HEADS UP: GCC 3.4.6 update in progress
Alexander Kabaev
update GCC 3.4 in the tree to the latest version available in FSF
SVN repository. After a short period in -current this version will be
MFC-ed to RELENG6, which will then be able to take advantage of all the
fixes that went info FSF sources since we did last compiler update.
The update will happen over next hour or so. I will post 'all clear'
message once the system is back to consistent state.
--
Alexander Kabaev
_______________________________________________
freebsd...@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-curre...@freebsd.org"
Alexander Kabaev
break because of it. I am very interested in being notified of anything
that contradicts with an above statement.
Andrew R. Reiter
On Sat, 26 Aug 2006, Alexander Kabaev wrote:
:All is clear. This update should be pretty uneventful and nothing should
:break because of it. I am very interested in being notified of anything
:that contradicts with an above statement.
:
Any specific benefits from the move minus just "the latest & greatest"?
Jsut curious.
peace,
andrew
:--
:Alexander Kabaev
:_______________________________________________
:freebsd...@freebsd.org mailing list
:http://lists.freebsd.org/mailman/listinfo/freebsd-current
:To unsubscribe, send any mail to "freebsd-curre...@freebsd.org"
:
:
Steve Kargl
>
> On Sat, 26 Aug 2006, Alexander Kabaev wrote:
>
> :All is clear. This update should be pretty uneventful and nothing should
> :break because of it. I am very interested in being notified of anything
> :that contradicts with an above statement.
> :
>
> Any specific benefits from the move minus just "the latest & greatest"?
> Jsut curious.
>
--
Steve
Alexander Kabaev
"Andrew R. Reiter" <a...@watson.org> wrote:
>
> On Sat, 26 Aug 2006, Alexander Kabaev wrote:
>
> :All is clear. This update should be pretty uneventful and nothing
> should :break because of it. I am very interested in being notified
> of anything :that contradicts with an above statement.
> :
>
> Any specific benefits from the move minus just "the latest &
> greatest"? Jsut curious.
>
> peace,
> andrew
>
> :--
> :Alexander Kabaev
> :_______________________________________________
> :freebsd...@freebsd.org mailing list
> :http://lists.freebsd.org/mailman/listinfo/freebsd-current
> :To unsubscribe, send any mail to
> "freebsd-curre...@freebsd.org" :
> :
>
> --
> a...@watson.org
Two releases worth of bugfixes mainly. Enough to justify 'latest and
greatest' sentiments.
--
Alexander Kabaev
Andrew R. Reiter
:On Sat, Aug 26, 2006 at 07:00:08PM -0400, Andrew R. Reiter wrote:
:>
:> On Sat, 26 Aug 2006, Alexander Kabaev wrote:
:>
:> :All is clear. This update should be pretty uneventful and nothing should
:> :break because of it. I am very interested in being notified of anything
:> :that contradicts with an above statement.
:> :
:>
:> Any specific benefits from the move minus just "the latest & greatest"?
:> Jsut curious.
:>
:
:http://gcc.gnu.org/gcc-3.4/
:
Thanks; good point :-)
Cheers,
--
a...@watson.org
Divacky Roman
> In order to pave the way for an upcoming GCC 4.1 import, we decided to
thnx for the work! when can we expect the 4.1 import? tentatively..
thnx
roman
Slawa Olhovchenkov
> All is clear. This update should be pretty uneventful and nothing should
> break because of it. I am very interested in being notified of anything
> that contradicts with an above statement.
buildkernel break with many "warning: 'XXX' might be used
uninitialized in this function"
Alexander Kabaev
Slawa Olhovchenkov <s...@zxy.spb.ru> wrote:
> On Sat, Aug 26, 2006 at 10:36:08PM +0000, Alexander Kabaev wrote:
>
> > All is clear. This update should be pretty uneventful and nothing
> > should break because of it. I am very interested in being notified
> > of anything that contradicts with an above statement.
>
> buildkernel break with many "warning: 'XXX' might be used
> uninitialized in this function"
>
<YOUR LOG GOES HERE>
______________________________________________
> freebsd...@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-current
> To unsubscribe, send any mail to
> "freebsd-curre...@freebsd.org"
--
Alexander Kabaev
Norikatsu Shigemura
Alexander Kabaev <k...@freebsd.org> wrote:
> All is clear. This update should be pretty uneventful and nothing should
> break because of it. I am very interested in being notified of anything
> that contradicts with an above statement.
I couldn't compile gcc, because of following error message.
===> gnu/usr.bin/cc/cc1plus (all)
cc -O2 -fno-strict-aliasing -pipe -DIN_GCC -DHAVE_CONFIG_H -DPREFIX=\"/usr/obj/usr/src/tmp/usr\" -I/usr/obj/usr/src/tmp/usr/src/gnu/usr.bin/cc/cc1plus/../cc_tools -I/usr/src/gnu/usr.bin/cc/cc1plus/../cc_tools -I/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc -I/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/config -I/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp -I. -I/usr/obj/usr/src/tmp/legacy/usr/include -c /usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/decl.c
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/decl.c: In function `grokfndecl':
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/decl.c:5721: error: too many arguments to function `grok_op_properties'
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/decl.c: At top level:
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/decl.c:9008: error: conflicting types for 'grok_op_properties'
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/cp-tree.h:3686: error: previous declaration of 'grok_op_properties' was here
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/decl.c:9008: error: conflicting types for 'grok_op_properties'
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/cp-tree.h:3686: error: previous declaration of 'grok_op_properties' was here
I can fix it such as following patch.
--- contrib/gcc/cp/cp-tree.h.orig Sun Aug 27 06:29:09 2006
+++ contrib/gcc/cp/cp-tree.h Mon Aug 28 07:42:35 2006
@@ -3683,7 +3683,7 @@
extern tree get_scope_of_declarator (tree);
extern void grok_special_member_properties (tree);
extern int grok_ctor_properties (tree, tree);
-extern bool grok_op_properties (tree, bool);
+extern bool grok_op_properties (tree, int, bool);
extern tree xref_tag (enum tag_types, tree, bool, bool);
extern tree xref_tag_from_type (tree, tree, int);
extern void xref_basetypes (tree, tree);
But....
cc -I/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/config -I/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp -I. -I/usr/obj/usr/src/tmp/legacy/usr/include -c /usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/pt.c
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/pt.c: In function `tsubst_decl':
/usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/pt.c:6315: error: too few arguments to function `grok_op_properties'
Sorry, I don't know how to fix:-(. I have no idea.
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
}
else if (IDENTIFIER_OPNAME_P (DECL_NAME (r)))
grok_op_properties (r, (complain & tf_error) != 0);
if (DECL_FRIEND_P (t) && DECL_FRIEND_CONTEXT (t))
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Norikatsu Shigemura
Norikatsu Shigemura <no...@freebsd.org> wrote:
> On Sat, 26 Aug 2006 22:36:08 +0000
> Alexander Kabaev <k...@freebsd.org> wrote:
> > All is clear. This update should be pretty uneventful and nothing should
> > break because of it. I am very interested in being notified of anything
> > that contradicts with an above statement.
> I couldn't compile gcc, because of following error message.
Humm... This may be mirror sites issue. I research following
sites:
cvsup.jp.freebsd.org
cvsup3.jp.freebsd.org
Alexander Kabaev
From /usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/CVS/Entries:
/decl.c/1.20/Sun Aug 27 01:02:46 2006//
From decl.c:9008
bool
grok_op_properties (tree decl, bool complain)
From cvs diff -r1.19 -r1.20 contrib/gcc/cp/decl.c:
bool
-grok_op_properties (tree decl, int friendp, bool complain)
+grok_op_properties (tree decl, bool complain)
{
Your src tree is out of whack. Update and try again. Change your cvsup
server if that will not be enough.
--
Alexander Kabaev
Norikatsu Shigemura
Alexander Kabaev <kab...@gmail.com> wrote:
> From /usr/src/gnu/usr.bin/cc/cc1plus/../../../../contrib/gcc/cp/CVS/Entries:
> /decl.c/1.20/Sun Aug 27 01:02:46 2006//
> From decl.c:9008
> bool
> grok_op_properties (tree decl, bool complain)
> From cvs diff -r1.19 -r1.20 contrib/gcc/cp/decl.c:
> bool
> -grok_op_properties (tree decl, int friendp, bool complain)
> +grok_op_properties (tree decl, bool complain)
> {
> Your src tree is out of whack. Update and try again. Change your cvsup
> server if that will not be enough.
All japanese cvsup mirror sites had same problem:-). (decl.c,v was
broken) kuriyama@ fixed this problem. And I confirmed OK:-).
Thank you!
Conrad J. Sabatier
wrote:
> In order to pave the way for an upcoming GCC 4.1 import, we decided to
> update GCC 3.4 in the tree to the latest version available in FSF
> SVN repository. After a short period in -current this version will be
> MFC-ed to RELENG6, which will then be able to take advantage of all
> the fixes that went info FSF sources since we did last compiler
> update.
>
> The update will happen over next hour or so. I will post 'all clear'
> message once the system is back to consistent state.
Cool!
Incidentally, I've been wondering for the longest time now: is/are there
any particular reason(s) why we exclude gcj from the gcc tree/build?
Sure would be a cool thing to have.
--
Conrad J. Sabatier <con...@cox.net> -- "In Unix veritas"
Steve Kargl
> On Sat, 26 Aug 2006 21:21:20 +0000, Alexander Kabaev <k...@freebsd.org>
> > update GCC 3.4 in the tree to the latest version available in FSF
> > SVN repository. After a short period in -current this version will be
> > MFC-ed to RELENG6, which will then be able to take advantage of all
> > the fixes that went info FSF sources since we did last compiler
> > update.
> >
> > The update will happen over next hour or so. I will post 'all clear'
> > message once the system is back to consistent state.
>
> Cool!
>
> Incidentally, I've been wondering for the longest time now: is/are there
> any particular reason(s) why we exclude gcj from the gcc tree/build?
> Sure would be a cool thing to have.
>
1) It's not needed to build anything in the base system.
2) It takes a long time to build.
3) You really want to use the gcc41 or gcc42 port.
Of course, IMHO.
--
Steve
Conrad J. Sabatier
<s...@troutmask.apl.washington.edu> wrote:
> On Tue, Aug 29, 2006 at 05:50:22PM -0500, Conrad J. Sabatier wrote:
> >
> > Incidentally, I've been wondering for the longest time now: is/are
> > there any particular reason(s) why we exclude gcj from the gcc
> > tree/build? Sure would be a cool thing to have.
> >
>
> 1) It's not needed to build anything in the base system.
> 2) It takes a long time to build.
> 3) You really want to use the gcc41 or gcc42 port.
>
> Of course, IMHO.
I've tried both ports, but gcj fails to run, complaining about a
missing "spec" file (I forget the exact filename, maybe
"gcjlib.spec" or something similar). :-(
Is this maybe an arch-specific (amd64) problem? Or?
--
Conrad J. Sabatier <con...@cox.net> -- "In Unix veritas"
Steve Kargl
> On Tue, 29 Aug 2006 16:04:47 -0700, Steve Kargl
> <s...@troutmask.apl.washington.edu> wrote:
>
> > On Tue, Aug 29, 2006 at 05:50:22PM -0500, Conrad J. Sabatier wrote:
> > >
> > > Incidentally, I've been wondering for the longest time now: is/are
> > > there any particular reason(s) why we exclude gcj from the gcc
> > > tree/build? Sure would be a cool thing to have.
> > >
> >
> > 1) It's not needed to build anything in the base system.
> > 2) It takes a long time to build.
> > 3) You really want to use the gcc41 or gcc42 port.
> >
> > Of course, IMHO.
>
> I've tried both ports, but gcj fails to run, complaining about a
> missing "spec" file (I forget the exact filename, maybe
> "gcjlib.spec" or something similar). :-(
>
> Is this maybe an arch-specific (amd64) problem? Or?
>
Contact Gerald. He is fairly quick at addressing port
issues. I don't build the ports because I build GCC
svn sources for 4.1.x and trunk several times a week.
I'll note that I don't normally build gjc. I can try
a bootstrap tomorrow.
--
Steve
Conrad J. Sabatier
<s...@troutmask.apl.washington.edu> wrote:
> On Tue, Aug 29, 2006 at 08:29:16PM -0500, Conrad J. Sabatier wrote:
> > On Tue, 29 Aug 2006 16:04:47 -0700, Steve Kargl
> > <s...@troutmask.apl.washington.edu> wrote:
> >
> > > On Tue, Aug 29, 2006 at 05:50:22PM -0500, Conrad J. Sabatier
> > > wrote:
> > > >
> > > > Incidentally, I've been wondering for the longest time now:
> > > > is/are there any particular reason(s) why we exclude gcj from
> > > > the gcc tree/build? Sure would be a cool thing to have.
> > > >
> > >
> > > 1) It's not needed to build anything in the base system.
> > > 2) It takes a long time to build.
> > > 3) You really want to use the gcc41 or gcc42 port.
> > >
> > > Of course, IMHO.
> >
> > I've tried both ports, but gcj fails to run, complaining about a
> > missing "spec" file (I forget the exact filename, maybe
> > "gcjlib.spec" or something similar). :-(
> >
> > Is this maybe an arch-specific (amd64) problem? Or?
> >
>
> Contact Gerald. He is fairly quick at addressing port
> issues. I don't build the ports because I build GCC
> svn sources for 4.1.x and trunk several times a week.
> I'll note that I don't normally build gjc. I can try
> a bootstrap tomorrow.
OK, will do. Thanks.
--
Conrad J. Sabatier <con...@cox.net> -- "In Unix veritas"
Steve Kargl
For those interesting in bootstrap time.
% time gmake bootstrap
4456.51 real 4098.55 user 349.61 sys
This includes C, C++, Fortran, and Java. These numbers are from
a dual-processor, dual-core opteron system running at 2.4 GHz
and with 16 GB of memory. The system was otherwise idle.
--
Steve