3.333333333333333148296163e-01
f precision 24 => +0.333333333333333314829616
e precision 25 => +3.3333333333333331482961626e-01
f precision 25 => +0.3333333333333333148296163

[Bug target/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[richard at netbsd dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #20 from Richard PALO <richard at netbsd dot org> ---
>From what I've been able to gather, 'f' precision 17 & 18 incorrectly terminate
with a '2' where they seemingly need be '1' and '5' respectfully
(along with corresponding 'e' with precision 16 & 17).

The default rounding is FE_TONEAREST on illumos. The other modes seem okay.

Therefore I filed https://www.illumos.org/issues/5224  (with an updated version
of the test program).

BTW, on a sidenote, with the test program on NetBSD (6.1) I notice that
rounding mode seems to be completely ignored. Does gnu-c runtime expect correct
behaviour?
(NB I made a version of the test program to avoid any constant 1.0/3.0 being
used)... It would be interesting to know about other platforms... 
I came across a sourceware bug where this was supposedly fixed
(https://sourceware.org/bugzilla/show_bug.cgi?id=5044)

[Bug target/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[richard at netbsd dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

--- Comment #21 from Richard PALO <richard at netbsd dot org> ---
The problem has been isolated in illumos with a preliminary patch available.

This test now passes with the fix applied, therefore I believe the issue can be
closed as an OS issue and not a bug.

cheers

[Bug target/63352] problem with fmt_g0_1.f08 on i386-pc-solaris2.11

[ro at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63352

Rainer Orth <ro at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
                 CC|                            |ro at gcc dot gnu.org
         Resolution|---                         |INVALID

--- Comment #22 from Rainer Orth <ro at gcc dot gnu.org> ---
Ok, thanks for the update.

  Rainer