[Bug other/54692] New: [4.8 Regression] gcc doesn't build with "-Og -g"

[Bug other/54692] New: [4.8 Regression] gcc doesn't build with "-Og -g"

[markus at trippelsdorf dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

             Bug #: 54692
           Summary: [4.8 Regression] gcc doesn't build with "-Og -g"
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: other
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: markus@trippelsdorf.de


With:
 gcc_build_dir % ~/gcc/configure --disable-werror --disable-multilib
--enable-languages=c,c++
 gcc_build_dir % make STAGE1_CFLAGS="-g -Og" all-stage1

I get:
...
g++ -c   -g -Og -DIN_GCC   -fno-exceptions -fno-rtti -W -Wall -Wno-narrowing
-Wwrite-strings -Wcast-qual -Wmissing-format-attribute -pedantic -Wno-long-long
-Wno-variadic-macros -Wno-overlength-strings -fno-common  -DHAVE_CONFIG_H
-DGENERATOR_FILE -I. -Ibuild -I/home/markus/gcc/gcc
-I/home/markus/gcc/gcc/build -I/home/markus/gcc/gcc/../include
-I/home/markus/gcc/gcc/../libcpp/include 
-I/home/markus/gcc/gcc/../libdecnumber
-I/home/markus/gcc/gcc/../libdecnumber/bid -I../libdecnumber    \
        -o build/genconstants.o /home/markus/gcc/gcc/genconstants.c
In file included from /home/markus/gcc/gcc/read-md.h:22:0,
                 from /home/markus/gcc/gcc/genconstants.c:32:
/home/markus/gcc/gcc/../include/obstack.h:153:40: error: attempt to use
poisoned "bcopy"
 #  define _obstack_memcpy(To, From, N) bcopy ((char *)(From), (To), (N))
                                        ^
In file included from ./bconfig.h:3:0,
                 from /home/markus/gcc/gcc/genconstants.c:28:
./auto-host.h:1988:17: error: multiple types in one declaration
 #define ssize_t int
                 ^
./auto-host.h:1988:17: error: declaration does not declare anything
[-fpermissive]
./auto-host.h:1976:15: error: multiple types in one declaration
 #define pid_t int
               ^
./auto-host.h:1976:15: error: declaration does not declare anything
[-fpermissive]
In file included from /home/markus/gcc/gcc/system.h:198:0,
                 from /home/markus/gcc/gcc/genconstants.c:29:
/usr/include/sys/types.h:116:26: error: expected unqualified-id before ‘;’
token
 typedef __caddr_t caddr_t;
                          ^
In file included from ./bconfig.h:3:0,
                 from /home/markus/gcc/gcc/genconstants.c:28:
./auto-host.h:1982:16: error: declaration does not declare anything
[-fpermissive]
 #define rlim_t long
                ^
In file included from /home/markus/gcc/gcc/genconstants.c:29:0:
/home/markus/gcc/gcc/system.h:447:48: error: new declaration ‘char*
strstr(const char*, const char*)’
 extern char *strstr (const char *, const char *);
                                                ^
In file included from
/usr/lib/gcc/x86_64-pc-linux-gnu/4.8.0/include/g++-v4/cstring:44:0,
                 from /home/markus/gcc/gcc/system.h:207,
                 from /home/markus/gcc/gcc/genconstants.c:29:
/usr/include/string.h:331:1: error: ambiguates old declaration ‘const char*
strstr(const char*, const char*)’
 strstr (const char *__haystack, const char *__needle) __THROW
 ^
In file included from /home/markus/gcc/gcc/genconstants.c:29:0:
/home/markus/gcc/gcc/system.h:499:34: error: declaration of C function ‘const
char* strsignal(int)’ conflicts with
 extern const char *strsignal (int);
                                  ^
In file included from
/usr/lib/gcc/x86_64-pc-linux-gnu/4.8.0/include/g++-v4/cstring:44:0,
                 from /home/markus/gcc/gcc/system.h:207,
                 from /home/markus/gcc/gcc/genconstants.c:29:
/usr/include/string.h:562:14: error: previous declaration ‘char*
strsignal(int)’ here
 extern char *strsignal (int __sig) __THROW;
              ^
In file included from /home/markus/gcc/gcc/system.h:639:0,
                 from /home/markus/gcc/gcc/genconstants.c:29:
/home/markus/gcc/gcc/../include/libiberty.h:110:36: error: new declaration
‘char* basename(const char*)’
 extern char *basename (const char *);
                                    ^
In file included from
/usr/lib/gcc/x86_64-pc-linux-gnu/4.8.0/include/g++-v4/cstring:44:0,
                 from /home/markus/gcc/gcc/system.h:207,
                 from /home/markus/gcc/gcc/genconstants.c:29:
/usr/include/string.h:599:26: error: ambiguates old declaration ‘const char*
basename(const char*)’
 extern "C++" const char *basename (const char *__filename)
                          ^
make[1]: *** [build/genconstants.o] Error 1
make[1]: Leaving directory `/var/tmp/gcc_build_dir/gcc'

[Bug other/54692] [4.8 Regression] gcc doesn't build with "-Og -g"

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.8.0

--- Comment #1 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-09-24 13:54:07 UTC ---
ISTR it worked for me when checking in -Og support with
BOOT_CFLAGS/BOOT_CXXFLAGS="-Og -g".  I don't have a host compiler with -Og
support around, but I wonder how -Og can make a difference for preprocessing?
Is it some configure tests giving different results?

[Bug other/54692] [4.8 Regression] gcc doesn't build with "-Og -g"

[markus at trippelsdorf dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

--- Comment #2 from Markus Trippelsdorf <markus at trippelsdorf dot de> 2012-09-24 14:07:06 UTC ---
(In reply to comment #1)
> ISTR it worked for me when checking in -Og support with
> BOOT_CFLAGS/BOOT_CXXFLAGS="-Og -g".  I don't have a host compiler with -Og
> support around, but I wonder how -Og can make a difference for preprocessing?
> Is it some configure tests giving different results?

Yes (from gcc/config.log):

configure:5224: checking for ANSI C header files
configure:5244: gcc -c -g g  conftest.c >&5
gcc: error: g: No such file or directory

and many more similar ones.

[Bug other/54692] [4.8 Regression] gcc doesn't build with "-Og -g"

[markus at trippelsdorf dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

--- Comment #3 from Markus Trippelsdorf <markus at trippelsdorf dot de> 2012-09-24 14:21:05 UTC ---
The following lines from gcc/configure don't know about -Og:

 4862 # Remove the -O2: for historical reasons, unless bootstrapping we prefer
 4863 # optimizations to be activated explicitly by the toplevel.
 4864 case "$CC" in
 4865   */prev-gcc/xgcc*) ;;
 4866   *) CFLAGS=`echo $CFLAGS | sed "s/-O[s0-9]* *//" `
 4867      CXXFLAGS=`echo $CXXFLAGS | sed "s/-O[s0-9]* *//" ` ;;
 4868 esac

When I comment them out, gcc builds fine.

[Bug other/54692] [4.8 Regression] gcc doesn't build with "-Og -g"

[markus at trippelsdorf dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

Markus Trippelsdorf <markus at trippelsdorf dot de> changed:

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

--- Comment #4 from Markus Trippelsdorf <markus at trippelsdorf dot de> 2012-09-24 14:32:16 UTC ---
Looks like Jakuk is responsible:

r191267
commit 56d581e9cde5c36d15b6d859abf6b2ea99f64ea0
Author: jakub <jakub@138bc75d-0d04-0410-961f-82ee72b054a4>
Date:   Thu Sep 13 16:30:17 2012 +0000

        * configure.ac (CXXFLAGS): Remove -O2 when not bootstrapping.
        * configure: Regenerated.

[Bug other/54692] [4.8 Regression] gcc doesn't build with "-Og -g"

[mpolacek at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

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

--- Comment #5 from Marek Polacek <mpolacek at gcc dot gnu.org> 2012-09-24 14:36:24 UTC ---
Does it work with "s/-O[sg0-9]* *//"?

[Bug other/54692] [4.8 Regression] gcc doesn't build with "-Og -g"

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> 2012-09-24 14:39:30 UTC ---
Guess
  *) CFLAGS=`echo $CFLAGS | sed "s/-O[[s0-9]]* *//" `
     CXXFLAGS=`echo $CXXFLAGS | sed "s/-O[[s0-9]]* *//" ` ;;
needs to be now
-O[[s0-9gf]] instead (also for -Ofast).
That said, I don't see how it is related to using STAGE1_CFLAGS (note missing
XX).

[Bug other/54692] [4.8 Regression] gcc doesn't build with "-Og -g"

[markus at trippelsdorf dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

--- Comment #7 from Markus Trippelsdorf <markus at trippelsdorf dot de> 2012-09-24 14:40:50 UTC ---
(In reply to comment #5)
> Does it work with "s/-O[sg0-9]* *//"?

Yes:
tmp % echo "-Og -g"| sed "s/-O[sg0-9]* *//"
-g
tmp %

[Bug other/54692] [4.8 Regression] gcc doesn't build with "-Og -g"

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

--- Comment #8 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-09-24 14:49:55 UTC ---
(In reply to comment #6)
> Guess
>   *) CFLAGS=`echo $CFLAGS | sed "s/-O[[s0-9]]* *//" `
>      CXXFLAGS=`echo $CXXFLAGS | sed "s/-O[[s0-9]]* *//" ` ;;
> needs to be now
> -O[[s0-9gf]] instead (also for -Ofast).
> That said, I don't see how it is related to using STAGE1_CFLAGS (note missing
> XX).

I wonder why we do the above at all?  I suppose that's for removing
a configure default, but the toplevel passes STAGE1_CFLAGS as CFLAGS to
gcc configure (that's why we need to re-specify CFLAGS on the make
command-line?!).

So - why not drop this and instead save/restore flags around AC_PROG_CC/CXX?

[Bug other/54692] [4.8 Regression] gcc doesn't build with "-Og -g"

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

--- Comment #9 from Jakub Jelinek <jakub at gcc dot gnu.org> 2012-09-24 15:04:06 UTC ---
(In reply to comment #8)
> (In reply to comment #6)
> > Guess
> >   *) CFLAGS=`echo $CFLAGS | sed "s/-O[[s0-9]]* *//" `
> >      CXXFLAGS=`echo $CXXFLAGS | sed "s/-O[[s0-9]]* *//" ` ;;
> > needs to be now
> > -O[[s0-9gf]] instead (also for -Ofast).
> > That said, I don't see how it is related to using STAGE1_CFLAGS (note missing
> > XX).
> 
> I wonder why we do the above at all?  I suppose that's for removing
> a configure default, but the toplevel passes STAGE1_CFLAGS as CFLAGS to
> gcc configure (that's why we need to re-specify CFLAGS on the make
> command-line?!).

The intent of this is to make sure that the toplevel Makefile has whatever
fancy
CXXFLAGS/CFLAGS is needed for bootstrapping, and gcc/Makefile has corresponding
CXXFLAGS/CFLAGS without -O2 or similar in it.  Thus, if in --disable-bootstrap
(or cross) gcc you do make in toplevel, you are building an optimized compiler,
while cd gcc; make after you tweak stuff here and there will default to no
optimization and thus hopefully better debugging experience.  If/when -Og is
better than -O0 for debug experience surely we can use there -Og instead.
>From toplevel make just passes down CXXFLAGS/CFLAGS, so the values stored in
gcc/Makefile are ignored.

[Bug other/54692] [4.8 Regression] gcc doesn't build with "-Og -g"

[markus at trippelsdorf dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

--- Comment #10 from Markus Trippelsdorf <markus at trippelsdorf dot de> 2012-09-24 15:08:27 UTC ---
(In reply to comment #6)

> -O[[s0-9gf]] instead (also for -Ofast).

"s/-O\([sg0-9]\|fast\) *//" should work.

[Bug other/54692] gcc doesn't build with "-Og -g"

[steven at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

Steven Bosscher <steven at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[4.8 Regression] gcc        |gcc doesn't build with "-Og
                   |doesn't build with "-Og -g" |-g"

--- Comment #11 from Steven Bosscher <steven at gcc dot gnu.org> 2012-09-24 18:54:55 UTC ---
Not a regression, -Og is new.

[Bug other/54692] gcc doesn't build with "-Og -g"

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.8.0                       |---

--- Comment #12 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-09-25 07:27:15 UTC ---
(In reply to comment #9)
> (In reply to comment #8)
> > (In reply to comment #6)
> > > Guess
> > >   *) CFLAGS=`echo $CFLAGS | sed "s/-O[[s0-9]]* *//" `
> > >      CXXFLAGS=`echo $CXXFLAGS | sed "s/-O[[s0-9]]* *//" ` ;;
> > > needs to be now
> > > -O[[s0-9gf]] instead (also for -Ofast).
> > > That said, I don't see how it is related to using STAGE1_CFLAGS (note missing
> > > XX).
> > 
> > I wonder why we do the above at all?  I suppose that's for removing
> > a configure default, but the toplevel passes STAGE1_CFLAGS as CFLAGS to
> > gcc configure (that's why we need to re-specify CFLAGS on the make
> > command-line?!).
> 
> The intent of this is to make sure that the toplevel Makefile has whatever
> fancy
> CXXFLAGS/CFLAGS is needed for bootstrapping, and gcc/Makefile has corresponding
> CXXFLAGS/CFLAGS without -O2 or similar in it.  Thus, if in --disable-bootstrap
> (or cross) gcc you do make in toplevel, you are building an optimized compiler,
> while cd gcc; make after you tweak stuff here and there will default to no
> optimization and thus hopefully better debugging experience.  If/when -Og is
> better than -O0 for debug experience surely we can use there -Og instead.
> From toplevel make just passes down CXXFLAGS/CFLAGS, so the values stored in
> gcc/Makefile are ignored.

Hm.  Doesn't make much sense to be - but anyway.

[Bug other/54692] gcc doesn't build with "-Og -g"

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

--- Comment #13 from Jakub Jelinek <jakub at gcc dot gnu.org> 2012-09-25 12:27:53 UTC ---
Author: jakub
Date: Tue Sep 25 12:27:47 2012
New Revision: 191702

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=191702
Log:
    PR other/54692
    * configure.ac (CFLAGS, CXXFLAGS): Remove -Ofast or -Og
    properly.
    * configure: Regenerated.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/configure
    trunk/gcc/configure.ac

[Bug other/54692] gcc doesn't build with "-Og -g"

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54692

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |FIXED
         AssignedTo|unassigned at gcc dot       |jakub at gcc dot gnu.org
                   |gnu.org                     |
   Target Milestone|---                         |4.8.0

--- Comment #14 from Jakub Jelinek <jakub at gcc dot gnu.org> 2012-09-25 12:49:57 UTC ---
Should be fixed now.