[Bug build/15414] New: Build fails with mingw32-w64 GCC-4.8.0

[Bug build/15414] New: Build fails with mingw32-w64 GCC-4.8.0

[freddie_chopin at op dot pl]

http://sourceware.org/bugzilla/show_bug.cgi?id=15414

             Bug #: 15414
           Summary: Build fails with mingw32-w64 GCC-4.8.0
           Product: gdb
           Version: HEAD
            Status: NEW
          Severity: normal
          Priority: P2
         Component: build
        AssignedTo: unassigned@sourceware.org
        ReportedBy: freddie_chopin@op.pl
    Classification: Unclassified


When building GDB from repo with mingw32-w64 compiler, GCC 4.8.0 there's a
problem with compilation flags:

i686-w64-mingw32-gcc
-I/home/freddie/bleeding-edge-toolchain/build-mingw/host-libs/zlib/include -O2
-D__USE_MINGW_ACCESS   -I. -I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/common
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/config
-DLOCALEDIR="\"/home/freddie/bleeding-edge-toolchain/install-mingw/share/locale\""
-DHAVE_CONFIG_H
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/../include/opcode
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/../opcodes/..
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/../readline/.. -I../bfd
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/../bfd
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/../include
-I../libdecnumber
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/../libdecnumber 
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/gnulib/import
-Ibuild-gnulib/import  
-I/home/freddie/bleeding-edge-toolchain/build-mingw/host-libs/zlib/include
-I/home/freddie/bleeding-edge-toolchain/build-mingw/host-libs/usr/include 
-Wall -Wdeclaration-after-statement -Wpointer-arith -Wformat-nonliteral
-Wpointer-sign -Wno-unused -Wunused-value -Wunused-function -Wno-switch
-Wno-char-subscripts -Wmissing-prototypes -Wdeclaration-after-statement
-Wempty-body -Wno-format -Werror -c -o gdb.o -MT gdb.o -MMD -MP -MF
.deps/gdb.Tpo /home/freddie/bleeding-edge-toolchain/src/gdb/gdb/gdb.c
i686-w64-mingw32-gcc
-I/home/freddie/bleeding-edge-toolchain/build-mingw/host-libs/zlib/include -O2
-D__USE_MINGW_ACCESS   -I. -I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/common
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/config
-DLOCALEDIR="\"/home/freddie/bleeding-edge-toolchain/install-mingw/share/locale\""
-DHAVE_CONFIG_H
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/../include/opcode
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/../opcodes/..
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/../readline/.. -I../bfd
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/../bfd
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/../include
-I../libdecnumber
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/../libdecnumber 
-I/home/freddie/bleeding-edge-toolchain/src/gdb/gdb/gnulib/import
-Ibuild-gnulib/import  
-I/home/freddie/bleeding-edge-toolchain/build-mingw/host-libs/zlib/include
-I/home/freddie/bleeding-edge-toolchain/build-mingw/host-libs/usr/include 
-Wall -Wdeclaration-after-statement -Wpointer-arith -Wformat-nonliteral
-Wpointer-sign -Wno-unused -Wunused-value -Wunused-function -Wno-switch
-Wno-char-subscripts -Wmissing-prototypes -Wdeclaration-after-statement
-Wempty-body -Wno-format -Werror -c -o arm-tdep.o -MT arm-tdep.o -MMD -MP -MF
.deps/arm-tdep.Tpo /home/freddie/bleeding-edge-toolchain/src/gdb/gdb/arm-tdep.c
cc1: error: -Wformat-nonliteral ignored without -Wformat
[-Werror=format-nonliteral]
cc1: error: -Wformat-nonliteral ignored without -Wformat
[-Werror=format-nonliteral]
cc1: all warnings being treated as errors
cc1: all warnings being treated as errors

As you see because this is mingw32 -Wno-format is added, but it seems that in
GCC this makes -Wformat-... option (or at least -Wformat-nonliteral) invalid.
No such problem exists with native Linux build, there was no problem when I
built GDB 2 months ago with mingw32-w64 GCC 4.7.x.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug build/15414] Build fails with mingw32-w64 GCC-4.8.0

[tromey at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=15414

Tom Tromey <tromey at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tromey at redhat dot com

--- Comment #1 from Tom Tromey <tromey at redhat dot com> 2013-04-29 13:46:41 UTC ---
According to the GCC documentation, -Wall implies -Wformat.
So offhand I would say it is a GCC bug.
However, it seems like a reasonably simple workaround would
be to add -Wformat to build_warnings in gdb/configure.ac.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug build/15414] Build fails with mingw32-w64 GCC-4.8.0

[freddie_chopin at op dot pl]

http://sourceware.org/bugzilla/show_bug.cgi?id=15414

--- Comment #2 from Freddie Chopin <freddie_chopin at op dot pl> 2013-04-29 14:45:50 UTC ---
(In reply to comment #1)
> According to the GCC documentation, -Wall implies -Wformat.
> So offhand I would say it is a GCC bug.
> However, it seems like a reasonably simple workaround would
> be to add -Wformat to build_warnings in gdb/configure.ac.

But there's an EXPLICIT -Wno-format and the GDB changelog says it's there
because with mingw32 there may be some false positives for -Wformat. But these
two options conflict with each other - one saying to check for non-literal
strings (-Wformat-nonliteral), other saying to not check at all (-Wno-format).
Previously GCC probably ignored that silently, now it gives a warning, which
triggers an error because of -Werror. I don't think there's a bug, this
behavior makes perfect sense - GCC cannot do what you ask it to do, so it gives
you a warning.

I also have GCC 4.7.3 (mingw32) and there's the same behavior:

> gcc test.c -Wno-format -Wformat-nonliteral -Werror
cc1.exe: error: -Wformat-nonliteral ignored without -Wformat
[-Werror=format-non
literal]
cc1.exe: all warnings being treated as errors

> gcc --version
gcc (rev0, Built by MinGW-builds project) 4.7.3
Copyright (C) 2012 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

>

If there's still need to disable these checks for mingw32 it has to be done
completely, so instead of just adding -Wno-format, -Wformat-nonliteral must be
reversed to -Wno-format-nonliteral. In that case there's no error:

> gcc test.c -Wno-format -Wno-format-nonliteral -Werror

>

test.c is just an empty main().

Previous - successful - compilations were done with GCC 4.7.2.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug build/15414] Build fails with mingw32-w64 GCC-4.8.0

[tromey at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=15414

--- Comment #3 from Tom Tromey <tromey at redhat dot com> 2013-04-29 17:05:34 UTC ---
Ok, I see.  Sorry about that, I misunderstood.

I think the logic in configure.ac for setting this up is not
correct:

build_warnings="-Wall -Wdeclaration-after-statement -Wpointer-arith \
-Wformat-nonliteral -Wpointer-sign \
-Wno-unused -Wunused-value -Wunused-function \
-Wno-switch -Wno-char-subscripts -Wmissing-prototypes \
-Wdeclaration-after-statement -Wempty-body"

# Enable -Wno-format by default when using gcc on mingw since many
# GCC versions complain about %I64.
case "${host}" in
  *-*-mingw32*) build_warnings="$build_warnings -Wno-format" ;;
esac


It seems to me it would be better to try -Wformat only on other
systems like:

diff --git a/gdb/configure.ac b/gdb/configure.ac
index bb7fbdd..7c2080c 100644
--- a/gdb/configure.ac
+++ b/gdb/configure.ac
@@ -1926,7 +1926,7 @@ fi
 # NOTE: If you change this list, remember to update
 # gdb/doc/gdbint.texinfo.
 build_warnings="-Wall -Wdeclaration-after-statement -Wpointer-arith \
--Wformat-nonliteral -Wpointer-sign \
+-Wpointer-sign \
 -Wno-unused -Wunused-value -Wunused-function \
 -Wno-switch -Wno-char-subscripts -Wmissing-prototypes \
 -Wdeclaration-after-statement -Wempty-body"
@@ -1935,6 +1935,7 @@ build_warnings="-Wall -Wdeclaration-after-statement
-Wpointer-arith \
 # GCC versions complain about %I64.
 case "${host}" in
   *-*-mingw32*) build_warnings="$build_warnings -Wno-format" ;;
+  *) build_warnings="$build_warnings -Wformat-nonliteral" ;;
 esac

 AC_ARG_ENABLE(build-warnings,


Could you give that a try?
You should be able to apply this patch directly to configure if you
can't re-run autoconf.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug build/15414] Build fails with mingw32-w64 GCC-4.8.0

[freddie_chopin at op dot pl]

http://sourceware.org/bugzilla/show_bug.cgi?id=15414

--- Comment #4 from Freddie Chopin <freddie_chopin at op dot pl> 2013-04-29 17:30:57 UTC ---
With GCC 4.8.0 (linux host and target) this command works:
$ gcc -Wall -Werror -Wno-format-nonliteral test.c
This doesn't:
$ gcc -Wall -Werror -Wformat-nonliteral -Wno-format test.c
cc1: error: -Wformat-nonliteral ignored without -Wformat
[-Werror=format-nonliteral]
cc1: all warnings being treated as errors

Interesting fact - this command works too:
$ gcc -Wall -Werror -Wformat-nonliteral -Wno-format-nonliteral test.c
I've tested that it in fact disables checking of non-literal format strings,
but it "looks bad" - like a hack (;

So I think the solution is to change "-Wformat-nonliteral" to
"-Wno-format-nonliteral" instead of adding "-Wno-format". Another possible
option is to add "-Wno-format-nonliteral" instead of "-Wno-format"...

<just saw the new comment>

I was thinking about the same thing - I'll try that in a moment and report
back.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug build/15414] Build fails with mingw32-w64 GCC-4.8.0

[freddie_chopin at op dot pl]

http://sourceware.org/bugzilla/show_bug.cgi?id=15414

--- Comment #5 from Freddie Chopin <freddie_chopin at op dot pl> 2013-04-29 18:15:57 UTC ---
(In reply to comment #3)
> Could you give that a try?
> You should be able to apply this patch directly to configure if you
> can't re-run autoconf.

I've added that change only to gdb/configure and that was enough for the build
to complete!

But for a complete patch it's probably necessary to change that in other places
that have exactly the same fragment, plus there are also a few places like
this:

bfd/configure-case "${host}" in
bfd/configure-  *-*-mingw32*)
bfd/configure-    if test "${GCC}" = yes -a -z "${ERROR_ON_WARNING}" ; then
bfd/configure:      GCC_WARN_CFLAGS="$GCC_WARN_CFLAGS -Wno-format"
bfd/configure-    fi
bfd/configure-    ;;
bfd/configure-  *) ;;

Thank you for your help! (;

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug build/15414] Build fails with mingw32-w64 GCC-4.8.0

[freddie_chopin at op dot pl]

http://sourceware.org/bugzilla/show_bug.cgi?id=15414

--- Comment #6 from Freddie Chopin <freddie_chopin at op dot pl> 2013-05-02 17:19:41 UTC ---
Created attachment 7011
  --> http://sourceware.org/bugzilla/attachment.cgi?id=7011
patch

above fix for all occurences of identical fragment

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug build/15414] Build fails with mingw32-w64 GCC-4.8.0

[tromey at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=15414

--- Comment #7 from Tom Tromey <tromey at redhat dot com> 2013-05-07 20:10:56 UTC ---
To send the patch for review, follow these instructions:
http://sourceware.org/gdb/contribute/
Thanks!

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug build/15414] Build fails with mingw32-w64 GCC-4.8.0

[cvs-commit at gcc dot gnu.org]

http://sourceware.org/bugzilla/show_bug.cgi?id=15414

--- Comment #8 from cvs-commit at gcc dot gnu.org <cvs-commit at gcc dot gnu.org> 2013-05-10 16:10:42 UTC ---
CVSROOT:    /cvs/src
Module name:    src
Changes by:    tromey@sourceware.org    2013-05-10 16:10:40

Modified files:
    sim/bfin       : ChangeLog configure 
    sim/common     : ChangeLog acinclude.m4 
    sim/cr16       : ChangeLog configure 
    sim/cris       : ChangeLog configure 
    sim/d10v       : ChangeLog configure 
    sim/igen       : ChangeLog configure 
    sim/m68hc11    : ChangeLog configure 
    sim/mips       : ChangeLog configure 
    sim/mn10300    : ChangeLog configure 
    sim/v850       : ChangeLog configure 
    gdb            : ChangeLog configure configure.ac 

Log message:
    gdb
    Freddie Chopin  <freddie_chopin@op.pl>
    PR build/15414:
    * configure: Rebuild.
    * configure.ac (build_warnings): Do not use -Wformat-nonliteral
    with -Wno-format.
    sim/common
    * acinclude.m4 (SIM_AC_OPTION_WARNINGS): Do not use
    -Wformat-nonliteral with -Wno-format.
    sim/bfin
    * configure: Rebuild.
    sim/cr16
    * configure: Rebuild.
    sim/cris
    * configure: Rebuild.
    sim/d10v
    * configure: Rebuild.
    sim/igen
    * configure: Rebuild.
    sim/m68hc11
    * configure: Rebuild.
    sim/mips
    * configure: Rebuild.
    sim/mn10300
    * configure: Rebuild.
    sim/v850
    * configure: Rebuild.

Patches:
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/bfin/ChangeLog.diff?cvsroot=src&r1=1.93&r2=1.94
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/bfin/configure.diff?cvsroot=src&r1=1.17&r2=1.18
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/common/ChangeLog.diff?cvsroot=src&r1=1.285&r2=1.286
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/common/acinclude.m4.diff?cvsroot=src&r1=1.10&r2=1.11
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/cr16/ChangeLog.diff?cvsroot=src&r1=1.19&r2=1.20
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/cr16/configure.diff?cvsroot=src&r1=1.12&r2=1.13
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/cris/ChangeLog.diff?cvsroot=src&r1=1.13&r2=1.14
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/cris/configure.diff?cvsroot=src&r1=1.20&r2=1.21
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/d10v/ChangeLog.diff?cvsroot=src&r1=1.49&r2=1.50
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/d10v/configure.diff?cvsroot=src&r1=1.24&r2=1.25
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/igen/ChangeLog.diff?cvsroot=src&r1=1.33&r2=1.34
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/igen/configure.diff?cvsroot=src&r1=1.11&r2=1.12
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/m68hc11/ChangeLog.diff?cvsroot=src&r1=1.63&r2=1.64
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/m68hc11/configure.diff?cvsroot=src&r1=1.26&r2=1.27
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/mips/ChangeLog.diff?cvsroot=src&r1=1.172&r2=1.173
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/mips/configure.diff?cvsroot=src&r1=1.49&r2=1.50
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/mn10300/ChangeLog.diff?cvsroot=src&r1=1.54&r2=1.55
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/mn10300/configure.diff?cvsroot=src&r1=1.28&r2=1.29
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/v850/ChangeLog.diff?cvsroot=src&r1=1.54&r2=1.55
http://sourceware.org/cgi-bin/cvsweb.cgi/src/sim/v850/configure.diff?cvsroot=src&r1=1.24&r2=1.25
http://sourceware.org/cgi-bin/cvsweb.cgi/src/gdb/ChangeLog.diff?cvsroot=src&r1=1.15545&r2=1.15546
http://sourceware.org/cgi-bin/cvsweb.cgi/src/gdb/configure.diff?cvsroot=src&r1=1.390&r2=1.391
http://sourceware.org/cgi-bin/cvsweb.cgi/src/gdb/configure.ac.diff?cvsroot=src&r1=1.202&r2=1.203

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug build/15414] Build fails with mingw32-w64 GCC-4.8.0

[tromey at redhat dot com]

http://sourceware.org/bugzilla/show_bug.cgi?id=15414

Tom Tromey <tromey at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|---                         |7.7

--- Comment #9 from Tom Tromey <tromey at redhat dot com> 2013-05-10 16:13:36 UTC ---
Fixed.

-- 
Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.