NetBSD Problem Report #15514

Received: (qmail 1410 invoked from network); 6 Feb 2002 22:08:51 -0000
Message-Id: <20020206220837.D7C252DA@jocelyn.sub-etha.wtal.de>
Date: Wed,  6 Feb 2002 23:08:37 +0100 (CET)
From: is@netbsd.org
Reply-To: is@netbsd.org
To: gnats-bugs@gnats.netbsd.org
Subject: i386->m68k doesn't cross-build
X-Send-Pr-Version: 3.95

>Number:         15514
>Category:       port-m68k
>Synopsis:       i386->m68k doesn't cross-build
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    thorpej
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Wed Feb 06 22:09:00 +0000 2002
>Closed-Date:    Sun Jun 02 15:55:46 +0000 2002
>Last-Modified:  Sun Jun 02 15:55:46 +0000 2002
>Originator:     Ignatios Souvatzis
>Release:        2002 Feb 1
>Organization:
	dis-
>Environment:

	Host: i386 netbsd-1.5ZA
	Target: amiga /m68k
>Description:
	changing share/mk/bsd.own.mk to make Amiga use the new toolchain,
	and build.sh -m amiga on a NetBSD/i386 (built end January), fails
	when trying to build the native compiler.
	logfile: http://www.kleinbus.org/~is/logfile.excerpt
>How-To-Repeat:
	build.sh -m amiga -R ... -D ... -T ... -O ...
>Fix:
	workaround: manipulate gnu/usr.bin/gcc/arch/m68k/tm.h 
	to contain the apropriate #define TARGET_CPU_DEFAULT
>Release-Note:
>Audit-Trail:
State-Changed-From-To: open->analyzed 
State-Changed-By: scw 
State-Changed-When: Wed Feb 6 22:29:59 GMT 2002 
State-Changed-Why:  

I will be commiting the results of re-running toolchain2netbsd on a 
new-toolchain m68k machine soon, once I've completed some testing on it. 


From: Frederick Bruckman <fredb@immanent.net>
To: gnats-bugs@gnats.NetBSD.ORG
Cc: scw@netbsd.org, <is@netbsd.org>
Subject: Re: port-m68k/15514: i386->m68k doesn't cross-build
Date: Wed, 6 Feb 2002 23:56:18 -0600 (CST)

 	State-Changed-From-To: open->analyzed
 	State-Changed-By: scw
 	State-Changed-When: Wed Feb 6 22:29:59 GMT 2002
 	State-Changed-Why:


 	I will be commiting the results of re-running toolchain2netbsd on a
 	new-toolchain m68k machine soon, once I've completed some testing on
 	it.

 I ran toolchain2netbsd on 1.5ZA/mac68k/ELF, and the build now dies for me in
 libbfd:

 /t/tools/bin/m68k--netbsdelf-gcc -O2   -Werror  -I/s/src-current/gnu/lib/libbfd/arch/m68k -I.  -DARCH_m68k -DHAVE_CONFIG_H -D_GNU_SOURCE -DNETBSD_CORE -DDEFAULT_VECTOR=bfd_elf32_m68k_vec -DSELECT_VECS='&bfd_elf32_m68k_vec,&m68knetbsd_vec,&m68k4knetbsd_vec,&hp300bsd_vec,&sunos_big_vec,&bfd_elf32_little_generic_vec,&bfd_elf32_big_generic_vec' -DSELECT_ARCHITECTURES='&bfd_m68k_arch' -DHAVE_bfd_elf32_m68k_vec -DHAVE_m68knetbsd_vec -DHAVE_m68k4knetbsd_vec -DHAVE_hp300bsd_vec -DHAVE_sunos_big_vec -DHAVE_bfd_elf3 2_little_generic_vec -DHAVE_bfd_elf32_big_generic_vec -I/s/src-current/gnu/lib/libbfd/../../dist/toolchain/bfd -I/s/src-current/gnu/lib/libbfd/../../dist/toolchain/bfd -I/s/src-current/gnu/lib/libbfd/../../dist/toolchain/bfd/../include -I/s/src-current/gnu/lib/libbfd/../../dist/toolchain/bfd/../intl -nostdinc -isystem /t/current-mac68k/usr/include  -c -fPIC -DPIC /s/src-current/gnu/lib/libbfd/../../dist/toolchain/bfd/aout32.c -o aout32.so
 In file included from /s/src-current/gnu/lib/libbfd/../../dist/toolchain/bfd/sysdep.h:27,
                  from /s/src-current/gnu/lib/libbfd/../../dist/toolchain/bfd/aoutx.h:125,
                  from /s/src-current/gnu/lib/libbfd/../../dist/toolchain/bfd/aout32.c:23:
 /s/src-current/gnu/lib/libbfd/arch/m68k/config.h:390: warning: `VERSION' redefined
 /s/src-current/gnu/lib/libbfd/arch/m68k/config.h:183: warning: this is the location of the previous definition
 *** Error code 1

 Stop.
 nbmake: stopped in /s/src-current/gnu/lib/libbfd
 *** Error code 1

 Stop.

 Here's a patch to toolchain2netbsd that lets me get past this:

 Index: toolchain2netbsd
 ===================================================================
 RCS file: /cvsroot/gnusrc/gnu/toolchain2netbsd,v
 retrieving revision 1.14
 diff -u -r1.14 toolchain2netbsd
 --- toolchain2netbsd	2002/01/22 13:13:00	1.14
 +++ toolchain2netbsd	2002/02/07 05:51:24
 @@ -107,7 +107,7 @@

  {
  	cat $_TMPDIR/bfd/config.h
 -	grep -v PACKAGE $_TMPDIR/opcodes/config.h
 +	fgrep -v "$(cat $_TMPDIR/bfd/config.h)" $_TMPDIR/opcodes/config.h
  } | write_c lib/libbfd/arch/$MACHINE_ARCH/config.h

  ##### gnu/lib/libf2c #####


 There aren't actually (currently) any definitions in opcodes/config.h
 that aren't also in bfd/config.h, but fgrep should catch them if there
 ever are.

 Frederick


From: Matthias Drochner <M.Drochner@fz-juelich.de>
To: is@jocelyn.sub-etha.wtal.de
Cc: gnats-bugs@gnats.netbsd.org
Subject: Re: port-m68k/15514: i386->m68k doesn't cross-build 
Date: Thu, 07 Feb 2002 14:06:31 +0100

 It didn't even build natively without the definition of
 TARGET_CPU_DEFAULT for me.
 I've rerun toolchain2netbsd; the only substancial changes
 are appended. After that, I've sucessfully rebuilt the
 world. (which tool a while as you can see from the file
 dates below...)
 And when I take back the change after that, gcc doesn't
 build again, so it is not just a bootstrap issue.
 So I'd call the change to tm.h necessary and tested...
 The second - G_PROTO_OBJS is now empty - works too.
 "protoize" and "unprotoize" still build, so the 3
 additional files are really not necessary. Whether these
 tools are useful at all whichout SYSCALLS.c.X is another
 question...
 I haven't found a place where G_HOST_LIBS is needed.

 best regards
 Matthias


 Index: tm.h
 ===================================================================
 RCS file: /cvsroot/gnusrc/gnu/usr.bin/gcc/arch/m68k/tm.h,v
 retrieving revision 1.3
 diff -c -r1.3 tm.h
 *** tm.h	2001/11/05 19:11:01	1.3
 --- tm.h	2002/02/05 10:20:32
 ***************
 *** 1,6 ****
   /* This file is automatically generated.  DO NOT EDIT! */
 ! /* Generated from: 	NetBSD: toolchain2netbsd,v 1.12 2001/08/14 05:17:59 tv 
 Exp  */

   #ifdef IN_GCC
   #include "gansidecl.h"
   #endif
 --- 1,7 ----
   /* This file is automatically generated.  DO NOT EDIT! */
 ! /* Generated from: 	NetBSD: toolchain2netbsd,v 1.14 2002/01/22 13:13:00 mrg 
 Exp  */

 + #define TARGET_CPU_DEFAULT (MASK_68020|MASK_68881|MASK_BITFIELD)
   #ifdef IN_GCC
   #include "gansidecl.h"
   #endif


 Index: defs.mk
 ===================================================================
 RCS file: /cvsroot/gnusrc/gnu/usr.bin/gcc/arch/m68k/defs.mk,v
 retrieving revision 1.2
 diff -c -r1.2 defs.mk
 *** defs.mk	2001/12/19 23:54:27	1.2
 --- defs.mk	2002/02/05 10:28:54
 ***************
 *** 1,5 ****
   # This file is automatically generated.  DO NOT EDIT!
 ! # Generated from: 	NetBSD: toolchain2netbsd,v 1.12 2001/08/14 05:17:59 tv 
 Exp
   #
   G_ALL_CFLAGS= -DIN_GCC    -g   -DHAVE_CONFIG_H
   G_ALL_CPPFLAGS=  
 --- 1,5 ----
   # This file is automatically generated.  DO NOT EDIT!
 ! # Generated from: 	NetBSD: toolchain2netbsd,v 1.14 2002/01/22 13:13:00 mrg 
 Exp
   #
   G_ALL_CFLAGS= -DIN_GCC    -g   -DHAVE_CONFIG_H
   G_ALL_CPPFLAGS=  
 ***************
 *** 7,13 ****
   G_C_OBJS=c-parse.o c-lang.o c-lex.o c-pragma.o c-decl.o c-typeck.o 
 c-convert.o  c-aux-info.o c-common.o c-iterate.o
   G_CCCP_OBJS=cccp.o cexp.o intl.o prefix.o version.o  mbchar.o
   G_GCOV_OBJS=gcov.o intl.o
 ! G_PROTO_OBJS=getpwd.o intl.o version.o
   G_HOST_PRINT=print-rtl.o
   G_HOST_RTL=rtl.o bitmap.o
   G_HOST_RTLANAL=rtlanal.o
 --- 7,13 ----
   G_C_OBJS=c-parse.o c-lang.o c-lex.o c-pragma.o c-decl.o c-typeck.o 
 c-convert.o  c-aux-info.o c-common.o c-iterate.o
   G_CCCP_OBJS=cccp.o cexp.o intl.o prefix.o version.o  mbchar.o
   G_GCOV_OBJS=gcov.o intl.o
 ! G_HOST_LIBS=` case "obstack.o" in ?*) echo obstack.o ;; esac ` ` case "" in 
 ?*) echo  ;; esac ` ` case "" in ?*) echo  ;; esac `    ` case "" in ?*) echo  
 ;; esac ` ` case "" in ?*) echo  ;; esac `
   G_HOST_PRINT=print-rtl.o
   G_HOST_RTL=rtl.o bitmap.o
   G_HOST_RTLANAL=rtlanal.o



From: Frederick Bruckman <fredb@immanent.net>
To: Matthias Drochner <M.Drochner@fz-juelich.de>
Cc: is@jocelyn.sub-etha.wtal.de, <gnats-bugs@gnats.netbsd.org>
Subject: Re: port-m68k/15514: i386->m68k doesn't cross-build
Date: Sat, 9 Feb 2002 09:00:31 -0600 (CST)

 From: Matthias Drochner <M.Drochner@fz-juelich.de>
 To: is@jocelyn.sub-etha.wtal.de
 Cc: gnats-bugs@gnats.netbsd.org
 Subject: Re: port-m68k/15514: i386->m68k doesn't cross-build
 Date: Thu, 07 Feb 2002 14:06:31 +0100

  It didn't even build natively without the definition of
  TARGET_CPU_DEFAULT for me.

 It had built for me, but it without the explicit default, it was
 generating soft-float function calls for extended float, which we don't
 have functions for.

  I've rerun toolchain2netbsd; the only substancial changes
  are appended. After that, I've sucessfully rebuilt the
  world. (which tool a while as you can see from the file
  dates below...)

 When I tried it, with the cross-built from i386 compiler, besides the
 duplicates in libbfd/arch/m68k/config.h (noted in the PR already),
 m68k--netbsdelf was changed everywhere to m68k-unknown-netbsdelf.
 Which is correct?

 Frederick


From: Steve Woodford <scw@netbsd.org>
To: Frederick Bruckman <fredb@immanent.net>
Cc: port-m68k-maintainer@netbsd.org, <gnats-bugs@gnats.netbsd.org>
Subject: Re: port-m68k/15514: i386->m68k doesn't cross-build
Date: Sat, 9 Feb 2002 16:19:33 +0000 (GMT)

 On 9 Feb 2002, Frederick Bruckman wrote:

 >  When I tried it, with the cross-built from i386 compiler, besides the
 >  duplicates in libbfd/arch/m68k/config.h (noted in the PR already),
 >  m68k--netbsdelf was changed everywhere to m68k-unknown-netbsdelf.
 >  Which is correct?

 I'd say the latter. I'm a bit behind in the commit for this bug, which
 will include the patch to toolchain2netbsd to prevent duplicate #defines.

 The plan is to commit it tomorrow. I'm still waiting for a native m68k
 build to finish and prove functional, as a test. (Call me Paranoid. ;-)

 Cheers, Steve



From: Jason R Thorpe <thorpej@wasabisystems.com>
To: Steve Woodford <scw@netbsd.org>
Cc: Frederick Bruckman <fredb@immanent.net>,
	port-m68k-maintainer@netbsd.org, gnats-bugs@gnats.netbsd.org
Subject: Re: port-m68k/15514: i386->m68k doesn't cross-build
Date: Sat, 9 Feb 2002 08:39:39 -0800

 On Sat, Feb 09, 2002 at 04:19:33PM +0000, Steve Woodford wrote:

  > >  When I tried it, with the cross-built from i386 compiler, besides the
  > >  duplicates in libbfd/arch/m68k/config.h (noted in the PR already),
  > >  m68k--netbsdelf was changed everywhere to m68k-unknown-netbsdelf.
  > >  Which is correct?
  > 
  > I'd say the latter. I'm a bit behind in the commit for this bug, which
  > will include the patch to toolchain2netbsd to prevent duplicate #defines.

 Err.  The latter is what the GNU tools use by default, but we've been
 using "m68k--netbsdelf" form for a while now in the in-tree tools as
 a nice short-hand.

 Actually, we probably should use "m68k-netbsdelf", since I believe the
 GNU configure scripts do deal properly with targets named that way.

  > 
  > The plan is to commit it tomorrow. I'm still waiting for a native m68k
  > build to finish and prove functional, as a test. (Call me Paranoid. ;-)
  > 
  > Cheers, Steve
  > 
  > 

 -- 
         -- Jason R. Thorpe <thorpej@wasabisystems.com>

From: Frederick Bruckman <fredb@immanent.net>
To: Jason R Thorpe <thorpej@wasabisystems.com>
Cc: Steve Woodford <scw@netbsd.org>, <port-m68k-maintainer@netbsd.org>,
   <gnats-bugs@gnats.netbsd.org>
Subject: Re: port-m68k/15514: i386->m68k doesn't cross-build
Date: Sat, 9 Feb 2002 11:23:04 -0600 (CST)

 On Sat, 9 Feb 2002, Jason R Thorpe wrote:

 > On Sat, Feb 09, 2002 at 04:19:33PM +0000, Steve Woodford wrote:
 >
 >  > >  When I tried it, with the cross-built from i386 compiler, besides the
 >  > >  duplicates in libbfd/arch/m68k/config.h (noted in the PR already),
 >  > >  m68k--netbsdelf was changed everywhere to m68k-unknown-netbsdelf.
 >  > >  Which is correct?
 >  >
 >  > I'd say the latter. I'm a bit behind in the commit for this bug, which
 >  > will include the patch to toolchain2netbsd to prevent duplicate #defines.
 >
 > Err.  The latter is what the GNU tools use by default, but we've been
 > using "m68k--netbsdelf" form for a while now in the in-tree tools as
 > a nice short-hand.

 Well, OK. That looks nicer. Seeing "unknown" makes you think you did
 something wrong. If that's the case, we should fix it in config.guess,
 and be consistent for other ports, too. I haven't actually run
 toolchain2netbsd with this patch, as I'm still in the middle of a build,
 too, but I think this will do it:

 Index: config.guess
 ===================================================================
 RCS file: /cvsroot/gnusrc/gnu/dist/toolchain/config.guess,v
 retrieving revision 1.10
 diff -u -r1.10 config.guess
 --- config.guess	2002/01/29 18:52:09	1.10
 +++ config.guess	2002/02/09 17:06:43
 @@ -136,14 +136,14 @@
  	# object file format.
  	#
  	# Note: NetBSD doesn't particularly care about the vendor
 -	# portion of the name.  We always set it to "unknown".
 +	# portion of the name.  We leave it empty.
  	UNAME_MACHINE_ARCH=`(uname -p) 2>/dev/null` || \
  	    UNAME_MACHINE_ARCH=unknown
  	case "${UNAME_MACHINE_ARCH}" in
 -	    arm*) machine=arm-unknown ;;
 -	    sh3el) machine=shle-unknown ;;
 -	    sh3eb) machine=sh-unknown ;;
 -	    *) machine=${UNAME_MACHINE_ARCH}-unknown ;;
 +	    arm*) machine=arm- ;;
 +	    sh3el) machine=shle- ;;
 +	    sh3eb) machine=sh- ;;
 +	    *) machine=${UNAME_MACHINE_ARCH}- ;;
  	esac
  	# The Operating System including object format, if it has switched
  	# to ELF recently, or will in the future.

 > Actually, we probably should use "m68k-netbsdelf", since I believe the
 > GNU configure scripts do deal properly with targets named that way.

 Old ones too? "pkgsrc" currently passes --host=${MACHINE_GNU_PLATFORM}
 to every configure script run, and

 fredb@tautology-> make show-var VARNAME=MACHINE_GNU_PLATFORM
 i386--netbsd

 (that should be i386--netbsdelf? I'll have to look over the cvs log to
 see if there's a reason that it isn't.) but

 root@deduction-> make show-var VARNAME=MACHINE_GNU_PLATFORM
 m68k--netbsdelf

 It would probably be best to get everyone on the same page with
 `uname -p`--netbsd, or "`uname -p`"--netbsdelf for platforms that
 started out a.out, as that's less likely to break things in pkgsrc than
 `uname -p`-netbsd{,elf}.

 Frederick


From: Jason R Thorpe <thorpej@wasabisystems.com>
To: Frederick Bruckman <fredb@immanent.net>
Cc: Steve Woodford <scw@netbsd.org>, port-m68k-maintainer@netbsd.org,
	gnats-bugs@gnats.netbsd.org
Subject: Re: port-m68k/15514: i386->m68k doesn't cross-build
Date: Sat, 9 Feb 2002 10:37:03 -0800

 On Sat, Feb 09, 2002 at 11:23:04AM -0600, Frederick Bruckman wrote:

  > Well, OK. That looks nicer. Seeing "unknown" makes you think you did
  > something wrong. If that's the case, we should fix it in config.guess,
  > and be consistent for other ports, too. I haven't actually run
  > toolchain2netbsd with this patch, as I'm still in the middle of a build,
  > too, but I think this will do it:

 This is totally unnecessary.  toolchain2netbsd should be passing the
 host and target explicitly.

 -- 
         -- Jason R. Thorpe <thorpej@wasabisystems.com>

From: Frederick Bruckman <fredb@immanent.net>
To: Jason R Thorpe <thorpej@wasabisystems.com>
Cc: Steve Woodford <scw@netbsd.org>, <port-m68k-maintainer@netbsd.org>,
   <gnats-bugs@gnats.netbsd.org>
Subject: Re: port-m68k/15514: i386->m68k doesn't cross-build
Date: Sat, 9 Feb 2002 21:18:28 -0600 (CST)

 On Sat, 9 Feb 2002, Jason R Thorpe wrote:

 > On Sat, Feb 09, 2002 at 11:23:04AM -0600, Frederick Bruckman wrote:
 >
 >  > Well, OK. That looks nicer. Seeing "unknown" makes you think you did
 >  > something wrong. If that's the case, we should fix it in config.guess,
 >  > and be consistent for other ports, too. I haven't actually run
 >  > toolchain2netbsd with this patch, as I'm still in the middle of a build,
 >  > too, but I think this will do it:
 >
 > This is totally unnecessary.  toolchain2netbsd should be passing the
 > host and target explicitly.

 I see the wisdom of that. "config.guess" comes from autoconf, so it's
 not likely that the gcc maintainers would want to maintain a custom
 patch to it.

 Here's a patch to toolchain2netbsd to obtain MACHINE_GNU_PLATFORM from
 bsd.own.mk, and to use it to pass the appropriate --host arguments to
 configure. I also identified one other place where there's a potential
 for config.h's to get duplicate defines, and gave it the fgrep
 treatment. I also changed the "fgrep" command(s)  to use `` instead of
 $(), to be consistent with the style of the rest of the file.

 This is not been tested at all, but it looks simple enough, that it
 might actually work without change. I'm still waiting for the first
 native build to finish on my mac.

 Index: toolchain2netbsd
 ===================================================================
 RCS file: /cvsroot/gnusrc/gnu/toolchain2netbsd,v
 retrieving revision 1.14
 diff -u -r1.14 toolchain2netbsd
 --- toolchain2netbsd	2002/01/22 13:13:00	1.14
 +++ toolchain2netbsd	2002/02/10 01:51:17
 @@ -8,6 +8,13 @@
  TMPDIR=${TMPDIR:-/tmp}
  MAKE=${MAKE:-make}

 +$MAKE -f - _x_ <<EOF | readline MACHINE_GNU_PLATFORM
 +_x_:
 +	echo ${MACHINE_GNU_PLATFORM}
 +
 +.include <bsd.own.mk>
 +EOF
 +
  # usage: getvars MAKEFILE VARNAME [VARNAME...]
  getvars () {
  	_mf="$1"; shift
 @@ -59,7 +66,7 @@
  if [ ! -f $_TMPDIR/gprof/Makefile ]; then
  	(mkdir -p $_TMPDIR && cd $_TMPDIR && \
  		LDFLAGS=-lintl LANGUAGES="c c++ objc f77" \
 -		$cwd/dist/toolchain/configure --disable-shared)
 +		$cwd/dist/toolchain/configure --disable-shared --host=$MACHINE_GNU_PLATFORM)
  fi

  # Library configuration, for after g++ or g77 are installed.
 @@ -72,9 +79,9 @@
  			(mkdir -p $_TMPDIR/$dir && cd $_TMPDIR/$dir && \
  			export CC="gcc -idirafter $_TMPDIR -idirafter $cwd/usr.bin/gcc/arch/$MACHINE_ARCH" && \
  			if [ -f $cwd/dist/toolchain/$dir/configure ]; then
 -				$cwd/dist/toolchain/$dir/configure --disable-shared
 +				$cwd/dist/toolchain/$dir/configure --disable-shared --host=$MACHINE_GNU_PLATFORM
  			else
 -				$cwd/dist/toolchain/configure --srcdir=$cwd/dist/toolchain/$dir --disable-shared
 +				$cwd/dist/toolchain/configure --srcdir=$cwd/dist/toolchain/$dir --disable-shared --host=$MACHINE_GNU_PLATFORM
  			fi)
  		fi
  	done
 @@ -107,7 +114,7 @@

  {
  	cat $_TMPDIR/bfd/config.h
 -	grep -v PACKAGE $_TMPDIR/opcodes/config.h
 +	fgrep -v "`cat $_TMPDIR/bfd/config.h`" $_TMPDIR/opcodes/config.h
  } | write_c lib/libbfd/arch/$MACHINE_ARCH/config.h

  ##### gnu/lib/libf2c #####
 @@ -273,8 +280,10 @@

  (cd $_TMPDIR/gdb && $MAKE init.c version.c)

 -cat $_TMPDIR/gdb/config.h $_TMPDIR/readline/config.h \
 -	| write_c usr.bin/gdb/arch/$MACHINE_ARCH/config.h
 +{
 +	cat $_TMPDIR/gdb/config.h
 +	fgrep -v "`cat $_TMPDIR/gdb/config.h`" $_TMPDIR/readline/config.h
 +} | write_c usr.bin/gdb/arch/$MACHINE_ARCH/config.h

  write_c usr.bin/gdb/arch/$MACHINE_ARCH/init.c <$_TMPDIR/gdb/init.c
  write_c usr.bin/gdb/arch/$MACHINE_ARCH/version.c <$_TMPDIR/gdb/version.c


 Frederick

Responsible-Changed-From-To: port-m68k-maintainer->thorpej 
Responsible-Changed-By: thorpej 
Responsible-Changed-When: Sun Feb 10 11:36:49 PST 2002 
Responsible-Changed-Why:  
The problem is likely my fault. 

From: Frederick Bruckman <fredb@immanent.net>
To: Jason R Thorpe <thorpej@wasabisystems.com>
Cc: Steve Woodford <scw@netbsd.org>, <port-m68k-maintainer@netbsd.org>,
   <gnats-bugs@gnats.netbsd.org>
Subject: Re: port-m68k/15514: i386->m68k doesn't cross-build
Date: Wed, 13 Feb 2002 15:51:44 -0600 (CST)

 On Sat, 9 Feb 2002, Frederick Bruckman wrote:

 > On Sat, 9 Feb 2002, Jason R Thorpe wrote:
 >
 > > This is totally unnecessary.  toolchain2netbsd should be passing the
 > > host and target explicitly.
 >
 > I see the wisdom of that. "config.guess" comes from autoconf, so it's
 > not likely that the gcc maintainers would want to maintain a custom
 > patch to it.
 >
 > Here's a patch to toolchain2netbsd to obtain MACHINE_GNU_PLATFORM from
 > bsd.own.mk, and to use it to pass the appropriate --host arguments to
 > configure.

 Yuck. That last was really broken. I have a better idea now, which I'll
 submit to tech-toolchain for review once I've tested it. For now, maybe
 we should just elide the duplicate definition of VERSION from
 .../libbfd/arch/m68k/config.h and close the PR?

 Frederick

State-Changed-From-To: analyzed->closed 
State-Changed-By: fredb 
State-Changed-When: Sun Jun 2 10:53:31 CDT 2002 
State-Changed-Why:  
This problem has been resolved -- I just built another 1.6_BETA1/mac68k 
snapshot from a 1.5.3/i386 host, to be sure. (There may be other problems 
with cross-builds -- see bin/17138 -- but no m68k problems, as such.) 
>Unformatted:

NetBSD Home
NetBSD PR Database Search

(Contact us) $NetBSD: query-full-pr,v 1.39 2013/11/01 18:47:49 spz Exp $
$NetBSD: gnats_config.sh,v 1.8 2006/05/07 09:23:38 tsutsui Exp $
Copyright © 1994-2007 The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.