NetBSD Problem Report #23240
Received: (qmail 19012 invoked by uid 605); 23 Oct 2003 07:52:20 -0000
Message-Id: <20031023075215.3123F2E4A1D@what-time-is-love.eterna.com.au>
Date: Thu, 23 Oct 2003 17:52:15 +1000 (EST)
From: mrg@eterna.com.au
Sender: gnats-bugs-owner@NetBSD.org
Reply-To: mrg@eterna.com.au
To: gnats-bugs@gnats.netbsd.org
Subject: mozilla/MozillaFirebird build failure on macppc-current
X-Send-Pr-Version: 3.95
>Number: 23240
>Category: pkg
>Synopsis: mozilla/MozillaFirebird build failure on macppc-current
>Confidential: no
>Severity: serious
>Priority: high
>Responsible: taya
>State: closed
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Thu Oct 23 07:53:00 +0000 2003
>Closed-Date: Sat Nov 20 06:57:09 +0000 2004
>Last-Modified: Sat Dec 04 02:18:01 +0000 2004
>Originator: matthew green
>Release: NetBSD 1.6ZD
>Organization:
people's front against (bozotic) www (softwar foundation)
>Environment:
System: NetBSD what-time-is-love.eterna.com.au 1.6ZD NetBSD 1.6ZD (_what_) #141: Sun Oct 19 21:03:51 EST 2003 mrg@phat-bass.eterna.com.au:/home/obj/macppc/usr/src/sys/arch/macppc/compile/_what_ macppc
Architecture: powerpc
Machine: macppc
>Description:
both mozilla and phoenix fail to build on my very recently upgraded
to -current macppc box. the mozilla/nsprpub/pr/src/io/prprf.c file
fails to build due to bad use of va_list. changing it to use va_copy(3)
on netbsd/powerpc makes this file build for both mozilla & firebird.
next problem - i don't know what is up here at all:
gmake[3]: Entering directory `/var/obj.pkgsrc/www/mozilla/work.what-time-is-love/mozilla/js/src/xpconnect/shell'
xpcshell.cpp
c++ -o xpcshell.o -c -DOSTYPE=\"NetBSD1\" -DOSARCH=\"NetBSD\" -DJS_THREADSAFE -I../../../../dist/include/xpconnect -I../../../../dist/include/xpcom -I../../../../dist/include/string -I../../../../dist/include/js -I../../../../dist/include/xpcshell -I../../../../dist/include -I/var/obj.pkgsrc/www/mozilla/work.what-time-is-love/mozilla/dist/include/nspr -I/usr/pkg/include -I/usr/pkg/include -I/usr/X11R6/include -fPIC -DPIC -I/usr/pkg/include/gdk-pixbuf-1.0 -I/usr/pkg/include -I/usr/X11R6/include -I/usr/X11R6/include -fno-rtti -fno-exceptions -Wall -Wconversion -Wpointer-arith -Wcast-align -Woverloaded-virtual -Wsynth -Wno-ctor-dtor-privacy -Wno-long-long -O2 -I/usr/pkg/include -I/usr/X11R6/include -Dunix -fshort-wchar -pthread -pipe -DNDEBUG -DTRIMMED -O2 -I/usr/pkg/include/gdk-pixbuf-1.0 -I/usr/pkg/include -I/usr/X11R6/include -I/usr/X11R6/include -DMOZILLA_CLIENT -include ../../../../mozilla-config.h -Wp,-MD,.deps/xpcshell.pp xpcshell.cpp
c++ -I/usr/pkg/include/gdk-pixbuf-1.0 -I/usr/pkg/include -I/usr/X11R6/include -I/usr/X11R6/include -fno-rtti -fno-exceptions -Wall -Wconversion -Wpointer-arith -Wcast-align -Woverloaded-virtual -Wsynth -Wno-ctor-dtor-privacy -Wno-long-long -O2 -I/usr/pkg/include -I/usr/X11R6/include -Dunix -fshort-wchar -pthread -pipe -DNDEBUG -DTRIMMED -O2 -o xpcshell xpcshell.o -Wl,-R/usr/pkg/lib/mozilla -L/usr/pkg/lib -Wl,-R/usr/pkg/lib -L/usr/X11R6/lib -Wl,-R/usr/X11R6/lib -L../../../../dist/bin -L../../../../dist/lib -L../../../../dist/bin -lmozjs -L../../../../dist/bin -lxpcom -L/var/obj.pkgsrc/www/mozilla/work.what-time-is-love/mozilla/dist/lib -lplds4 -lplc4 -lnspr4 -L/usr/pkg/lib -L/usr/X11R6/lib -pthread -lm -lintl -Wl,--export-dynamic
../../../../dist/bin/libxpcom.so: undefined reference to `nsXPTCStubBase::Stub203()'
../../../../dist/bin/libxpcom.so: undefined reference to `nsXPTCStubBase::Stub156()'
../../../../dist/bin/libxpcom.so: undefined reference to `nsXPTCStubBase::Stub184()'
../../../../dist/bin/libxpcom.so: undefined reference to `nsXPTCStubBase::Stub128()'
[ ... ]
../../../../dist/bin/libxpcom.so: undefined reference to `nsXPTCStubBase::Stub102()'
../../../../dist/bin/libxpcom.so: undefined reference to `nsXPTCStubBase::Stub229()'
gmake[3]: *** [xpcshell] Error 1
gmake[3]: Leaving directory `/var/obj.pkgsrc/www/mozilla/work.what-time-is-love/mozilla/js/src/xpconnect/shell'
(there are about 250 of these "StubN" functions missing.) the exactly same
problem occurs in mozilla & phoenix.
this is with GCC 3.3.2-nb1.
>How-To-Repeat:
build either of www/{mozilla,MozillaFirebird} on maccpc/current.
>Fix:
for prprf.c make it use va_copy(3) instead of either raw assignment or
structure assignment.
>Release-Note:
>Audit-Trail:
From: Jarkko Teppo <jarkko.teppo@er-grp.com>
To: gnats-bugs@netbsd.org
Cc:
Subject: Re: pkg/23240
Date: Mon, 10 Nov 2003 11:04:13 +0200
FYI, the missing nsXPTCStubBase::StubXXX-functions show up with alpha
(1.6ZD, ~latest pkgsrc) as well. I'm not sure if it's related with moving
from gcc-2.9 to gcc-3.
--
jht
From: Arto Huusko <armihu@utu.fi>
To: gnats-bugs@netbsd.org
Cc: netbsd-bugs@netbsd.org
Subject: Re: pkg/23240
Date: Thu, 04 Dec 2003 17:53:31 +0200 (EET)
Most likely this bug is caused by newer GCC.
The problem is that
mozilla/xpcom/reflect/src/xptcall/src/md/unix/xptcstubs_linux_alpha.cpp
defines STUB_ENTRY like this:
#define STUB_ENTRY(n) \
__asm__( \
"#### Stub"#n" ####\n" \
".text\n\t" \
".align 5\n\t" \
".globl Stub"#n"__14nsXPTCStubBase\n\t" \
".ent Stub"#n"__14nsXPTCStubBase\n" \
"Stub"#n"__14nsXPTCStubBase:\n\t" \
".frame $30,0,$26,0\n\t" \
"ldgp $29,0($27)\n" \
"$Stub"#n"__14nsXPTCStubBase..ng:\n\t" \
".prologue 1\n\t" \
"lda $1,"#n"\n\t" \
"br $31,$SharedStub..ng\n\t" \
".end Stub"#n"__14nsXPTCStubBase" \
);
However, newer GCC apparently mangles C++ names differently.
Running nm on libxptcall.a shows that it expects symbols like
_ZN14nsXPTCStubBase5Stub3Ev
_ZN14nsXPTCStubBase6Stub10Ev
_ZN14nsXPTCStubBase7Stub100Ev
So, the solution to the problem is to modify the STUBS_ENTRY macro.
Since I'm no CPP guru I fixed it like this:
191c191
< #define STUB_ENTRY(n) \
---
> #define STUB_ENTRY(n, m) \
196,198c196,198
< ".globl Stub"#n"__14nsXPTCStubBase\n\t" \
< ".ent Stub"#n"__14nsXPTCStubBase\n" \
< "Stub"#n"__14nsXPTCStubBase:\n\t" \
---
> ".globl _ZN14nsXPTCStubBase"#m"Stub"#n"Ev\n\t" \
> ".ent _ZN14nsXPTCStubBase"#m"Stub"#n"Ev\n" \
> "_ZN14nsXPTCStubBase"#m"Stub"#n"Ev:\n\t" \
201c201
< "$Stub"#n"__14nsXPTCStubBase..ng:\n\t" \
---
> "$_ZN14nsXPTCStubBase"#m"Stub"#n"Ev..ng:\n\t" \
205c205
< ".end Stub"#n"__14nsXPTCStubBase" \
---
> ".end _ZN14nsXPTCStubBase"#m"Stub"#n"Ev" \
I couldn't figure out how to extract length of n in CPP, so I added a new
parameter m. This also requires patching
mozilla/xpcom/reflect/xptcall/public/xptcstubsdef.inc
to add the lengths (5, 6 and 7) to all the STUBS_ENTRY calls.
Alternative is, if it's possible, to make STUBS_ENTRY define
an actual C++ function, like STUBS_SENTINEL.
With this, the build is now at least past the point where it last bombed.
-- Arto
State-Changed-From-To: open->feedback
State-Changed-By: grant
State-Changed-When: Sat Dec 20 08:49:18 UTC 2003
State-Changed-Why:
could you please confirm whether this problem is fixed or still exists
in MozillaFirebird 0.7.
State-Changed-From-To: feedback->open
State-Changed-By: mrg
State-Changed-When: Mon Jan 5 05:10:41 UTC 2004
State-Changed-Why:
this problem still exists. the original compile-time error
is still there, the meantioned fix for Stubxxx may be enough tough.
From: matthew green <mrg@eterna.com.au>
To: gnats-bugs@netbsd.org
Cc:
Subject: Re: pkg/23240
Date: Sun, 04 Apr 2004 17:05:50 +1000
BTW:
the prprf.c problem still exists today for both mozilla & firefox.
.mrg.
From: matthew green <mrg@eterna.com.au>
To: gnats-bugs@netbsd.org
Cc:
Subject: Re: pkg/23240
Date: Mon, 05 Apr 2004 19:25:16 +1000
with the patch below prprf.c works. as it changes all __NetBSD__
platforms i don't want to just commit it... but i still have the
Stub* lossage with 2.0B/macppc even with this change, so none of
the original problems are solved.
this patch probably applies directly to firefox as well. it
certainly has the exact same lossage in the same file...
.mrg.
Index: distinfo
===================================================================
RCS file: /cvsroot/pkgsrc/www/mozilla/distinfo,v
retrieving revision 1.50
diff -p -u -r1.50 distinfo
--- distinfo 22 Feb 2004 12:02:55 -0000 1.50
+++ distinfo 5 Apr 2004 09:23:56 -0000
@@ -9,6 +9,7 @@ SHA1 (patch-ad) = 5408db4b7585dcf5855f64
SHA1 (patch-ae) = e10936ff84d8f5b631c5c9f18271f21e902ea351
SHA1 (patch-af) = 8fcedb0bdd48f9527fff8c9b970a0e6ea68711e4
SHA1 (patch-ag) = c52586f92fc801571ccaf5a80c081103a99ed4e7
+SHA1 (patch-ah) = b79f1bdc12b64538ca18a9ca872a259e88b3b74c
SHA1 (patch-am) = 6646e8f1e5dbc3532ba45945ad2054756de2186a
SHA1 (patch-aw) = 5beeefa214a3003a74ccc68776a97a414c6acff1
SHA1 (patch-ba) = e11ab3e2baa19c9db9e320868a9860fedf5c8cec
Index: patches/patch-ah
===================================================================
RCS file: patches/patch-ah
diff -N patches/patch-ah
--- /dev/null 1 Jan 1970 00:00:00 -0000
+++ patches/patch-ah 5 Apr 2004 09:23:56 -0000
@@ -0,0 +1,18 @@
+$NetBSD$
+
+--- nsprpub/pr/src/io/prprf.c.orig 2003-02-27 10:53:42.000000000 +1100
++++ nsprpub/pr/src/io/prprf.c 2004-04-05 18:44:09.000000000 +1000
+@@ -54,10 +54,10 @@
+ #if (defined(LINUX) && defined(__powerpc__)) || \
+ (defined(LINUX) && defined(__s390__)) || \
+ (defined(LINUX) && defined(__s390x__)) || \
+- defined(WIN16) || defined(QNX) || \
+- (defined(__NetBSD__) && defined(__powerpc__) && \
+- __NetBSD_Version__ < 105000000)
++ defined(WIN16) || defined(QNX)
+ #define VARARGS_ASSIGN(foo, bar) foo[0] = bar[0]
++#elif defined(__NetBSD__)
++#define VARARGS_ASSIGN(foo, bar) va_copy((foo), (bar))
+ #else
+ #define VARARGS_ASSIGN(foo, bar) (foo) = (bar)
+ #endif
From: matthew green <mrg@eterna.com.au>
To: gnats-bugs@netbsd.org
Cc: vincent@labri.fr
Subject: re: pkg/23240
Date: Tue, 06 Apr 2004 23:10:48 +1000
with this patch for pkgsrc mozilla builds on macppc current, but it
dumps core at install time. i haven't yet attempted to look why,
but i'm going to try building with -g and/or -O1 next.
.mrg.
Index: distinfo
===================================================================
RCS file: /cvsroot/pkgsrc/www/mozilla/distinfo,v
retrieving revision 1.50
diff -p -r1.50 distinfo
*** distinfo 22 Feb 2004 12:02:55 -0000 1.50
--- distinfo 6 Apr 2004 13:08:15 -0000
*************** SHA1 (mozilla-source-1.6.tar.bz2) = baec
*** 4,14 ****
Size (mozilla-source-1.6.tar.bz2) = 31037452 bytes
SHA1 (patch-aa) = 03f158fe673a54b8060d541c80f9ebda2977afc5
SHA1 (patch-ab) = 3abad84d0768de60f845f26c61cc2629a6e25ce3
! SHA1 (patch-ac) = 4cddfd0bfc8acd97db3b974289a79c9ae41c9edf
SHA1 (patch-ad) = 5408db4b7585dcf5855f64a342cc721a23aec449
SHA1 (patch-ae) = e10936ff84d8f5b631c5c9f18271f21e902ea351
SHA1 (patch-af) = 8fcedb0bdd48f9527fff8c9b970a0e6ea68711e4
SHA1 (patch-ag) = c52586f92fc801571ccaf5a80c081103a99ed4e7
SHA1 (patch-am) = 6646e8f1e5dbc3532ba45945ad2054756de2186a
SHA1 (patch-aw) = 5beeefa214a3003a74ccc68776a97a414c6acff1
SHA1 (patch-ba) = e11ab3e2baa19c9db9e320868a9860fedf5c8cec
--- 4,17 ----
Size (mozilla-source-1.6.tar.bz2) = 31037452 bytes
SHA1 (patch-aa) = 03f158fe673a54b8060d541c80f9ebda2977afc5
SHA1 (patch-ab) = 3abad84d0768de60f845f26c61cc2629a6e25ce3
! SHA1 (patch-ac) = bc0aad7fc0ab64b1e2c9673144bbd3609a375731
SHA1 (patch-ad) = 5408db4b7585dcf5855f64a342cc721a23aec449
SHA1 (patch-ae) = e10936ff84d8f5b631c5c9f18271f21e902ea351
SHA1 (patch-af) = 8fcedb0bdd48f9527fff8c9b970a0e6ea68711e4
SHA1 (patch-ag) = c52586f92fc801571ccaf5a80c081103a99ed4e7
+ SHA1 (patch-ah) = b79f1bdc12b64538ca18a9ca872a259e88b3b74c
+ SHA1 (patch-ai) = e5c40bb3639b5ff6d7e4f87ddf33f7cb51d4a42a
+ SHA1 (patch-aj) = aa6fe6a0ad514028596b9e246595192970f78f13
SHA1 (patch-am) = 6646e8f1e5dbc3532ba45945ad2054756de2186a
SHA1 (patch-aw) = 5beeefa214a3003a74ccc68776a97a414c6acff1
SHA1 (patch-ba) = e11ab3e2baa19c9db9e320868a9860fedf5c8cec
Index: patches/patch-ac
===================================================================
RCS file: /cvsroot/pkgsrc/www/mozilla/patches/patch-ac,v
retrieving revision 1.21
diff -p -r1.21 patch-ac
*** patches/patch-ac 17 Jan 2004 00:04:49 -0000 1.21
--- patches/patch-ac 6 Apr 2004 13:08:15 -0000
***************
*** 1,8 ****
$NetBSD: patch-ac,v 1.21 2004/01/17 00:04:49 taya Exp $
! diff -ru ../Orig/mozilla/xpcom/reflect/xptcall/src/md/unix/Makefile.in ./xpcom/reflect/xptcall/src/md/unix/Makefile.in
! --- ../Orig/mozilla/xpcom/reflect/xptcall/src/md/unix/Makefile.in 2003-08-15 13:00:23.000000000 +0900
! +++ ./xpcom/reflect/xptcall/src/md/unix/Makefile.in 2004-01-16 23:17:21.000000000 +0900
@@ -135,7 +135,7 @@
# NetBSD/ARM
#
--- 1,20 ----
$NetBSD: patch-ac,v 1.21 2004/01/17 00:04:49 taya Exp $
! --- xpcom/reflect/xptcall/src/md/unix/Makefile.in.orig 2003-08-15 14:00:23.000000000 +1000
! +++ xpcom/reflect/xptcall/src/md/unix/Makefile 2004-04-06 16:47:38.000000000 +1000
! @@ -20,9 +20,9 @@
! #
!
! DEPTH = ../../../../../..
! -topsrcdir = @top_srcdir@
! -srcdir = @srcdir@
! -VPATH = @srcdir@
! +topsrcdir = ../../../../../..
! +srcdir = .
! +VPATH = .
!
! include $(DEPTH)/config/autoconf.mk
!
@@ -135,7 +135,7 @@
# NetBSD/ARM
#
*************** diff -ru ../Orig/mozilla/xpcom/reflect/x
*** 21,27 ****
CPPSRCS := xptcinvoke_netbsd_m68k.cpp xptcstubs_netbsd_m68k.cpp
endif
endif
! @@ -295,6 +295,13 @@
CPPSRCS := xptcinvoke_sparc_netbsd.cpp xptcstubs_sparc_netbsd.cpp
ASFILES := xptcinvoke_asm_sparc_netbsd.s xptcstubs_asm_sparc_netbsd.s
endif
--- 33,47 ----
CPPSRCS := xptcinvoke_netbsd_m68k.cpp xptcstubs_netbsd_m68k.cpp
endif
endif
! @@ -245,6 +245,7 @@
! # NetBSD/PPC
! #
! ifneq (,$(filter NetBSDmacppc NetBSDbebox NetBSDofppc NetBSDprep NetBSDamigappc,$(OS_ARCH)$(OS_TEST)))
! +AS := $(CC) -c -x assembler-with-cpp
! CPPSRCS := xptcinvoke_ppc_netbsd.cpp xptcstubs_ppc_netbsd.cpp
! ASFILES := xptcinvoke_asm_ppc_netbsd.s xptcstubs_asm_ppc_netbsd.s
! endif
! @@ -295,6 +296,13 @@
CPPSRCS := xptcinvoke_sparc_netbsd.cpp xptcstubs_sparc_netbsd.cpp
ASFILES := xptcinvoke_asm_sparc_netbsd.s xptcstubs_asm_sparc_netbsd.s
endif
Index: patches/patch-ah
===================================================================
RCS file: patches/patch-ah
diff -N patches/patch-ah
*** /dev/null 1 Jan 1970 00:00:00 -0000
--- patches/patch-ah 6 Apr 2004 13:08:15 -0000
***************
*** 0 ****
--- 1,18 ----
+ $NetBSD$
+
+ --- nsprpub/pr/src/io/prprf.c.orig 2003-02-27 10:53:42.000000000 +1100
+ +++ nsprpub/pr/src/io/prprf.c 2004-04-05 18:44:09.000000000 +1000
+ @@ -54,10 +54,10 @@
+ #if (defined(LINUX) && defined(__powerpc__)) || \
+ (defined(LINUX) && defined(__s390__)) || \
+ (defined(LINUX) && defined(__s390x__)) || \
+ - defined(WIN16) || defined(QNX) || \
+ - (defined(__NetBSD__) && defined(__powerpc__) && \
+ - __NetBSD_Version__ < 105000000)
+ + defined(WIN16) || defined(QNX)
+ #define VARARGS_ASSIGN(foo, bar) foo[0] = bar[0]
+ +#elif defined(__NetBSD__)
+ +#define VARARGS_ASSIGN(foo, bar) va_copy((foo), (bar))
+ #else
+ #define VARARGS_ASSIGN(foo, bar) (foo) = (bar)
+ #endif
Index: patches/patch-ai
===================================================================
RCS file: patches/patch-ai
diff -N patches/patch-ai
*** /dev/null 1 Jan 1970 00:00:00 -0000
--- patches/patch-ai 6 Apr 2004 13:08:15 -0000
***************
*** 0 ****
--- 1,14 ----
+ $NetBSD$
+
+ --- xpcom/reflect/xptcall/src/md/unix/xptcinvoke_asm_ppc_netbsd.s.orig 2001-03-12 04:22:17.000000000 +1100
+ +++ xpcom/reflect/xptcall/src/md/unix/xptcinvoke_asm_ppc_netbsd.s 2004-04-06 11:39:54.000000000 +1000
+ @@ -86,7 +86,9 @@
+ lwz r4,12(r31) # r4 <= methodIndex
+ lwz r5,0(r3) # r5 <= vtable ( == *that )
+ slwi r4,r4,3 # convert to offset ( *= 8 )
+ +#if !((__GNUC__ == 3 && __GNUC_MINOR__ < 2) || __GXX_ABI_VERSION >= 100) // G++ pre-V3 ABI
+ addi r4,r4,8 # skip first two vtable entries
+ +#endif
+ add r4,r4,r5
+ lhz r0,0(r4) # virtual base offset
+ extsh r0,r0
Index: patches/patch-aj
===================================================================
RCS file: patches/patch-aj
diff -N patches/patch-aj
*** /dev/null 1 Jan 1970 00:00:00 -0000
--- patches/patch-aj 6 Apr 2004 13:08:15 -0000
***************
*** 0 ****
--- 1,60 ----
+ $NetBSD$
+
+ --- xpcom/reflect/xptcall/src/md/unix/xptcstubs_ppc_netbsd.cpp.orig 2001-09-29 06:12:53.000000000 +1000
+ +++ xpcom/reflect/xptcall/src/md/unix/xptcstubs_ppc_netbsd.cpp 2004-04-06 11:39:08.000000000 +1000
+ @@ -195,6 +195,8 @@
+ // however, it's quick, dirty, and'll break when the ABI changes on
+ // us, which is what we want ;-).
+
+ +#if __GXX_ABI_VERSION < 100
+ +// gcc-2 version
+ #define STUB_ENTRY(n) \
+ __asm__ ( \
+ ".section \".text\" \n\t" \
+ @@ -206,6 +208,46 @@
+ "li 11,"#n" \n\t" \
+ "b SharedStub@local \n" \
+ );
+ +#else
+ +// gcc-3 version
+ +//
+ +// As G++3 ABI contains the length of the functionname in the mangled
+ +// name, it is difficult to get a generic assembler mechanism like
+ +// in the G++ 2.95 case.
+ +// Create names would be like:
+ +// _ZN14nsXPTCStubBase5Stub1Ev
+ +// _ZN14nsXPTCStubBase6Stub12Ev
+ +// _ZN14nsXPTCStubBase7Stub123Ev
+ +// _ZN14nsXPTCStubBase8Stub1234Ev
+ +// etc.
+ +// Use assembler directives to get the names right...
+ +
+ +# define STUB_ENTRY(n) \
+ +__asm__ ( \
+ + ".align 2 \n\t" \
+ + ".if "#n" < 10 \n\t" \
+ + ".globl _ZN14nsXPTCStubBase5Stub"#n"Ev \n\t" \
+ + ".type _ZN14nsXPTCStubBase5Stub"#n"Ev,@function \n\n" \
+ +"_ZN14nsXPTCStubBase5Stub"#n"Ev: \n\t" \
+ + \
+ + ".elseif "#n" < 100 \n\t" \
+ + ".globl _ZN14nsXPTCStubBase6Stub"#n"Ev \n\t" \
+ + ".type _ZN14nsXPTCStubBase6Stub"#n"Ev,@function \n\n" \
+ +"_ZN14nsXPTCStubBase6Stub"#n"Ev: \n\t" \
+ + \
+ + ".elseif "#n" < 1000 \n\t" \
+ + ".globl _ZN14nsXPTCStubBase7Stub"#n"Ev \n\t" \
+ + ".type _ZN14nsXPTCStubBase7Stub"#n"Ev,@function \n\n" \
+ +"_ZN14nsXPTCStubBase7Stub"#n"Ev: \n\t" \
+ + \
+ + ".else \n\t" \
+ + ".err \"stub number "#n" >= 1000 not yet supported\"\n" \
+ + ".endif \n\t" \
+ + \
+ + "li 11,"#n" \n\t" \
+ + "b SharedStub@local \n" \
+ +);
+ +#endif
+
+ #define SENTINEL_ENTRY(n) \
+ nsresult nsXPTCStubBase::Sentinel##n() \
From: Aymeric Vincent <vincent@labri.fr>
To: matthew green <mrg@eterna.com.au>
Cc: gnats-bugs@netbsd.org
Subject: Re: pkg/23240
Date: Tue, 6 Apr 2004 16:45:19 +0200
I couldn't resist the temptation to give it a look and I found two
typos:
In patch-ac, this should really be Makefile.in otherwise the patch
won't apply
> ! +++ xpcom/reflect/xptcall/src/md/unix/Makefile
> + slwi r4,r4,3 # convert to offset ( *= 8 )
> + +#if !((__GNUC__ == 3 && __GNUC_MINOR__ < 2) || __GXX_ABI_VERSION
> >= 100) // G++ pre-V3 ABI
> + addi r4,r4,8 # skip first two vtable entries
If the comments are correct, this should be addi r4,r4,16, right?
I'll be starting a build of mozilla, but it will be from scratch so I
don't expect useful results before a long time + a few fsck's... :-/
Aymeric
From: matthew green <mrg@eterna.com.au>
To: Aymeric Vincent <vincent@labri.fr>
Cc: gnats-bugs@netbsd.org
Subject: re: pkg/23240
Date: Wed, 07 Apr 2004 01:01:11 +1000
I couldn't resist the temptation to give it a look and I found two
typos:
In patch-ac, this should really be Makefile.in otherwise the patch
won't apply
> ! +++ xpcom/reflect/xptcall/src/md/unix/Makefile
yeah i noticed that in my -g build....
> + slwi r4,r4,3 # convert to offset ( *= 8 )
> + +#if !((__GNUC__ == 3 && __GNUC_MINOR__ < 2) || __GXX_ABI_VERSION
> >= 100) // G++ pre-V3 ABI
> + addi r4,r4,8 # skip first two vtable entries
If the comments are correct, this should be addi r4,r4,16, right?
I'll be starting a build of mozilla, but it will be from scratch so I
don't expect useful results before a long time + a few fsck's... :-/
hmmm the code from linux just made that addi not exist for new abi.
are you saying it needs to be skip 4 entries?
From: Aymeric Vincent <Aymeric.Vincent@labri.fr>
To: matthew green <mrg@eterna.com.au>
Cc: gnats-bugs@netbsd.org
Subject: Re: pkg/23240
Date: 06 Apr 2004 17:45:04 +0200
matthew green <mrg@eterna.com.au> writes:
> hmmm the code from linux just made that addi not exist for new abi.
> are you saying it needs to be skip 4 entries?
Nope, I don't know. I just understood after sending the mail that
actually your patch was removing that instruction... So it doesn't
matter that the comment is wrong or whatever. It just striked me as
incorrect that entries were 8 bytes long (instead of 4 in linux) and
that we added 8 to jump over two (2) entries.
Actually, I was wondering why we just don't use the linux code? I
didn't see any place in gcc where netbsd differs from linux from the
ABI point of view. I didn't look very thoroughly and I'm not used to
the code. Do you know where I could find the place where the virtual
table's layout is decided within gcc?
Aymeric
From: matthew green <mrg@eterna.com.au>
To: Aymeric Vincent <Aymeric.Vincent@labri.fr>
Cc: gnats-bugs@netbsd.org
Subject: re: pkg/23240
Date: Wed, 26 May 2004 03:13:46 +1000
> hmmm the code from linux just made that addi not exist for new abi.
> are you saying it needs to be skip 4 entries?
Nope, I don't know. I just understood after sending the mail that
actually your patch was removing that instruction... So it doesn't
matter that the comment is wrong or whatever. It just striked me as
incorrect that entries were 8 bytes long (instead of 4 in linux) and
that we added 8 to jump over two (2) entries.
Actually, I was wondering why we just don't use the linux code? I
didn't see any place in gcc where netbsd differs from linux from the
ABI point of view. I didn't look very thoroughly and I'm not used to
the code. Do you know where I could find the place where the virtual
table's layout is decided within gcc?
afaik it should probably work?
From: Aymeric Vincent <vincent@labri.fr>
To: mrg@netbsd.org
Cc: gnats-bugs@netbsd.org
Subject: pkg/23240 (mozilla on macppc)
Date: 30 Jun 2004 14:13:04 +0200
Hi Matthew,
I have good news. I finally did what I suggested: ie. use the stubs
from linux and it gives me a working mozilla, except for the menus and
more generally all (gtk?) widgets which are broken.
So there is still some work to do, but at least we have a decent way
to read HTML.
The following discussion takes place in
pkgsrc/www/mozilla/work.powerpc/mozilla/xpcom/reflect/xptcall/src/md/unix
Replacing the small Makefile fragment with this one and recompiling
and installing libxpcom.so produces a kind-of-working mozilla.
ifneq (,$(filter NetBSDmacppc NetBSDbebox NetBSDofppc NetBSDprep NetBSDamigappc,$(OS_ARC
H)$(OS_TEST)))
CPPSRCS := xptcinvoke_ppc_linux.cpp xptcstubs_ppc_linux.cpp
ASFILES := xptcinvoke_asm_ppc_linux.s xptcstubs_asm_ppc_linux.s
AS := $(CC) -c -x assembler-with-cpp
endif
I'm wondering if we shouldn't just remove the
xptcinvoke_asm_ppc_netbsd.s and xptcstubs_asm_ppc_netbsd.s files
altogether.
You are welcome to investigate this further: I don't expect to be
working on that very much in the coming days.
Regards,
Aymeric
From: Aymeric Vincent <vincent@labri.fr>
To: mrg@netbsd.org
Cc: gnats-bugs@netbsd.org
Subject: Re: pkg/23240 (mozilla on macppc)
Date: 30 Jun 2004 14:15:38 +0200
Oh, joy! On second invocation, everything works fine. It must be
related to the fact that the installation process couldn't finish
properly the first time I did it (regxpcom and regchrome dumped core).
Aymeric
From: matthew green <mrg@eterna.com.au>
To: Aymeric Vincent <vincent@labri.fr>
Cc: gnats-bugs@netbsd.org
Subject: re: pkg/23240 (mozilla on macppc)
Date: Wed, 30 Jun 2004 22:44:07 +1000
hiya.
i'm unlikely to look at this at all for months if ever. i don't use
mozilla anymore and my mac is mostly a frontroom a/v machine not a
desktop anymore...
glad you got forward somewhere tho.
.mrg.
Responsible-Changed-From-To: pkg-manager->taya
Responsible-Changed-By: wiz
Responsible-Changed-When: Tue Jul 6 10:48:12 UTC 2004
Responsible-Changed-Why:
over to maintainer
From: Aymeric Vincent <aymeric@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc:
Subject: pr/23240 CVS commit: pkgsrc/www/mozilla
Date: Wed, 7 Jul 2004 09:08:31 +0000 (UTC)
Module Name: pkgsrc
Committed By: aymeric
Date: Wed Jul 7 09:08:31 UTC 2004
Modified Files:
pkgsrc/www/mozilla: distinfo
pkgsrc/www/mozilla/patches: patch-ac patch-bx
Added Files:
pkgsrc/www/mozilla/patches: patch-bz patch-ca patch-cb
Log Message:
. on PowerPC, update files so that Mozilla works properly when compiled with
gcc version 3+.
. generally reduce diffs to Linux version
. retain compatibility with older ABI (AIX-like) thanks to useful comments
from Charles Hannum
Thanks to Matthew Green for the fruitful discussion. This should address
PR#23240 as far as mozilla is concerned.
To generate a diff of this commit:
cvs rdiff -r1.59 -r1.60 pkgsrc/www/mozilla/distinfo
cvs rdiff -r1.23 -r1.24 pkgsrc/www/mozilla/patches/patch-ac
cvs rdiff -r1.1 -r1.2 pkgsrc/www/mozilla/patches/patch-bx
cvs rdiff -r0 -r1.1 pkgsrc/www/mozilla/patches/patch-bz \
pkgsrc/www/mozilla/patches/patch-ca pkgsrc/www/mozilla/patches/patch-cb
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
From: Aymeric Vincent <vincent@labri.fr>
To: gnats-bugs@gnats.NetBSD.org
Cc: mrg@netbsd.org
Subject: pkg/23240 (mozilla and firefox fail on macppc)
Date: Tue, 24 Aug 2004 18:45:00 +0200
With my latest commits to firefox, I think this PR can be closed/put in
feedback state.
However, I still see some inconsistent GTk layouts in many windows, but
"not always". I have no clue for this issue and it may be specific to
my setup.
Regards,
Aymeric
State-Changed-From-To: open->closed
State-Changed-By: mbw@netbsd.org
State-Changed-When: Sat, 20 Nov 2004 06:57:09 +0000
State-Changed-Why:
From: Shin'ichiro TAYA <taya@netbsd.org>
To: gnats-bugs@netbsd.org
Cc:
Subject: PR/23240 CVS commit: pkgsrc/mail
Date: Sat, 4 Dec 2004 02:16:03 +0000 (UTC)
Module Name: pkgsrc
Committed By: taya
Date: Sat Dec 4 02:16:03 UTC 2004
Modified Files:
pkgsrc/mail/thunderbird: Makefile-thunderbird.common PLIST distinfo
pkgsrc/mail/thunderbird-gtk2: PLIST
pkgsrc/mail/thunderbird/patches: patch-ab patch-ac patch-br patch-bx
Added Files:
pkgsrc/mail/thunderbird/patches: patch-bz patch-ca patch-cb patch-cc
Removed Files:
pkgsrc/mail/thunderbird/patches: patch-bn
Log Message:
1)Generate extension directory at installation time.
2)sync patches from www/mozilla.
- update patch-ab
from commit log
>> date: 2004/08/31 02:37:57; author: danw; state: Exp; lines: +13 -6
>> darwin fixes (tested against firefox-gtk2)
- update patch-ac
- update patch-bx
- add patch-bz
- add patch-ca
- add patch-cb
from commit log
>> date: 2004/07/07 09:08:31; author: aymeric; state: Exp; lines: +14 -7
>> . on PowerPC, update files so that Mozilla works properly when compiled with
>> gcc version 3+.
>> . generally reduce diffs to Linux version
>> . retain compatibility with older ABI (AIX-like) thanks to useful comments
>> from Charles Hannum
>>
>> Thanks to Matthew Green for the fruitful discussion. This should address
>> PR#23240 as far as mozilla is concerned.
- remove patch-bn
enable HAVE_SOCKLEN_T
- update patch-br
from commit log
>> date: 2004/10/04 11:52:45; author: grant; state: Exp; lines: +10 -6
>> bring across a patch in Firefox for using thread-safe resolver library
>> functions on NetBSD >=2.0F.
- update patch-cc
make mozilla work on NetBSD-current/alpha
3)bump PKGREVISION
To generate a diff of this commit:
cvs rdiff -r1.3 -r1.4 pkgsrc/mail/thunderbird/Makefile-thunderbird.common
cvs rdiff -r1.6 -r1.7 pkgsrc/mail/thunderbird/PLIST
cvs rdiff -r1.9 -r1.10 pkgsrc/mail/thunderbird/distinfo
cvs rdiff -r1.5 -r1.6 pkgsrc/mail/thunderbird-gtk2/PLIST
cvs rdiff -r1.3 -r1.4 pkgsrc/mail/thunderbird/patches/patch-ab \
pkgsrc/mail/thunderbird/patches/patch-ac \
pkgsrc/mail/thunderbird/patches/patch-br
cvs rdiff -r1.3 -r0 pkgsrc/mail/thunderbird/patches/patch-bn
cvs rdiff -r1.1 -r1.2 pkgsrc/mail/thunderbird/patches/patch-bx
cvs rdiff -r0 -r1.1 pkgsrc/mail/thunderbird/patches/patch-bz \
pkgsrc/mail/thunderbird/patches/patch-ca \
pkgsrc/mail/thunderbird/patches/patch-cb \
pkgsrc/mail/thunderbird/patches/patch-cc
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
>Unformatted:
(Contact us)
$NetBSD: query-full-pr,v 1.51 2026/08/10 02:28:17 riastradh Exp $
$NetBSD: gnats_config.sh,v 1.10 2026/05/13 22:00:09 riastradh Exp $
Copyright © 1994-2026
The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.