Re: make buildworld fails

[Philip M. Gollucci]

Jamie Avery wrote:
> put KERNCONF=/kernelname/ into /etc/make.conf.  However, when I make
> buildkernel, I get the following error
7.x+ src.conf, below make.conf

grep KERNCONF /etc/make.conf /etc/src.conf
/etc/src.conf:KERNCONF?=  RIDERWAY

NO /s neccessary.



-- 

Philip M. Gollucci ([EMAIL PROTECTED])
o:703.549.2050x206
Senior System Admin - Riderway, Inc.
http://riderway.com / http://ridecharge.com
1024D/EC88A0BF 0DE5 C55C 6BF3 B235 2DAB  B89E 1324 9B4F EC88 A0BF

Work like you don't need the money,
love like you'll never get hurt,
and dance like nobody's watching.

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails on 6.2-STABLE

[JD Bronson]

At 08:19 PM 7/27/2007 +1200, Jonathan Chen wrote:

On Thu, Jul 26, 2007 at 11:12:26AM -0500, J.D. Bronson wrote:
> so I deleted /usr/src
> redownloaded from a different mirror and tried
> make buildworld again...
>
> It still failed -but this time at a different point:

Standard behaviour of failing hardware - most likely memory.


yes. I removed/cleaned/replaced the RAM and it built.
Not sure if the issue will return or not :)

-JD  


___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails on 6.2-STABLE

[Jonathan Chen]

On Thu, Jul 26, 2007 at 11:12:26AM -0500, J.D. Bronson wrote:
> so I deleted /usr/src
> redownloaded from a different mirror and tried
> make buildworld again...
> 
> It still failed -but this time at a different point:

Standard behaviour of failing hardware - most likely memory.
-- 
Jonathan Chen <[EMAIL PROTECTED]>
--
  Experience is a hard teacher
   because she gives the test first, the lesson afterwards
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails on 6.2-STABLE

[J.D. Bronson]

so I deleted /usr/src
redownloaded from a different mirror and tried
make buildworld again...

It still failed -but this time at a different point:


mkdep -f .depend -a-DIN_GCC -DHAVE_CONFIG_H -DPREFIX=\"/usr\" 
-I/usr/obj/usr/src/gnu/usr.bin/cc/cc/../cc_tools 
-I/usr/src/gnu/usr.bin/cc/cc/../cc_tools 
-I/usr/src/gnu/usr.bin/cc/cc/../../../../contrib/gcc 
-I/usr/src/gnu/usr.bin/cc/cc/../../../../contrib/gcc/config 
-DDEFAULT_TARGET_VERSION=\"3.4.6\" -DDEFAULT_TARGET_MACHINE=\"\" 
/usr/src/gnu/usr.bin/cc/cc/../../../../contrib/gcc/gcc.c 
/usr/src/gnu/usr.bin/cc/cc/../../../../contrib/gcc/gccspec.c
echo cc: /usr/obj/usr/src/tmp/usr/lib/libc.a 
/usr/obj/usr/src/gnu/usr.bin/cc/cc/../cc_int/libcc_int.a >> .depend

===> gnu/usr.bin/cc/cc1 (depend)
sed -e "/^@@ifobjc.*/,/^@@end_ifobjc.*/d"  -e "/^@@ifc.*/d" -e 
"/^@@end_ifc.*/d" 
/usr/src/gnu/usr.bin/cc/cc1/../../../../contrib/gcc/c-parse.in > c-parse.y

yacc -d c-parse.y
yacc: e - line 1811 of "c-parse.y", syntax error
{ if ($1 == error_}ark_node)
   ^
*** Error code 1

Stop in /usr/src/gnu/usr.bin/cc/cc1.
*** Error code 1

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails on 6.2-STABLE

[J.D. Bronson]

At 09:16 AM 07/26/2007, Heiko Wundram (Beenic) wrote:

Am Donnerstag 26 Juli 2007 15:54:36 schrieb J.D. Bronson:
> internal compiler error: Segmentation fault: 11
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See http://gcc.gnu.org/bugs.html> for instructions.

Most probably a (physical) memory error.

As the message says, this has pretty much nothing to do with the upping of
world, but is an "internal" compiler error, which I've only seen on
development snapshots of gcc (improbable that these are distributed with
STABLE), or flaky memory (which is much more likely the cause).

--


thanks - ironically I have never had ANY issue building world on this 
machine until today. I have deleted /usr/src and re cvs'd from a diff 
mirror as a test.


-JD 


___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails on 6.2-STABLE

[Heiko Wundram (Beenic)]

Am Donnerstag 26 Juli 2007 15:54:36 schrieb J.D. Bronson:
> internal compiler error: Segmentation fault: 11
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See http://gcc.gnu.org/bugs.html> for instructions.

Most probably a (physical) memory error.

As the message says, this has pretty much nothing to do with the upping of 
world, but is an "internal" compiler error, which I've only seen on 
development snapshots of gcc (improbable that these are distributed with 
STABLE), or flaky memory (which is much more likely the cause).

-- 
Heiko Wundram
Product & Application Development
-
Office Germany - EXPO PARK HANNOVER
 
Beenic Networks GmbH
Mailänder Straße 2
30539 Hannover
 
Fon+49 511 / 590 935 - 15
Fax+49 511 / 590 935 - 29
Mail   [EMAIL PROTECTED]


Beenic Networks GmbH
-
Sitz der Gesellschaft: Hannover
Geschäftsführer: Jorge Delgado
Registernummer: HRB 61869
Registergericht: Amtsgericht Hannover
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make Buildworld fails...why?

[Josh Paetzel]

Mark Stout wrote:
> Hello,
> 
> I've been trying to upgrade a 5.4-RELEASE to 6.2-RELEASE and having a bit of
> a problem.
> 
> Downloaded and untar's the source files for base, catpages, dict, doc,
> games, info, manpages, proflibs, and src directories per INSTALL.TXT in the
> releases/.../src directory after backing up my local /etc and /usr/local/etc
> directories.
> 
> I've tweaked the GENERIC kernel config file and renamed is RADIUS2
> 
> While in /usr/src I did a 'make buildworld'.
> 
>   I do okay until I get to "stage 2.3: build tools".  There I get to the
> following error when compiling make_hash.  Can anyone help me determine why
> this is failing?  Were my upgrade procedures
> 
> Thank you,
> Mark Stout
> VPM Global Internet Services, Inc.
> 530-626-4218 x205 Office
> 530-626-7182 Fax
> 530-554-9295 VoIP
> 916-240-2850 Cell
> www.vpm.com

Generally speaking the best supported upgrade path across major
version numbers is from the last release of the older version to the
first of the newer, which in your case would mean upgrading from 5.4
-> 5.5 -> 6.0 -> 6.2

Is there any particular reason you aren't using cvsup/csup to update
your source tree?

-- 
Thanks,

Josh Paetzel
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: [SOLVED] Re: make buildworld fails with CPUTYPE set

[Pietro Cerutti]

On 5/9/06, RW <[EMAIL PROTECTED]> wrote:

On Tuesday 09 May 2006 22:21, Pietro Cerutti wrote:
> On 5/9/06, Pietro Cerutti <[EMAIL PROTECTED]> wrote:
> > CPUTYPE=pentium3[m]
>
> I solved the problem by putting pentium3 instead of pentium3[m].
>
> So why it's stated "pentium3[m]" in line 35 of
> /usr/src/share/examples/etc/make.conf ?

[] is usually shorthand for an optional variant

i.e.  pentium3[m] means pentium3 or pentium3m


Thanx Erick and RW.
I knew the notation, but I didn't realize it was used there

--
Pietro Cerutti
<[EMAIL PROTECTED]>
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: [SOLVED] Re: make buildworld fails with CPUTYPE set

[RW]

On Tuesday 09 May 2006 22:21, Pietro Cerutti wrote:
> On 5/9/06, Pietro Cerutti <[EMAIL PROTECTED]> wrote:
> > CPUTYPE=pentium3[m]
>
> I solved the problem by putting pentium3 instead of pentium3[m].
>
> So why it's stated "pentium3[m]" in line 35 of
> /usr/src/share/examples/etc/make.conf ?

[] is usually shorthand for an optional variant

i.e.  pentium3[m] means pentium3 or pentium3m
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: [SOLVED] Re: make buildworld fails with CPUTYPE set

[Erik Trulsson]

On Tue, May 09, 2006 at 11:21:52PM +0200, Pietro Cerutti wrote:
> On 5/9/06, Pietro Cerutti <[EMAIL PROTECTED]> wrote:
> >CPUTYPE=pentium3[m]
> 
> I solved the problem by putting pentium3 instead of pentium3[m].
> 
> So why it's stated "pentium3[m]" in line 35 of
> /usr/src/share/examples/etc/make.conf ?
> 
> #   (Intel CPUs) nocona pentium4[m] prescott pentium3[m] pentium-m
> #pentium2 pentiumpro pentium-mmx pentium i486 i386
> 

It is just a shorthand for 'pentium3 pentium3m'.
The [xxx] syntax is often used in used in Unix documentation to denote
an optional part of a command line (or similar.)  In this case it is the
letter 'm' which can be added or not.



-- 

Erik Trulsson
[EMAIL PROTECTED]
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

[SOLVED] Re: make buildworld fails with CPUTYPE set

[Pietro Cerutti]

On 5/9/06, Pietro Cerutti <[EMAIL PROTECTED]> wrote:

CPUTYPE=pentium3[m]


I solved the problem by putting pentium3 instead of pentium3[m].

So why it's stated "pentium3[m]" in line 35 of
/usr/src/share/examples/etc/make.conf ?

#   (Intel CPUs) nocona pentium4[m] prescott pentium3[m] pentium-m
#pentium2 pentiumpro pentium-mmx pentium i486 i386

Thanx,



--
Pietro Cerutti
<[EMAIL PROTECTED]>


--
Pietro Cerutti
<[EMAIL PROTECTED]>
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails

[László Nagy]

Here's what I would try.  Upgrade to RELENG_4 via the cvsup/buildworld 
process,

read Bruce Mah's 4.11-5.X Migration Guide, and be sure and include
"mergemaster -p" in your buildworld cycle.


Can you tell me where is that migration guide? When I google for it, I 
only get posts about that guide. But where is the guide itself?


Another problem: I only have remote access to that machine. Is it 
possible to upgrade to 5.X without booting into single user mode?
Yes, theoretically, I shouldn't do that but I'm in Hungary and the 
computer is in the US...

Thanks,

 Laszlo

 
___

freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails

[Kevin Kinsey]

gandalf wrote:



 Hello,

I have this file /etc/sup.sys:

*default host=cvsup.hu.FreeBSD.org
*default base=/usr
*default prefix=/usr
*default release=cvs tag=RELENG_6
*default delete use-rel-suffix
*default compress
src-all


Then I traverse to /usr and execute this command:

cvsup /etc/sup.sys

After successful checkout, I do

cd /usr/src
make buildworld

This is what I get:

--
>>> Building an up-to-date make(1)
--
rm -f .depend
mkdep -f .depend -a-I/usr/src/usr.bin/make 
-DMAKE_VERSION=\"5200408120\" -D__FBSDID=__RCSID 
-DDEFSHELLNAME=\"sh\"  /usr/src/usr.bin/make/arch.c 
/usr/src/usr.bin/make/buf.c /usr/src/usr.bin/make/cond.c 
/usr/src/usr.bin/make/dir.c /usr/src/usr.bin/make/for.c 
/usr/src/usr.bin/make/hash.c /usr/src/usr.bin/make/hash_tables.c 
/usr/src/usr.bin/make/job.c /usr/src/usr.bin/make/lst.c 
/usr/src/usr.bin/make/main.c /usr/src/usr.bin/make/make.c 
/usr/src/usr.bin/make/parse.c /usr/src/usr.bin/make/proc.c 
/usr/src/usr.bin/make/shell.c /usr/src/usr.bin/make/str.c 
/usr/src/usr.bin/make/suff.c /usr/src/usr.bin/make/targ.c 
/usr/src/usr.bin/make/util.c /usr/src/usr.bin/make/var.c

In file included from /usr/src/usr.bin/make/arch.c:107:
/usr/src/usr.bin/make/globals.h:49: stdint.h: No such file or directory
In file included from /usr/src/usr.bin/make/cond.c:59:
/usr/src/usr.bin/make/globals.h:49: stdint.h: No such file or directory
In file included from /usr/src/usr.bin/make/dir.c:97:
/usr/src/usr.bin/make/globals.h:49: stdint.h: No such file or directory
In file included from /usr/src/usr.bin/make/for.c:55:
/usr/src/usr.bin/make/globals.h:49: stdint.h: No such file or directory
In file included from /usr/src/usr.bin/make/job.c:131:
/usr/src/usr.bin/make/globals.h:49: stdint.h: No such file or directory
In file included from /usr/src/usr.bin/make/main.c:84:
/usr/src/usr.bin/make/globals.h:49: stdint.h: No such file or directory
In file included from /usr/src/usr.bin/make/make.c:80:
/usr/src/usr.bin/make/globals.h:49: stdint.h: No such file or directory
In file included from /usr/src/usr.bin/make/parse.c:86:
/usr/src/usr.bin/make/globals.h:49: stdint.h: No such file or directory
In file included from /usr/src/usr.bin/make/str.c:49:
/usr/src/usr.bin/make/globals.h:49: stdint.h: No such file or directory
In file included from /usr/src/usr.bin/make/suff.c:102:
/usr/src/usr.bin/make/globals.h:49: stdint.h: No such file or directory
In file included from /usr/src/usr.bin/make/targ.c:84:
/usr/src/usr.bin/make/globals.h:49: stdint.h: No such file or directory
In file included from /usr/src/usr.bin/make/util.c:59:
/usr/src/usr.bin/make/globals.h:49: stdint.h: No such file or directory
In file included from /usr/src/usr.bin/make/var.c:96:
/usr/src/usr.bin/make/globals.h:49: stdint.h: No such file or directory
mkdep: compile failed
*** Error code 1

Stop in /usr/src/usr.bin/make.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.


What am I doing wrong?



Possibly, and I say possibly, and this isn't to be taken personally:
diving in before depth of the water. ;-)

At the very least, you should have run "mergemaster -p" before
make buildworld as this is a giant step (e.g. 4.8->4.11)..  Did you?

Probably the issue is simply the huge distance in time/development
between 4.8 and 6.0.  See below.


Additional questions:

- Can I upgrade from 4.8 to 6.0 in one step? Or should I upgrade to 
5.4 first?



Probably not.  A fresh install would have a benefit: you could re-format
your disks as UFS2.  However, some people have been able to get from 4.11 to
5.X with cvsup/buildworld, and the 5.X -> 6.X jump is "a piece of 
cake".  You

just have to do the jump from 4 to 5 carefully, and from a much more recent
4.X codebase than 4.8 .

- I have a .so file for 4.8. I do not have the sources, will it work 
on 6.0?




Dunno.  Probably not natively, but maybe; there is, of course, the 
COMPAT_FREEBSD4

stuff, too.


Thanks,

  Laszlo



Here's what I would try.  Upgrade to RELENG_4 via the cvsup/buildworld 
process,

read Bruce Mah's 4.11-5.X Migration Guide, and be sure and include
"mergemaster -p" in your buildworld cycle.

HTH,

Kevin Kinsey

--
In the next world, you're on your own.


___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails from 5.4->6.0

[Anthony Philipp]

On Mon, Dec 05, 2005 at 08:12:50AM -1000, 
Robert Marella <[EMAIL PROTECTED]> wrote:
> Hello Anthony
> 
> Did you remove everything under the /usr/obj directory before make
> buildworld?
> 
> Good luck
> Robert

I now have, and that fixed the problem. Thanks for you help!

Anthony
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails from 5.4->6.0

[Robert Marella]

On Mon, 5 Dec 2005 03:11:46 -0600
Anthony Philipp <[EMAIL PROTECTED]> wrote:

> Hello,
> 
> I am trying to upgrade from 5.4 to 6.0. I ran cvsup and tried a make
> buildworld, but it failed. So I checked out the handbook and noticed
> that it mentioned I should run a mergemaster -p first. So I did that
> and I still failed. Here are the last few lines:
> 
> : undefined reference to `Buf_AddByte'
> var.o(.text+0x2b5e): In function `Var_Dump':
> : undefined reference to `Buf_Data'
> var.o(.text+0x2ba2): In function `Var_Dump':
> : undefined reference to `Buf_Data'
> var.o(.text+0x2c31): In function `Var_Print':
> : undefined reference to `Buf_Peel'
> *** Error code 1
> 
> Stop in /usr/src/usr.bin/make.
> *** Error code 1
> 
> Stop in /usr/src.
> *** Error code 1
> 
> Stop in /usr/src.
> 
> I've searched the archives and couldn't find anything relating to
> this, but hopefully someone knows the answer. Also I have have done
> minor version changes before, but never a major one, so maybe I am
> just missing a key step.
> 
> Thanks in advance!
> 
> Anthony Philipp
Hello Anthony

Did you remove everything under the /usr/obj directory before make
buildworld?

Good luck
Robert
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails from 5.4->6.0

[Benjamin Sobotta]

On Monday 05 December 2005 09:51, Anthony Philipp wrote:
> On Mon, Dec 05, 2005 at 10:41:15AM +,
>
> Benjamin Sobotta <[EMAIL PROTECTED]> wrote:
> > Hey!
> >
> > You might want to have a look at
> >
> > http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/makeworld.html
> >
> > especially subsection
> >
> > 20.4.16.6. What do I do if something goes wrong
>
> I will try this and get back do you. Thanks for the tip.
>
> > Further, I believe that you run mergemaster -p before make installworld.
> > Not buildworld, after all I believe buildworld has to finish so you know
> > what to merge in the first place. So if buildworld fails, that has
> > nothing to do with mergemaster.
>
> I had looked at the handbook and saw this:
>
> Note: There are a few rare cases when an extra run of mergemaster -p is
> needed before the buildworld step. These are described in UPDATING. In
> general, though, you can safely omit this step if you are not updating
> across one or more major FreeBSD versions.
>
> This is the only reason I tried it.
>
> Thanks for the response!
>
> Anthony

Interesting! I didn't know that.

Thank you!

>
> > On Monday 05 December 2005 09:11, Anthony Philipp wrote:
> > > Hello,
> > >
> > > I am trying to upgrade from 5.4 to 6.0. I ran cvsup and tried a make
> > > buildworld, but it failed. So I checked out the handbook and noticed
> > > that it mentioned I should run a mergemaster -p first. So I did that
> > >
> > > and I still failed. Here are the last few lines:
> > > : undefined reference to `Buf_AddByte'
> > >
> > > var.o(.text+0x2b5e): In function `Var_Dump':
> > > : undefined reference to `Buf_Data'
> > >
> > > var.o(.text+0x2ba2): In function `Var_Dump':
> > > : undefined reference to `Buf_Data'
> > >
> > > var.o(.text+0x2c31): In function `Var_Print':
> > > : undefined reference to `Buf_Peel'
> > >
> > > *** Error code 1
> > >
> > > Stop in /usr/src/usr.bin/make.
> > > *** Error code 1
> > >
> > > Stop in /usr/src.
> > > *** Error code 1
> > >
> > > Stop in /usr/src.
> > >
> > > I've searched the archives and couldn't find anything relating to
> > > this, but hopefully someone knows the answer. Also I have have done
> > > minor version changes before, but never a major one, so maybe I am
> > > just missing a key step.
> > >
> > > Thanks in advance!
> > >
> > > Anthony Philipp
> > > ___
> > > freebsd-questions@freebsd.org mailing list
> > > http://lists.freebsd.org/mailman/listinfo/freebsd-questions
> > > To unsubscribe, send any mail to
> > > "[EMAIL PROTECTED]"
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails from 5.4->6.0

[Anthony Philipp]

On Mon, Dec 05, 2005 at 10:41:15AM +, 
Benjamin Sobotta <[EMAIL PROTECTED]> wrote:

> Hey!
> 
> You might want to have a look at
> 
> http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/makeworld.html
> 
> especially subsection 
> 
> 20.4.16.6. What do I do if something goes wrong
> 

I will try this and get back do you. Thanks for the tip.

> Further, I believe that you run mergemaster -p before make installworld. Not 
> buildworld, after all I believe buildworld has to finish so you know what to 
> merge in the first place. So if buildworld fails, that has nothing to do with 
> mergemaster.

I had looked at the handbook and saw this:

Note: There are a few rare cases when an extra run of mergemaster -p is needed
before the buildworld step. These are described in UPDATING. In general,
though, you can safely omit this step if you are not updating across one or
more major FreeBSD versions.

This is the only reason I tried it.

Thanks for the response!

Anthony

> On Monday 05 December 2005 09:11, Anthony Philipp wrote:
> > Hello,
> >
> > I am trying to upgrade from 5.4 to 6.0. I ran cvsup and tried a make
> > buildworld, but it failed. So I checked out the handbook and noticed
> > that it mentioned I should run a mergemaster -p first. So I did that
> >
> > and I still failed. Here are the last few lines:
> > : undefined reference to `Buf_AddByte'
> >
> > var.o(.text+0x2b5e): In function `Var_Dump':
> > : undefined reference to `Buf_Data'
> >
> > var.o(.text+0x2ba2): In function `Var_Dump':
> > : undefined reference to `Buf_Data'
> >
> > var.o(.text+0x2c31): In function `Var_Print':
> > : undefined reference to `Buf_Peel'
> >
> > *** Error code 1
> >
> > Stop in /usr/src/usr.bin/make.
> > *** Error code 1
> >
> > Stop in /usr/src.
> > *** Error code 1
> >
> > Stop in /usr/src.
> >
> > I've searched the archives and couldn't find anything relating to
> > this, but hopefully someone knows the answer. Also I have have done
> > minor version changes before, but never a major one, so maybe I am
> > just missing a key step.
> >
> > Thanks in advance!
> >
> > Anthony Philipp
> > ___
> > freebsd-questions@freebsd.org mailing list
> > http://lists.freebsd.org/mailman/listinfo/freebsd-questions
> > To unsubscribe, send any mail to
> > "[EMAIL PROTECTED]"
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails from 5.4->6.0

[Benjamin Sobotta]

Hey!

You might want to have a look at

http://www.freebsd.org/doc/en_US.ISO8859-1/books/handbook/makeworld.html

especially subsection 

20.4.16.6. What do I do if something goes wrong

Further, I believe that you run mergemaster -p before make installworld. Not 
buildworld, after all I believe buildworld has to finish so you know what to 
merge in the first place. So if buildworld fails, that has nothing to do with 
mergemaster.

HTH,

Ben

On Monday 05 December 2005 09:11, Anthony Philipp wrote:
> Hello,
>
> I am trying to upgrade from 5.4 to 6.0. I ran cvsup and tried a make
> buildworld, but it failed. So I checked out the handbook and noticed
> that it mentioned I should run a mergemaster -p first. So I did that
>
> and I still failed. Here are the last few lines:
> : undefined reference to `Buf_AddByte'
>
> var.o(.text+0x2b5e): In function `Var_Dump':
> : undefined reference to `Buf_Data'
>
> var.o(.text+0x2ba2): In function `Var_Dump':
> : undefined reference to `Buf_Data'
>
> var.o(.text+0x2c31): In function `Var_Print':
> : undefined reference to `Buf_Peel'
>
> *** Error code 1
>
> Stop in /usr/src/usr.bin/make.
> *** Error code 1
>
> Stop in /usr/src.
> *** Error code 1
>
> Stop in /usr/src.
>
> I've searched the archives and couldn't find anything relating to
> this, but hopefully someone knows the answer. Also I have have done
> minor version changes before, but never a major one, so maybe I am
> just missing a key step.
>
> Thanks in advance!
>
> Anthony Philipp
> ___
> freebsd-questions@freebsd.org mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-questions
> To unsubscribe, send any mail to
> "[EMAIL PROTECTED]"
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails on step 1.2 and questions

[Garrett Cooper]

On Sep 20, 2005, at 7:01 PM, Chris wrote:



A few things - If you are using CURRENT (anything) you have posted to
the wrong list - try FreeBSD-CURRENT.


K.


If your a newb (as you seem to be), you should be using either
5.4-RELEASE or 4.11-RELEASE


Meh. I just want to get my hands dirty a bit and learn more  
about FreeBSD to the point where I feel confident about the OS inside  
out.



If you want to use the patch branch of 5.4, your cvs tag will be:
*default release=cvs tag=RELENG_5_4

If you want the 4.11 patch branch:
*default release=cvs tag=RELENG_4_11

I have been using FreeBSD since 2.2.8 - and only recently have I felt
comfy enough to use CURRENT.

For you - admirable, yes - foolish, you betcha. Stick to something  
solid

until you know what you're doing...


Still, one must make mistakes in order to learn. Besides, my  
important data is on a separate partition just in case anything goes  
awry, and I do have a LiveCD on hand.


Thank you though, I do appreciate it.
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails on step 1.2 and questions

[Chris]

Garrett Cooper wrote:
> Hello,
> I'm building my OS from source from the first time, and I  realize
> that I must have missed a step in configuring it properly. I  just hate
> treading through a lot of documentation when there isn't a  really nice
> search feature involved, so if someone can give me the  steps they
> used-or even better-the section in a handbook or FAQ on  how to do the
> buildworld properly, I would appreciate it.
> Also, I was wondering if there was any specific syntax I should  use
> for specifying the version of FreeBSD that I want to use, ie 6.0-
> CURRENT vs 7.0-CURRENT (because I think that 7.0 has just come out,  but
> I may be wrong because I haven't been keeping completely up to  date).
> All that I have done so far is CVSUP'ed my files into /usr and 
> that's about it. I am not sure where to go from here.
> I greatly appreciate any and all help provided to me.
> Thanks,
> -Garrett
> 
> Command I use for building the system:
> 
> [EMAIL PROTECTED] cd /usr/src/ && make update && make buildworld &&  make
> kernel
> 
> Here is the error that comes up:

... Snip

A few things - If you are using CURRENT (anything) you have posted to
the wrong list - try FreeBSD-CURRENT.

If your a newb (as you seem to be), you should be using either
5.4-RELEASE or 4.11-RELEASE

If you want to use the patch branch of 5.4, your cvs tag will be:
*default release=cvs tag=RELENG_5_4

If you want the 4.11 patch branch:
*default release=cvs tag=RELENG_4_11

I have been using FreeBSD since 2.2.8 - and only recently have I felt
comfy enough to use CURRENT.

For you - admirable, yes - foolish, you betcha. Stick to something solid
until you know what you're doing...

-- 
Best regards,
Chris

1) You can't win
2) You can't break even
3) You can't even quit the game
___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails in openssl/colldef [SOLVED]

[Alex Zbyslaw]

Marius Korsmo wrote:

That does not look like an *old* version, it looks like a completely 
different file.  Did you install anything not from ports that might 
have overwritten it?  Did you try and install a port into a target 
hierarchy that /usr/local?  (The same copyright header as the real 
5.4 one exists in err.h from 4.11).


Operating system bugs are rare compared to user errors, I'm afraid.

--Alex


As far as I can remember, I have not installed anything that didn't 
come from
the ports collection. I have only used the ports, and always installed 
them to

the default directory.

I did not mean that this was a OS bug, and it would be really strange 
if it was
since I can't find any other person that has had the same problem. 
There must
be an application that overwrote my err.h. I've tried to provoke this 
error
again, but I can't. I tried to install all the ports on another server 
(until i
ran out of space), but the error did not occur. But with that said, 
there was 20
ports I didn't have enough space to install, so the problem might lie 
there.


I might be the problem as well, maybe I did something I shouldn't have 
done. But
I can't really see what that should be :) Anyway, the solution is here 
and I
hope it can help other people that end up with the same problem (if 
that ever

happens).


Then put it down as a mystery :-)

If you are worried about it happening again, then just mail yourself the 
head of the file every day from cron.  If it does change, at least 
you'll know better when it happened and might be better able to figure 
out what caused it.


You could also force re-installation of your current ports on the same 
machine, but that might be more time than you care to spend.


--Alex

PS I suppose the clever way would be to head the file, egrep for 
Copyright then egrep -v the correct date (2003) since that will be empty 
when the file is correct.  So you should be able to arrange for the mail 
to arrive only when there is a problem (which might be never again!).


___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails in openssl/colldef [SOLVED]

[Marius Korsmo]

Quoting Alex Zbyslaw <[EMAIL PROTECTED]>:


Marius Korsmo wrote:

It took two weeks to fix this problem, and therefore I do have a few 
questions.


The error turned out to be in err.h. The file located in /usr/include was
totally different from the one located in /usr/src/include.

< /*-
<  * Copyright (c) 1993
<  *The Regents of the University of California.  All rights reserved.
---


/* crypto/err/err.h */
/* Copyright (C) 1995-1998 Eric Young ([EMAIL PROTECTED])
* All rights reserved.



You can see the entire diff at http://pastebin.com/322918

My question is: I deleted /usr/src, and I cvsup'ed everything. When 
I do a make

buildworld, why does not the new err.h get copied from /usr/src/include to
/usr/include? This would have solved my problem two weeks ago :)


build = recompile things
install = put them where they belong in the filesystem

buildworld doesn't install the file because it's not supposed to.


Does err.h get copied only when you do a make installworld?


I would expect so.  Why don't you try it?

Another question, why on earth did I have an old version of err.h? I 
was running

5.4 RELEASE, and it was installed from an ISO downloaded at FreeBSD.org

That does not look like an *old* version, it looks like a completely 
different file.  Did you install anything not from ports that might 
have overwritten it?  Did you try and install a port into a target 
hierarchy that /usr/local?  (The same copyright header as the real 
5.4 one exists in err.h from 4.11).


Operating system bugs are rare compared to user errors, I'm afraid.

--Alex



As far as I can remember, I have not installed anything that didn't come from
the ports collection. I have only used the ports, and always installed them to
the default directory.

I did not mean that this was a OS bug, and it would be really strange 
if it was

since I can't find any other person that has had the same problem. There must
be an application that overwrote my err.h. I've tried to provoke this error
again, but I can't. I tried to install all the ports on another server 
(until i
ran out of space), but the error did not occur. But with that said, 
there was 20

ports I didn't have enough space to install, so the problem might lie there.

I might be the problem as well, maybe I did something I shouldn't have 
done. But

I can't really see what that should be :) Anyway, the solution is here and I
hope it can help other people that end up with the same problem (if that ever
happens).

Marius




This message was sent using IMP, the Internet Messaging Program.

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails in openssl/colldef [SOLVED]

[Alex Zbyslaw]

Marius Korsmo wrote:


It took two weeks to fix this problem, and therefore I do have a few questions.

The error turned out to be in err.h. The file located in /usr/include was
totally different from the one located in /usr/src/include.

< /*-
<  * Copyright (c) 1993
<  *The Regents of the University of California.  All rights reserved.
---
 


/* crypto/err/err.h */
/* Copyright (C) 1995-1998 Eric Young ([EMAIL PROTECTED])
* All rights reserved.
   



You can see the entire diff at http://pastebin.com/322918

My question is: I deleted /usr/src, and I cvsup'ed everything. When I do a make
buildworld, why does not the new err.h get copied from /usr/src/include to
/usr/include? This would have solved my problem two weeks ago :)
 


build = recompile things
install = put them where they belong in the filesystem

buildworld doesn't install the file because it's not supposed to.


Does err.h get copied only when you do a make installworld?
 


I would expect so.  Why don't you try it?


Another question, why on earth did I have an old version of err.h? I was running
5.4 RELEASE, and it was installed from an ISO downloaded at FreeBSD.org
 

That does not look like an *old* version, it looks like a completely 
different file.  Did you install anything not from ports that might have 
overwritten it?  Did you try and install a port into a target hierarchy 
that /usr/local?  (The same copyright header as the real 5.4 one exists 
in err.h from 4.11).


Operating system bugs are rare compared to user errors, I'm afraid.

--Alex

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails

[Valerio Daelli]

Yes.  You're using unsupported optimizations.

Please read very carefully the comments above CFLAGS in

/usr/share/examples/etc/make.conf

- Giorgos



Thanks a lot! I'll use a safer optimization.


Valeiro


___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails

[Giorgos Keramidas]

On 2005-06-22 14:38, Valerio Daelli <[EMAIL PROTECTED]> wrote:
>
> Then I started to make buildworld. But it fails with this error:
> __
>
> cc -O3 -mtune=pentium4 -march=pentium4 -pipe -funroll-loops -ffast-math
> -march=pentium4 -I/usr/src/lib/libc/include
> -I/usr/src/lib/libc/../../include -I/usr/src/lib/libc/i386
> -D__DBINTERFACE_PRIVATE -I/usr/src/lib/libc/../../contrib/gdtoa -DINET6
> -I/usr/obj/usr/src/lib/libc -DPOSIX_MISTAKE -I/usr/src/lib/libc/locale
> -DBROKEN_DES -DPORTMAP -DDES_BUILTIN -I/usr/src/lib/libc/rpc -DYP -DHESIOD
> -Wsystem-headers -Werror -Wall -Wno-format-y2k -Wno-uninitialized -c
> /usr/src/lib/libc/net/res_debug.c
> In file included from /usr/src/lib/libc/net/res_debug.c:786:
> /usr/src/lib/libc/net/res_debug.c: In function `__loc_aton':
> /usr/src/lib/libc/net/res_debug.c:576: warning: passing arg 1 of
> `precsize_aton' from incompatible pointer type
> In file included from /usr/src/lib/libc/net/res_debug.c:797:
> /usr/src/lib/libc/net/res_debug.c:576: warning: passing arg 1 of
> `precsize_aton' from incompatible pointer type
> In file included from /usr/src/lib/libc/net/res_debug.c:808:
> /usr/src/lib/libc/net/res_debug.c:576: warning: passing arg 1 of
> `precsize_aton' from incompatible pointer type
>
> __
>
> Any idea what's happened?

Yes.  You're using unsupported optimizations.

Please read very carefully the comments above CFLAGS in

/usr/share/examples/etc/make.conf

- Giorgos

___
freebsd-questions@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

RE: make buildworld fails and a reboot?

[Eric F Crist]

-Original Message-
From: Chuck Swiger [mailto:[EMAIL PROTECTED]
Sent: Wednesday, February 18, 2004 10:50 PM
To: [EMAIL PROTECTED]
Cc: FreeBSD questions List
Subject: Re: make buildworld fails and a reboot?


Eric F Crist wrote:
> Hey all.  I figured out why my system doesn't seem to be working.  For

> some
> reason, make world failed and the system rebooted.  Now, I can't cvsup
any
> new sources.  How do I recover from this?

"make world" doesn't always result in a working system, which is why the

procedure in UPDATING has many more steps (but which should be safe[r]).

You haven't provided an error message as to what happens when you try to

cvsup; what happens when you try?  For that matter, you haven't told us
which
version of FreeBSD you are running and what branch of the sources you
were
building world with...

> Please tell me I don't have to reinstall from scratch.

Oh, it's absolutely certain that you could recover in other ways, but it
also
might be the case that doing a clean reinstall is faster than any of
these
alternatives.

--
-Chuck

Chuck, sorry about the lack of information.  Here's the dilly-oh.

I've got a FreeBSD 5.2 system.  I've figured out that cvsup doesn't work
because nslookup doesn't exist on the system.  I figure I could just
copy the binary from my 4.9 system and try from there.  I use the
RELENG_5_2 tag for CVSUP.  I'm thinking this is wrong.  Bah.  In the
mean time, I'm going to try copying the nslookup exec from my 4.9 laptop
and see if I can't get cvsup to work like that.  Dig works just fine,
which is why I think this method could work.

If all else fails, I'll just start a rebuild tonight.

Thanks,

Eric F Crist



___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails and a reboot?

[Chuck Swiger]

Eric F Crist wrote:
Hey all.  I figured out why my system doesn't seem to be working.  For some 
reason, make world failed and the system rebooted.  Now, I can't cvsup any 
new sources.  How do I recover from this?  
"make world" doesn't always result in a working system, which is why the 
procedure in UPDATING has many more steps (but which should be safe[r]).

You haven't provided an error message as to what happens when you try to 
cvsup; what happens when you try?  For that matter, you haven't told us which 
version of FreeBSD you are running and what branch of the sources you were 
building world with...

Please tell me I don't have to reinstall from scratch.
Oh, it's absolutely certain that you could recover in other ways, but it also 
might be the case that doing a clean reinstall is faster than any of these 
alternatives.

--
-Chuck
___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails

[Kent Stewart]

On Monday 19 January 2004 01:56 pm, Nick Rogness wrote:
> On Mon, 19 Jan 2004, Kent Stewart wrote:
> > On Monday 19 January 2004 11:55 am, Nick Rogness wrote:
> > > On Mon, 19 Jan 2004, Kevin D. Kinsey, DaleCo, S.P. wrote:
> > > > Nick Rogness wrote:
> > > > >I cvsup'd just a few minutes ago and tried a make buildworld. 
> > > > > Any hints as to why I'm failing out?  Details below:
> > > > >
> > > > >/usr/src/gnu/usr.bin/binutils/objdump/../../../../contrib/binu
> > > > >tils /binutils/objdump.c:1553: unterminated character constant
> > > > > mkdep: compile failed
> > > >
> > > > Yes  ;-)
> > > >
> > > > There is an unterminated character constant in line 1553 of
> > > > "objdump.c" in the directory given by the error message.
> > > >
> > > > Suppose there's a semicolon missing there?
> > > >
> > > > However, as Kent Stewart has noticed, it's a real good idea to
> > > > make sure you're building the correct src tree first.  AFAIK,
> > > > 5.x is using different build tools.
> > >
> > >   Forgot to mention, this is the 4 branch.
> >
> > All of the RELENG_4 have the following line at objdump.c:1553
> >   fatal (_("Can't use supplied machine %s"), machine);
> >
> > What do you have? That line has a matched pair of "'s. It sounds
> > like you have a corrupted source. The usual fix is to delete it and
> > recvsup. Since, you still really haven't told us what you cvsup'ed,
> > you should include "in-line" a copy of your cvsup file, so that we
> > know what we are trying to diagnose :).
>
>   I cvsup'd the whole system from the stable-supfile
>   provided in /usr/share/examples/cvsup.  That is what was ment by
>   'the 4 branch' above.  The entire source tree (RELENG_4) like I've
>   done thousands of times.

Ah, but you could have been following RELENG_4_9, which is the security 
version of 4.9-release, or RELENG_4_9_0_RELEASE and etc.

>
>   It appears to be dying in another place this time.  I think it is
>   probably memory or some other hardware problem.  Anybody had
>   this experience with hardware and building?

That is a characteristic of bad hardware. FWIW, just in case you were 
following RELENG_4, I cvsup'ed RELENG_4 and successfully did a 
buildworld. 

Kent

-- 
Kent Stewart
Richland, WA

http://users.owt.com/kstewart/index.html

___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails

[Nick Rogness]

On Mon, 19 Jan 2004, Kent Stewart wrote:

> On Monday 19 January 2004 11:55 am, Nick Rogness wrote:
> > On Mon, 19 Jan 2004, Kevin D. Kinsey, DaleCo, S.P. wrote:
> > > Nick Rogness wrote:
> > > >I cvsup'd just a few minutes ago and tried a make buildworld.  Any
> > > > hints as to why I'm failing out?  Details below:
> > > >
> > > >/usr/src/gnu/usr.bin/binutils/objdump/../../../../contrib/binutils
> > > >/binutils/objdump.c:1553: unterminated character constant
> > > >mkdep: compile failed
> > >
> > > Yes  ;-)
> > >
> > > There is an unterminated character constant in line 1553 of
> > > "objdump.c" in the directory given by the error message.
> > >
> > > Suppose there's a semicolon missing there?
> > >
> > > However, as Kent Stewart has noticed, it's a real good idea to make
> > > sure you're building the correct src tree first.  AFAIK, 5.x is
> > > using different build tools.
> >
> > Forgot to mention, this is the 4 branch.
> 
> All of the RELENG_4 have the following line at objdump.c:1553
>   fatal (_("Can't use supplied machine %s"), machine);
> 
> What do you have? That line has a matched pair of "'s. It sounds like
> you have a corrupted source. The usual fix is to delete it and recvsup.  
> Since, you still really haven't told us what you cvsup'ed, you should
> include "in-line" a copy of your cvsup file, so that we know what we are
> trying to diagnose :).

I cvsup'd the whole system from the stable-supfile 
provided in /usr/share/examples/cvsup.  That is what was ment by 
'the 4 branch' above.  The entire source tree (RELENG_4) like I've 
done thousands of times.

It appears to be dying in another place this time.  I think it is 
probably memory or some other hardware problem.  Anybody had 
this experience with hardware and building?

-- 
Nick Rogness <[EMAIL PROTECTED]>
-
  How many people here have telekenetic powers? Raise my hand.
-Emo Philips
 

___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails

[Kent Stewart]

On Monday 19 January 2004 11:55 am, Nick Rogness wrote:
> On Mon, 19 Jan 2004, Kevin D. Kinsey, DaleCo, S.P. wrote:
> > Nick Rogness wrote:
> > >I cvsup'd just a few minutes ago and tried a make buildworld.  Any
> > > hints as to why I'm failing out?  Details below:
> > >
> > >/usr/src/gnu/usr.bin/binutils/objdump/../../../../contrib/binutils
> > >/binutils/objdump.c:1553: unterminated character constant
> > >mkdep: compile failed
> >
> > Yes  ;-)
> >
> > There is an unterminated character constant in line 1553 of
> > "objdump.c" in the directory given by the error message.
> >
> > Suppose there's a semicolon missing there?
> >
> > However, as Kent Stewart has noticed, it's a real good idea to make
> > sure you're building the correct src tree first.  AFAIK, 5.x is
> > using different build tools.
>
>   Forgot to mention, this is the 4 branch.

All of the RELENG_4 have the following line at objdump.c:1553
  fatal (_("Can't use supplied machine %s"), machine);

What do you have? That line has a matched pair of "'s. It sounds like 
you have a corrupted source. The usual fix is to delete it and recvsup. 
Since, you still really haven't told us what you cvsup'ed, you should 
include "in-line" a copy of your cvsup file, so that we know what we 
are trying to diagnose :).

Kent

-- 
Kent Stewart
Richland, WA

http://users.owt.com/kstewart/index.html

___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails

[Nick Rogness]

On Mon, 19 Jan 2004, Kevin D. Kinsey, DaleCo, S.P. wrote:

> Nick Rogness wrote:
> 
> >I cvsup'd just a few minutes ago and tried a make buildworld.  Any hints 
> >as to why I'm failing out?  Details below:
> >
> >/usr/src/gnu/usr.bin/binutils/objdump/../../../../contrib/binutils/binutils/objdump.c:1553:
> > 
> >unterminated character constant
> >mkdep: compile failed
> >  
> >
> 
> Yes  ;-)
> 
> There is an unterminated character constant in line 1553 of "objdump.c"
> in the directory given by the error message.
> 
> Suppose there's a semicolon missing there?
> 
> However, as Kent Stewart has noticed, it's a real good idea to make sure
> you're building the correct src tree first.  AFAIK, 5.x is using
> different build tools.

Forgot to mention, this is the 4 branch.



-- 
Nick Rogness <[EMAIL PROTECTED]>
-
  How many people here have telekenetic powers? Raise my hand.
-Emo Philips
 

___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails

[Kevin D. Kinsey, DaleCo, S.P.]

Nick Rogness wrote:

I cvsup'd just a few minutes ago and tried a make buildworld.  Any hints 
as to why I'm failing out?  Details below:

/usr/src/gnu/usr.bin/binutils/objdump/../../../../contrib/binutils/binutils/objdump.c:1553: 
unterminated character constant
mkdep: compile failed
 

Yes  ;-)

There is an unterminated
character constant in line 1553 of "objdump.c" in the
directory given by the error message.
Suppose there's a semicolon missing there?

However, as Kent Stewart has noticed, it's a real good
idea to make sure you're building the correct src tree
first.  AFAIK, 5.x is using different build tools.
Take a look at the [EMAIL PROTECTED], as you're running 4.x,
and see if anyone else is having this trouble.  AAMOF,
it'd probably be good to post this there.
HTH,

Kevin Kinsey

___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails

[Kent Stewart]

On Monday 19 January 2004 10:12 am, Nick Rogness wrote:
> I cvsup'd just a few minutes ago and tried a make buildworld.  Any
> hints as to why I'm failing out?  Details below:

For starters, what did you cvsup?

Kent

>
>
> # uname -a
> FreeBSD cody.jharris.com 4.9-RELEASE FreeBSD 4.9-RELEASE #0: Mon Oct
> 27 17:51:09 GMT 2003
> [EMAIL PROTECTED]:/usr/obj/usr/src/sys/GENERIC  i386
>
> # cd /usr/src && make buildworld
>
> 
> rm -f .depend
> mkdep -f .depend -a-D_GNU_SOURCE -I.
> -I/usr/src/gnu/usr.bin/binutils/objdump
> -I/usr/src/gnu/usr.bin/binutils/objdump/../libbfd/i386
> -I/usr/src/gnu/usr.bin/binutils/objdump/../../../../contrib/binutils/
>include -I/usr/src/gnu/usr.bin/binutils/objdump/../libbinutils
> -I/usr/src/gnu/usr.bin/binutils/objdump/../../../../contrib/binutils/
>binutils -DBFD_VERSION_STRING=\""2.12.1 [FreeBSD] 2002-07-20"\"
> -D__FBSDID=__RCSID
> /usr/src/gnu/usr.bin/binutils/objdump/../../../../contrib/binutils/bi
>nutils/objdump.c
> /usr/src/gnu/usr.bin/binutils/objdump/../../../../contrib/binutils/bi
>nutils/prdbg.c
> /usr/src/gnu/usr.bin/binutils/objdump/../../../../contrib/binutils/bi
>nutils/objdump.c:1553: unterminated character constant
> mkdep: compile failed
> *** Error code 1
>
> Stop in /usr/src/gnu/usr.bin/binutils/objdump.
> *** Error code 1
>
> Stop in /usr/src/gnu/usr.bin/binutils.
> *** Error code 1
>
> Stop in /usr/src.
> *** Error code 1
>
> Stop in /usr/src.
> *** Error code 1
>
> Stop in /usr/src.

-- 
Kent Stewart
Richland, WA

http://users.owt.com/kstewart/index.html

___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: make buildworld fails

[Kris Kennaway]

On Mon, Jun 30, 2003 at 04:12:14PM +0200, Maarten de Vries wrote:
> Hi,
> 
> I'm trying to upgrade an old P166 from 5.0 to 5.1-R using freshly cvsup'ed 
> sources. After a couple of hours (heh...) it suddenly halts:
> 
> /usr/obj/usr/src/i386/usr/lib/libtermlib_p.a -> libncurses_p.a
> /usr/obj/usr/src/i386/usr/lib/libmytinfo_p.a -> libncurses_p.a
> /usr/obj/usr/src/i386/usr/lib/libtinfo_p.a -> libncurses_p.a
> 1 error
> *** Error code 2
> 1 error
> *** Error code 2
> 1 error
> *** Error code 2
> 1 error
> 
> Bummer. Anything I can do to fix this?

You could start by showing us the actual error, instead of some random
segment of the logs that doesn't show the command that failed :-)
Re-run the make without a -j option, and try again.

Kris


pgp0.pgp
Description: PGP signature

Re: Make buildworld fails (many times)

[Kris Kennaway]

On Sun, Jun 22, 2003 at 05:25:33PM -0700, Charlie Schluting wrote:
> 
> On Sun, 22 Jun 2003, Kris Kennaway wrote:
> 
> > On Sun, Jun 22, 2003 at 02:15:46PM -0700, Charlie Schluting wrote:
> >
> > > Other info:
> > > AMD K6-2 450MHz, 192MB.
> > > Current install: 5.0-RELEASE-p7
> >
> > This is likely to be a hardware issue.  AMD k6's are very sensitive to
> > temperature and require lots of CPU cooling.  Or maybe you have bad
> > RAM or other failing hardware.
> >
> > Kris
> >
> 
> Well, I tried to run make again, and it stopped in another place.
> Unfortunately, it wasn't a signal 11:

Doesn't matter, the random nature of the crashes is a dead give-away.

Kris


pgp0.pgp
Description: PGP signature

Re: Make buildworld fails (many times)

[Charlie Schluting]

Kent Stewart wrote:

It doesn't matter. My rule is that if you are the only one with problems, it 
is your computer that has the problem. There haven't been a number of 
complaints of buildworld failures in -current, which is where they would be 
if 5.x was failing.

In the past, IIRC, many of the K6 450 users had to underclock them to get them 
stable. 

Kent

Thanks for the help. I also found some info in the Sig11 doc about 
having to run these POS procs @ 400MHz (still more than enough for my 
NAT router / firewall). Oh well, that will have to wait for another 
weekend :)
--Charlie

___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: Make buildworld fails (many times)

[Kent Stewart]

On Sunday 22 June 2003 05:25 pm, Charlie Schluting wrote:
> On Sun, 22 Jun 2003, Kris Kennaway wrote:
> > On Sun, Jun 22, 2003 at 02:15:46PM -0700, Charlie Schluting wrote:
> > > Other info:
> > > AMD K6-2 450MHz, 192MB.
> > > Current install: 5.0-RELEASE-p7
> >
> > This is likely to be a hardware issue.  AMD k6's are very sensitive to
> > temperature and require lots of CPU cooling.  Or maybe you have bad
> > RAM or other failing hardware.
> >
> > Kris
>
> Well, I tried to run make again, and it stopped in another place.
> Unfortunately, it wasn't a signal 11:

It doesn't matter. My rule is that if you are the only one with problems, it 
is your computer that has the problem. There haven't been a number of 
complaints of buildworld failures in -current, which is where they would be 
if 5.x was failing.

In the past, IIRC, many of the K6 450 users had to underclock them to get them 
stable. 

Kent

> cc -fpic -DPIC -O -pipe -mcpu=pentiumpro -DHAVE_CONFIG_H=1
> -DSTDC_HEADERS=1 -DHAVE_TEMPNAM=1 -DNO_EOF_CHAR_CHECK=1 -DRETSIGTYPE=void
> -DSkip_f2c_Undefs=1 -DIEEE_drem=1 -DAllow_TYQUAD -DPedantic
> -I/usr/src/gnu/lib/libg2c/../../../contrib/libf2c
> -I/usr/src/gnu/lib/libg2c/../../../contrib/libf2c/libF77
> -I/usr/src/gnu/lib/libg2c/../../../contrib/libf2c/libI77
> -I/usr/src/gnu/lib/libg2c/../../../contrib/libf2c/libU77
> -I/usr/src/gnu/lib/libg2c  -c -DLperror -o perror.So
> /usr/src/gnu/lib/libg2c/../../../contrib/libf2c/f2cext.c
> *** Signal 10
>
> Stop in /usr/src/gnu/lib/libg2c.
> *** Error code 1
>
>
> Does this still sound like hardware issues?
> ___
> [EMAIL PROTECTED] mailing list
> http://lists.freebsd.org/mailman/listinfo/freebsd-questions
> To unsubscribe, send any mail to
> "[EMAIL PROTECTED]"

-- 
Kent Stewart
Richland, WA

http://users.owt.com/kstewart/index.html

___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: Make buildworld fails (many times)

[Charlie Schluting]

On Sun, 22 Jun 2003, Kris Kennaway wrote:

> On Sun, Jun 22, 2003 at 02:15:46PM -0700, Charlie Schluting wrote:
>
> > Other info:
> > AMD K6-2 450MHz, 192MB.
> > Current install: 5.0-RELEASE-p7
>
> This is likely to be a hardware issue.  AMD k6's are very sensitive to
> temperature and require lots of CPU cooling.  Or maybe you have bad
> RAM or other failing hardware.
>
> Kris
>

Well, I tried to run make again, and it stopped in another place.
Unfortunately, it wasn't a signal 11:
cc -fpic -DPIC -O -pipe -mcpu=pentiumpro -DHAVE_CONFIG_H=1
-DSTDC_HEADERS=1 -DHAVE_TEMPNAM=1 -DNO_EOF_CHAR_CHECK=1 -DRETSIGTYPE=void
-DSkip_f2c_Undefs=1 -DIEEE_drem=1 -DAllow_TYQUAD -DPedantic
-I/usr/src/gnu/lib/libg2c/../../../contrib/libf2c
-I/usr/src/gnu/lib/libg2c/../../../contrib/libf2c/libF77
-I/usr/src/gnu/lib/libg2c/../../../contrib/libf2c/libI77
-I/usr/src/gnu/lib/libg2c/../../../contrib/libf2c/libU77
-I/usr/src/gnu/lib/libg2c  -c -DLperror -o perror.So
/usr/src/gnu/lib/libg2c/../../../contrib/libf2c/f2cext.c
*** Signal 10

Stop in /usr/src/gnu/lib/libg2c.
*** Error code 1


Does this still sound like hardware issues?
___
[EMAIL PROTECTED] mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-questions
To unsubscribe, send any mail to "[EMAIL PROTECTED]"

Re: Make buildworld fails (many times)

[Kris Kennaway]

On Sun, Jun 22, 2003 at 02:15:46PM -0700, Charlie Schluting wrote:

> Other info:
> AMD K6-2 450MHz, 192MB.
> Current install: 5.0-RELEASE-p7

This is likely to be a hardware issue.  AMD k6's are very sensitive to
temperature and require lots of CPU cooling.  Or maybe you have bad
RAM or other failing hardware.

Kris


pgp0.pgp
Description: PGP signature

Re: Make buildworld fails for 5.0

[taxman]

On Tuesday 04 March 2003 09:40 pm, Mike Loiterman wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> I'm getting the following error in my:
>
> 'make buildworld'
>
> I'm currently running 5.0-RELEASE #0 and I've cvsup'ed to '.'.

So then this belongs on -current.  See 
http://www.FreeBSD.org/doc/en_US.ISO8859-1/books/handbook/current-stable.html
They'll help you out more there,

Tim

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

RE: Make buildworld fails for 5.0

[Andy Farkas]

On Tue, 4 Mar 2003, Mike Loiterman wrote:

> [21:02:07 [EMAIL PROTECTED]: /usr/src]# swapinfo
> Device  1K-blocks UsedAvail Capacity  Type
> /dev/ad2s1b 78008515362647266%Interleaved
>
> I don't know what to make of either of these commands.
>
> Which logs should I be looking at?  Here is `tail' of /var/log/messages:
>
> Mar  3 18:14:50 little_boy kernel: swap_pager: out of swap space
> Mar  3 18:14:50 little_boy kernel: swap_pager_getswapspace: failed
> Mar  3 18:14:50 little_boy kernel: pid 40733 (genattrtab), uid 0, was killed: out of 
> swap space
> Mar  3 20:09:38 little_boy kernel: swap_pager: out of swap space
> Mar  3 20:09:55 little_boy kernel: swap_pager_getswapspace: failed
> Mar  3 20:09:55 little_boy kernel: pid 58759 (genattrtab), uid 0, was killed: out of 
> swap space
>
> Which seems that I am out of space, but I'm not sure why that is
> happening.  My machine is old and slow, but I never had problems
> compiling 4.x?

Yup, you are out of swap space.

Fix: add more RAM or swap space.

As to why this is happening.probably because you have to many
processes running. Do 'ps aux' and check the %MEM column for memory pigs..

Workaround: drop to single-user mode and try buildworld again.

--

 :{ [EMAIL PROTECTED]

Andy Farkas
System Administrator
   Speednet Communications
 http://www.speednet.com.au/




To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

RE: Make buildworld fails for 5.0

[Mike Loiterman]

 
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On Tuesday, March 04, 2003 8:47 PM Andy Farkas  wrote:

>> I'm getting the following error in my:
>> 
>> 'make buildworld'
>> 
>> ===> gnu/usr.bin/cc/cc_int
>> /usr/obj/usr/src/i386/usr/src/gnu/usr.bin/cc/cc_int/../cc_tools/genattrtab
>> /usr/src/gnu/usr.bin/cc/cc_int/../../../../contrib/gcc/config/i386/i386.md
>> > insn-attrtab.c Killed  
>> *** Error code 137
>> 
> 
> Sounds like you're out of memory/swap or hit some other process limit.
> 
> what does 'limits' and/or 'swapinfo' say?  Anything in the logfiles?


[20:42:37 [EMAIL PROTECTED]: /usr/src]# limits
Resource limits (current):
  cputime  infinity secs
  filesize infinity kb
  datasize   524288 kb
  stacksize   65536 kb
  coredumpsize infinity kb
  memoryuseinfinity kb
  memorylocked infinity kb
  maxprocesses  478
  openfiles 957
  sbsize   infinity bytes
  vmemoryuse   infinity kb

[21:02:07 [EMAIL PROTECTED]: /usr/src]# swapinfo
Device  1K-blocks UsedAvail Capacity  Type
/dev/ad2s1b 78008515362647266%Interleaved

I don't know what to make of either of these commands.

Which logs should I be looking at?  Here is `tail' of /var/log/messages:

Mar  3 18:14:50 little_boy kernel: swap_pager: out of swap space
Mar  3 18:14:50 little_boy kernel: swap_pager_getswapspace: failed
Mar  3 18:14:50 little_boy kernel: pid 40733 (genattrtab), uid 0, was killed: out of 
swap space
Mar  3 20:09:38 little_boy kernel: swap_pager: out of swap space
Mar  3 20:09:55 little_boy kernel: swap_pager_getswapspace: failed
Mar  3 20:09:55 little_boy kernel: pid 58759 (genattrtab), uid 0, was killed: out of 
swap space

Which seems that I am out of space, but I'm not sure why that is happening.  My 
machine is old and slow, but I never had problems compiling 4.x?

- ---
Randomly Generated Quote:
'There are lies, damned lies, and
statistics.' -Mark Twain

Mike Loiterman
PGP Key 0xD1B9D18E
http://www.ascendency.net



-BEGIN PGP SIGNATURE-
Version: PGP 8.0
Comment: This message has been digitally signed by Mike Loiterman

iQA/AwUBPmVpjmjZbUnRudGOEQK4IwCg04FEx8LEtBs5/Nl4VMKITgMf6dsAn0Mn
ZNwlHOrMTDGDx4D/cdRbLqFQ
=kibk
-END PGP SIGNATURE-


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: Make buildworld fails for 5.0

[Andy Farkas]

>
> I'm getting the following error in my:
>
> 'make buildworld'
>
> ===> gnu/usr.bin/cc/cc_int
> /usr/obj/usr/src/i386/usr/src/gnu/usr.bin/cc/cc_int/../cc_tools/genattrtab 
> /usr/src/gnu/usr.bin/cc/cc_int/../../../../contrib/gcc/config/i386/i386.md > 
> insn-attrtab.c
> Killed
> *** Error code 137
>

Sounds like you're out of memory/swap or hit some other process limit.

what does 'limits' and/or 'swapinfo' say?  Anything in the logfiles?

--

 :{ [EMAIL PROTECTED]

Andy Farkas
System Administrator
   Speednet Communications
 http://www.speednet.com.au/




To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails on RELENG_4_7 - solved, but now buildkernel fails...

[E.S.]

Well, It appears that my problem in running "make buildworld" was that I had 
uncommented the line:

#NO_OPENSSL= true

in my /etc/make.conf.  This makes sense (except with respect to libncurses) 
given how many libs and apps rely depend on OpenSSL...  But 4.7 has just 
successfully built world.


But now "make buildkernel KERNCONF=KERNELCONF4" (KERNELCONF4 being the name of 
my kernel config file, of course) fails immediately.  Here's all of the 
very-short output:


--
>>> Kernel build for KERNELCONF4 started on Tue Nov 12 03:50:20 CST 2002
--
===> KERNELCONF4
mkdir -p /usr/obj/usr/src/sys
cd /usr/src/sys/i386/conf;  
PATH=/usr/obj/usr/src/i386/usr/sbin:/usr/obj/usr/src/i386/usr/bin:/usr/obj/usr/src/i386/usr/games:/sbin:/bin:/usr/sbin:/usr/bin
  
config  -d /usr/obj/usr/src/sys/KERNELCONF4 KERNELCONF4
Don't forget to do a ``make depend''
Warning: device "atapicam" is unknown
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.


I know the atapicam patches were merged into the 4.7-STABLE (RELENG_4) tree as 
of 11/01, but I'm cvsup'ing RELENG_4_7...  So I tried patching the kernel 
("patch < foo.diff") using Thomas Cuivre's atapicam patches:

atapicam-20021031.diff
atapicam-STABLE-config-20021031.diff

I built 4.6.2 with the patches from 8/20 (and so now cdrecord works with my 
IDE CD-burner), and "device atapicam" didn't present any config errors on 
that build, so why is it a problem now?

If I comment out "device atapicam", the kernel appears to build OK.  But I 
don't really want to upgrade to 4.7 unless I can get cdrecord working 
again...

Here is my kernel config file (KERNELCONF4):

machine i386
#cpuI386_CPU
#cpuI486_CPU
#cpuI586_CPU
cpu I686_CPU
ident   KERNELCONF1 
maxusers64

#makeoptionsDEBUG=-g#Build kernel with gdb(1) debug 
symbols

#optionsMATH_EMULATE#Support for x87 emulation
options INET#InterNETworking
options INET6   #IPv6 communications protocols
options FFS #Berkeley Fast Filesystem
options FFS_ROOT#FFS usable as root device [keep 
this!]
options SOFTUPDATES #Enable FFS soft updates support
options UFS_DIRHASH #Improve performance on big 
directories
options MFS #Memory Filesystem
options MD_ROOT #MD is a potential root device
options NFS #Network Filesystem
options NFS_ROOT#NFS usable as root device, NFS 
required
options MSDOSFS #MSDOS Filesystem
options CD9660  #ISO 9660 Filesystem
options CD9660_ROOT #CD-ROM usable as root, CD9660 
required
options PROCFS  #Process filesystem
options COMPAT_43   #Compatible with BSD 4.3 [KEEP THIS!]
#optionsSCSI_DELAY=15000#Delay (in ms) before probing SCSI
options UCONSOLE#Allow users to grab the console
options USERCONFIG  #boot -c editor
options VISUAL_USERCONFIG   #visual boot -c editor
options KTRACE  #ktrace(1) support
options SYSVSHM #SYSV-style shared memory
options SYSVMSG #SYSV-style message queues
options SYSVSEM #SYSV-style semaphores
options P1003_1B#Posix P1003_1B real-time extensions
options _KPOSIX_PRIORITY_SCHEDULING
options ICMP_BANDLIM#Rate limit bad replies
options KBD_INSTALL_CDEV# install a CDEV entry in /dev

# To make an SMP kernel, the next two are needed
#optionsSMP # Symmetric MultiProcessor Kernel
#optionsAPIC_IO # Symmetric (APIC) I/O

device  isa
device  eisa
device  pci

# Floppy drives
device  fdc0at isa? port IO_FD1 irq 6 drq 2
device  fd0 at fdc0 drive 0
device  fd1 at fdc0 drive 1
#
# If you have a Toshiba Libretto with its Y-E Data PCMCIA floppy,
# don't use the above line for fdc0 but the following one:
#device fdc0

# ATA and ATAPI devices
device  ata0at isa? port IO_WD1 irq 14
device  ata1at isa? port IO_WD2 irq 15
device  ata
device  atadisk # ATA disk drives
device  atapicd # ATAPI CDROM drives
device  atapifd # ATAPI floppy drives
#device atapist # ATAPI tape drives
options ATA_STATIC_ID   #Static device numbering

# SCSI Controllers
#device ahb # EISA AHA1742 family
#device ahc 

Re: make buildworld fails on RELENG_4_7

[Kent Stewart]

E.S. wrote:

I'm not sure if this is the right list for this question, so I'll ask here
first...

I've been cvsup'ing /usr/src/ for the last couple of days, and every time I
try "make buildworld' the build fails when building libncurses (it failed on
Nov. 8 on libm, but that appears to have since been fixed).

I've also noticed that Kerberos5 doesn't build due to
/usr/src/kerberos5/include/crypto-headers.h having bad paths for the headers
that crypto-headers.h #include's in it (bad path for #include
, for #include , and so on...), but I commented
out MAKE_KERBEROS5 in my /etc/make.conf to avoid this problem (I don't need
it anyway, but I'd rather not have to make buildworld again to get Krb5
support if I ever need it)...

Some (hopefully-relevant) info:
Hardware:
Dell's i815-based mobo, P733, 512MB PC133 RAM, 30GB Quantum Fireball HDD,
Geforce256SDR, 3COM 3c905C-TX NIC

FreeBSD:
* presently running 4.6.2-RELEASE
* cvsup'ing /usr/src/ to RELENG_4_7 using src-all in my src-supfile
* /usr/src/ contains no source patched by me (my source tree is only what I
get from the CVS server's src tree, i.e. no atapicam patches (since they're
in the tree as of 11/01)
* I'm building /usr/obj/ in using a 445MByte MFS filesystem ("mount_mfs -s
912600 /dev/ad0s3b /usr/obj/") -- this has worked in the past when building
4.5-p10, 4.6.0, 4.6.1, and 4.6.2 (keeps /usr/obj/ from cluttering my HDD!)
and doesn't seem to be a factor, since I've tried building /usr/obj/
"normally" to my HDD and I still get errors in buildworld at the same
place...


% cc -v
Using builtin specs.
gcc version 2.95.3 20010315 (release) [FreeBSD]

% perl -v

This is perl, v5.8.0 built for i386-freebsd




Well, I haven't been having any problems for quite awhile. My 
environment is as follows:

ruby# cc -v
Using builtin specs.
gcc version 2.95.4 20020320 [FreeBSD]
ruby# perl -v

This is perl, version 5.005_03 built for i386-freebsd

ruby# cat /etc/make.conf
USA_RESIDENT=YES
HAVE_MOTIF=YES
USE_128BIT=YES
COMPAT22=yes
COMPAT3x=yes
COMPAT4x=yes
XFREE86_VERSION=  4

NOPROFILE=true

KERNCONF=RUBY
#KERNCONF=RUBY GENERIC
#KERNCONF=GENERIC

#CFLAGS= -O -pipe -I/usr/local/include
#COPTFLAGS= -O -pipe

I think you dug a deep hole and fell in.

Kent



--
Kent Stewart
Richland, WA

http://users.owt.com/kstewart/index.html


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Chris Pressey]

On Thu, 7 Nov 2002 12:09:43 -0600
Chris Pressey <[EMAIL PROTECTED]> wrote:

> On Wed, 30 Oct 2002 08:11:34 +
> Matthew Seaman <[EMAIL PROTECTED]> wrote:
> 
> > I think it's worth trying NO_OPENSSH=true in /etc/make.conf together
> > with your C_INCLUDE_PATH thing to see if you can get to the end of the
> > buildworld process.  If you can get the rest of the system up to
> > 4.7-STABLE, you should have no trouble compiling normally.  Means two
> > buildworld cycles to get everything up to date, I'm afraid.
> > 
> > Cheers,
> > 
> > Matthew
> 
> Good news - I finally got make buildworld to build.
> 
> I had to add /usr/src/include:/usr/src/sys to the beginning of my
> C_INCLUDE_PATH.  I also had to remove several libraries from /usr/lib/*,
> and replace them with copies from /usr/src/*, so that symbols could
> resolve.  Apparently it was trying to build using my existing include
> files and library files, when they are clearly insufficient, older
> versions.  I have no idea why it was trying to do this, but hopefully
> after the second build cycle it will revert to the old correct
> behaviour. I'm about to try building and installing the kernel, and I'll
> post a second followup after the dust has cleared.
> 
> Thanks once again to everyone for their help.
> -Chris
> 
> To Unsubscribe: send mail to [EMAIL PROTECTED]
> with "unsubscribe freebsd-questions" in the body of the message

Well, good news and bad news.  I updated to 4.7-RELEASE p1 (RELENG_4_7)
and can make buildworld with it no problems.  But if I try following
4.7-STABLE (RELENG_4), my make buildworld fails:

===> usr.bin/kdump
cc -O -pipe  -I/usr/src/usr.bin/kdump/../ktrace
-I/usr/src/usr.bin/kdump/../..-c /usr/src/usr.bin/kdump/kdump.c
cc -O -pipe  -I/usr/src/usr.bin/kdump/../ktrace
-I/usr/src/usr.bin/kdump/../..-c ioctl.c
In file included from ioctl.c:92:
/usr/include/sys/memrange.h:19: warning: `MDF_ACTIVE' redefined
/usr/include/pccard/cardinfo.h:115: warning: this is the location of the
previous definition
In file included from ioctl.c:104:
/usr/include/sys/wormio.h:102: warning: `CDRIOCBLANK' redefined
/usr/include/sys/cdrio.h:88: warning: this is the location of the previous
definition
ioctl.c: In function `ioctlname':
ioctl.c:1335: `USB_GET_REPORT_ID' undeclared (first use in this function)
ioctl.c:1335: (Each undeclared identifier is reported only once
ioctl.c:1335: for each function it appears in.)
*** Error code 1

Stop in /usr/src/usr.bin/kdump.
*** Error code 1

Stop in /usr/src/usr.bin.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.

Mysteriously, 'grep -r ioctlname /usr/src' finds only:

/usr/src/usr.bin/kdump/kdump.c: char *ioctlname();
/usr/src/usr.bin/kdump/kdump.c: if ((cp = ioctlname(*ip)) != NULL)
/usr/src/usr.bin/kdump/mkioctls: print "ioctlname(register_t val)"
/usr/src/usr.bin/truss/extern.h: extern const char *ioctlname(register_t
val);
/usr/src/usr.bin/truss/syscalls.c: const char *temp =
ioctlname(args[sc->offset]);

No sign of a file called ioctl.c, that I can see.

I think I'll stick with 4.7-RELEASE for a while...

-Chris

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Chris Pressey]

On Wed, 30 Oct 2002 08:11:34 +
Matthew Seaman <[EMAIL PROTECTED]> wrote:

> I think it's worth trying NO_OPENSSH=true in /etc/make.conf together
> with your C_INCLUDE_PATH thing to see if you can get to the end of the
> buildworld process.  If you can get the rest of the system up to
> 4.7-STABLE, you should have no trouble compiling normally.  Means two
> buildworld cycles to get everything up to date, I'm afraid.
> 
>   Cheers,
> 
>   Matthew

Good news - I finally got make buildworld to build.

I had to add /usr/src/include:/usr/src/sys to the beginning of my
C_INCLUDE_PATH.  I also had to remove several libraries from /usr/lib/*,
and replace them with copies from /usr/src/*, so that symbols could
resolve.  Apparently it was trying to build using my existing include
files and library files, when they are clearly insufficient, older
versions.  I have no idea why it was trying to do this, but hopefully
after the second build cycle it will revert to the old correct behaviour. 
I'm about to try building and installing the kernel, and I'll post a
second followup after the dust has cleared.

Thanks once again to everyone for their help.
-Chris

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Chris Pressey]

On Fri, 1 Nov 2002 14:11:10 +0200
Giorgos Keramidas <[EMAIL PROTECTED]> wrote:

> On 2002-10-31 20:26, Chris Pressey <[EMAIL PROTECTED]> wrote:
> > On Thu, 31 Oct 2002 14:32:40 +0200
> > Lefteris Tsintjelis <[EMAIL PROTECTED]> wrote:
> > > Compile with OPENSSH and use sshd_program="/usr/local/sbin/sshd" at
> > > ur rc.conf if you are trying to use the new sshd. U will also need
> > > to change the PATH so that /usr/local/. comes first. Its not the
> > > proper way but still, its a workaround. :-)))
> >
> > Changing the PATH so that /usr/local/bin comes before /usr/bin
> > resulted in a couple of differences in attempting to build it that
> > seemed to go away once I dealt with /usr/local/bin/make and
> > /usr/local/bin/grep (they were the GNU versions.)  Does the PATH need
> > to be this way to buildworld, or just to run ssh?
> 
> That's not a good idea.  To try building world+kernel with a PATH that
> is different from the one in /etc/login.conf:
> 
> /sbin:/bin:/usr/sbin:/usr/bin:/usr/games:/usr/local/sbin:/usr/local/bin
> :/usr/X11R6/bin:~/bin

OK.  I changed my path to this, cvsup'ed this morning's sources,
chflags'ed and destroyed /usr/obj, did a 'make buildworld' and got the
same error as reported in my last post.

> On a sidenote, why is your GNU make installed as "make" and not
> "gmake"?  If you install GNU make as "gmake", and not "make" you won't
> have to worry about /usr/local/bin/make overriding /usr/bin/make.
> That's what the devel/gmake does already:
> 
>   keramida@gray[14:06]/home/keramida$ ls -l /usr/local/bin/*make
>   -r-xr-xr-x  1 root  wheel  132860 Oct 30 02:10
>   /usr/local/bin/gmake keramida@gray[14:07]/home/keramida$ pkg_info
>   | grep gmake
>   gmake-3.80  GNU version of 'make' utility

It was probably left over from an early install of the GNU toolchain, from
before I was using the ports tree.  I do use gmake from ports these days.

This is certainly not a stock FreeBSD install - e.g. I'm running qmail -
but I can't ever recall making a change to the base system.

Thanks again to everyone for all your help,
-Chris

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Lefteris Tsintjelis]

Chris Pressey wrote:
> 
> On Thu, 31 Oct 2002 14:32:40 +0200
> Lefteris Tsintjelis <[EMAIL PROTECTED]> wrote:
> 
> > > I've done this now with NO_OPENSSH=true in make.conf, but I *still*
> > > get errors!
> > >
> > I had the exact same results when I tried NO_OPENSSH=true. Kernel
> >   ^World
> > wouldn't compile. Seems like kernel is broken with NO_OPENSSH=true.
> > ^World
> > Compile with OPENSSH and use sshd_program="/usr/local/sbin/sshd" at ur
> > rc.conf if you are trying to use the new sshd. U will also need to
> > change the PATH so that /usr/local/. comes first. Its not the proper
> > way but still, its a workaround. :-)))
> 
> Changing the PATH so that /usr/local/bin comes before /usr/bin resulted in
> a couple of differences in attempting to build it that seemed to go away
> once I dealt with /usr/local/bin/make and /usr/local/bin/grep (they were
> the GNU versions.)  Does the PATH need to be this way to buildworld, or
> just to run ssh?

Just for ssh or any other port u want to run first and coexists in your
world, and u just can't get rid of it. :) U should normally make
world/kernel with normal PATH unless u know what u r doing. I would
actually recommend to leave PATH at defaults when building world/kernel.
Otherwise it could break dependencies down. Sorry I forgot to mention
that before.

> Thanks again for everybody's help,
> -Chris


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Giorgos Keramidas]

On 2002-10-31 20:26, Chris Pressey <[EMAIL PROTECTED]> wrote:
> On Thu, 31 Oct 2002 14:32:40 +0200
> Lefteris Tsintjelis <[EMAIL PROTECTED]> wrote:
> > Compile with OPENSSH and use sshd_program="/usr/local/sbin/sshd" at ur
> > rc.conf if you are trying to use the new sshd. U will also need to
> > change the PATH so that /usr/local/. comes first. Its not the proper
> > way but still, its a workaround. :-)))
>
> Changing the PATH so that /usr/local/bin comes before /usr/bin resulted in
> a couple of differences in attempting to build it that seemed to go away
> once I dealt with /usr/local/bin/make and /usr/local/bin/grep (they were
> the GNU versions.)  Does the PATH need to be this way to buildworld, or
> just to run ssh?

That's not a good idea.  To try building world+kernel with a PATH that
is different from the one in /etc/login.conf:

/sbin:/bin:/usr/sbin:/usr/bin:/usr/games:/usr/local/sbin:/usr/local/bin:/usr/X11R6/bin:~/bin

On a sidenote, why is your GNU make installed as "make" and not
"gmake"?  If you install GNU make as "gmake", and not "make" you won't
have to worry about /usr/local/bin/make overriding /usr/bin/make.
That's what the devel/gmake does already:

keramida@gray[14:06]/home/keramida$ ls -l /usr/local/bin/*make
-r-xr-xr-x  1 root  wheel  132860 Oct 30 02:10 /usr/local/bin/gmake
keramida@gray[14:07]/home/keramida$ pkg_info | grep gmake
gmake-3.80  GNU version of 'make' utility

> Now when I run make buildworld, I get errors that seem to be related to
> glob.h once again:

I'll try a buildworld with a make.conf that differs from the
distributed make.conf in one line: NO_OPENSSH=yes.

Giorgos.

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Chris Pressey]

On Thu, 31 Oct 2002 14:32:40 +0200
Lefteris Tsintjelis <[EMAIL PROTECTED]> wrote:

> > I've done this now with NO_OPENSSH=true in make.conf, but I *still*
> > get errors!
> > 
> I had the exact same results when I tried NO_OPENSSH=true. Kernel
>   ^World
> wouldn't compile. Seems like kernel is broken with NO_OPENSSH=true.
> ^World
> Compile with OPENSSH and use sshd_program="/usr/local/sbin/sshd" at ur
> rc.conf if you are trying to use the new sshd. U will also need to
> change the PATH so that /usr/local/. comes first. Its not the proper
> way but still, its a workaround. :-)))

Changing the PATH so that /usr/local/bin comes before /usr/bin resulted in
a couple of differences in attempting to build it that seemed to go away
once I dealt with /usr/local/bin/make and /usr/local/bin/grep (they were
the GNU versions.)  Does the PATH need to be this way to buildworld, or
just to run ssh?

Now when I run make buildworld, I get errors that seem to be related to
glob.h once again:

cc -O -pipe  -DLIBC_RCS -DSYSLIBC_RCS -I/usr/src/lib/libc/include
-D__DBINTERFACE_PRIVATE -DINET6 -DPOSIX_MISTAKE
-I/usr/src/lib/libc/../libc/locale -DBROKEN_DES -DYP  -c
/usr/src/lib/libc/../libc/gen/glob.c -o glob.o
/usr/src/lib/libc/../libc/gen/glob.c: In function `glob':
/usr/src/lib/libc/../libc/gen/glob.c:169: `GLOB_APPEND' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c:169: (Each undeclared identifier is
reported only once
/usr/src/lib/libc/../libc/gen/glob.c:169: for each function it appears
in.)
/usr/src/lib/libc/../libc/gen/glob.c:172: `GLOB_DOOFFS' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c:175: `GLOB_LIMIT' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c:181: `GLOB_MAGCHAR' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c:187: `GLOB_NOESCAPE' undeclared
(first use in this function)
/usr/src/lib/libc/../libc/gen/glob.c:205: `GLOB_BRACE' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c: In function `globtilde':
/usr/src/lib/libc/../libc/gen/glob.c:361: `GLOB_TILDE' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c: In function `glob0':
/usr/src/lib/libc/../libc/gen/glob.c:463: `GLOB_MAGCHAR' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c:498: `GLOB_NOCHECK' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c:499: `GLOB_NOMAGIC' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c:503: `GLOB_NOMATCH' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c:505: `GLOB_NOSORT' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c: In function `glob2':
/usr/src/lib/libc/../libc/gen/glob.c:558: `GLOB_MARK' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c:564: `GLOB_ABORTED' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c: In function `glob3':
/usr/src/lib/libc/../libc/gen/glob.c:618: `GLOB_ABORTED' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c:628: `GLOB_ERR' undeclared (first use
in this function)
/usr/src/lib/libc/../libc/gen/glob.c:637: `GLOB_ALTDIRFUNC' undeclared
(first use in this function)
/usr/src/lib/libc/../libc/gen/glob.c: In function `globextend':
/usr/src/lib/libc/../libc/gen/glob.c:698: `GLOB_NOSPACE' undeclared (first
use in this function)
/usr/src/lib/libc/../libc/gen/glob.c: In function `g_opendir':
/usr/src/lib/libc/../libc/gen/glob.c:822: `GLOB_ALTDIRFUNC' undeclared
(first use in this function)
/usr/src/lib/libc/../libc/gen/glob.c: In function `g_lstat':
/usr/src/lib/libc/../libc/gen/glob.c:840: `GLOB_ALTDIRFUNC' undeclared
(first use in this function)
/usr/src/lib/libc/../libc/gen/glob.c: In function `g_stat':
/usr/src/lib/libc/../libc/gen/glob.c:857: `GLOB_ALTDIRFUNC' undeclared
(first use in this function)
*** Error code 1

Stop in /usr/src/lib/libc.
*** Error code 1

Stop in /usr/src/lib.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.


This is with /usr/src/include at the start of my C_INCLUDE_PATH.  (With my
usual C_INCLUDE_PATH this gives me the first error message I posted, which
is slightly different.  Both versions of glob.h seem to define at least
GLOB_APPEND.)

Thanks again for everybody's help,
-Chris

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Lefteris Tsintjelis]

Giorgos Keramidas wrote:
> 
> On 2002-10-30 19:12, Chris Pressey <[EMAIL PROTECTED]> wrote:
> > On Wed, 30 Oct 2002 23:41:49 +0200
> > Giorgos Keramidas <[EMAIL PROTECTED]> wrote:
> > > + Remove everything under /usr/obj.
> > >
> > > # chflags -R noschg /usr/obj
> > > # rm -fr /usr/obj/*
> > > + CVSup your sources.
> > > + Start a new buildworld.  Avoid using -DNOCLEAN or similar flags.
> >
> > I've done this now with NO_OPENSSH=true in make.conf, but I *still* get
> > errors!
> >
> > cc -O -pipe  -o ppp acf.o arp.o async.o auth.o bundle.o cbcp.o ccp.o
> > chap.o chat.o command.o datalink.o deflate.o defs.o exec.o filter.o fsm.o
> > hdlc.o iface.o ip.o ipcp.o ipv6cp.o iplist.o lcp.o link.o log.o lqr.o
> > main.o mbuf.o mp.o ncp.o ncpaddr.o pap.o physical.o pred.o probe.o
> > prompt.o proto.o route.o server.o sig.o slcompress.o sync.o systems.o
> > tcp.o tcpmss.o throughput.o timer.o tty.o tun.o udp.o vjcomp.o nat_cmd.o
> > atm.o id.o chap_ms.o mppe.o radius.o i4b.o ether.o -lcrypt -lmd -lutil -lz
> > -lalias -lcrypto -lradius -lnetgraph
> > radius.o: In function `demangle':
> > radius.o(.text+0xd1): undefined reference to `rad_request_authenticator'
> > radius.o(.text+0x102): undefined reference to `rad_server_secret'
> > radius.o: In function `radius_Process':
> > radius.o(.text+0x84d): undefined reference to `rad_get_vendor_attr'
> > radius.o: In function `radius_Authenticate':
> > radius.o(.text+0x1217): undefined reference to `rad_put_vendor_attr'
> > radius.o(.text+0x12a3): undefined reference to `rad_put_vendor_attr'
> > radius.o(.text+0x1323): undefined reference to `rad_put_vendor_attr'
> > *** Error code 1
> 
> Odd.  This should work fine.  Try backing up your data and installing
> a -RELEASE version (preferrably 4.6.2-RELEASE).  It's obvious that
> your build system is at least, uhm, non-standard.  If that doesn't
> work either, you can ask around in freebsd-stable to see if other
> people have similar problems.  But this is very unlikely with -STABLE.

Giorgo, I am using 4.7-STABLE FreeBSD 4.7-STABLE #0: Tue Oct 29 02:30:34
EET 2002 and I still get the same exact error if I set NO_OPENSSH=true
in make.conf. I am also forwarding this to STABLE discusion.


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Giorgos Keramidas]

On 2002-10-30 19:12, Chris Pressey <[EMAIL PROTECTED]> wrote:
> On Wed, 30 Oct 2002 23:41:49 +0200
> Giorgos Keramidas <[EMAIL PROTECTED]> wrote:
> > + Remove everything under /usr/obj.
> > 
> > # chflags -R noschg /usr/obj
> > # rm -fr /usr/obj/*
> > + CVSup your sources.
> > + Start a new buildworld.  Avoid using -DNOCLEAN or similar flags.
> 
> I've done this now with NO_OPENSSH=true in make.conf, but I *still* get
> errors!
> 
> cc -O -pipe  -o ppp acf.o arp.o async.o auth.o bundle.o cbcp.o ccp.o
> chap.o chat.o command.o datalink.o deflate.o defs.o exec.o filter.o fsm.o
> hdlc.o iface.o ip.o ipcp.o ipv6cp.o iplist.o lcp.o link.o log.o lqr.o
> main.o mbuf.o mp.o ncp.o ncpaddr.o pap.o physical.o pred.o probe.o
> prompt.o proto.o route.o server.o sig.o slcompress.o sync.o systems.o
> tcp.o tcpmss.o throughput.o timer.o tty.o tun.o udp.o vjcomp.o nat_cmd.o
> atm.o id.o chap_ms.o mppe.o radius.o i4b.o ether.o -lcrypt -lmd -lutil -lz
> -lalias -lcrypto -lradius -lnetgraph
> radius.o: In function `demangle':
> radius.o(.text+0xd1): undefined reference to `rad_request_authenticator'
> radius.o(.text+0x102): undefined reference to `rad_server_secret'
> radius.o: In function `radius_Process':
> radius.o(.text+0x84d): undefined reference to `rad_get_vendor_attr'
> radius.o: In function `radius_Authenticate':
> radius.o(.text+0x1217): undefined reference to `rad_put_vendor_attr'
> radius.o(.text+0x12a3): undefined reference to `rad_put_vendor_attr'
> radius.o(.text+0x1323): undefined reference to `rad_put_vendor_attr'
> *** Error code 1

Odd.  This should work fine.  Try backing up your data and installing
a -RELEASE version (preferrably 4.6.2-RELEASE).  It's obvious that
your build system is at least, uhm, non-standard.  If that doesn't
work either, you can ask around in freebsd-stable to see if other
people have similar problems.  But this is very unlikely with -STABLE.

Giorgos.

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Lefteris Tsintjelis]

> I've done this now with NO_OPENSSH=true in make.conf, but I *still* get
> errors!
> 
I had the exact same results when I tried NO_OPENSSH=true. Kernel
  ^World
wouldn't compile. Seems like kernel is broken with NO_OPENSSH=true.
^World
Compile with OPENSSH and use sshd_program="/usr/local/sbin/sshd" at ur
rc.conf if you are trying to use the new sshd. U will also need to
change the PATH so that /usr/local/. comes first. Its not the proper
way but still, its a workaround. :-)))


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Lefteris Tsintjelis]

Chris Pressey wrote:
> 
> On Wed, 30 Oct 2002 23:41:49 +0200
> Giorgos Keramidas <[EMAIL PROTECTED]> wrote:
> 
> > On 2002-10-29 19:26, Chris Pressey <[EMAIL PROTECTED]> wrote:
> > > On Tue, 29 Oct 2002 14:28:39 +
> > > Matthew Seaman <[EMAIL PROTECTED]> wrote:
> > > > I think what has happened is that one or more of the Makefiles in
> > > > the src tree hasn't been updated properly, possibly because cvsup(1)
> > > > doesn't think it owns the file.  Try doing what the cvsup FAQ says,
> > > > and running an update with the 'list=cvs:RELENG_4' added to the
> > > > supfile.
> > > >
> > > > With any luck you should see updates to files matching '*.mk' and
> > > > 'Makefile*', and then the build will succeed.
> > >
> > > Unfortunately, no luck there.  No Makefile-type files were updated.
> > > So, I toasted my *entire* source tree (including the 'checkouts'
> > > files) and grabbed it all anew, tried rebuilding the world... and
> > > got the same error.
> >
> > To start with as clean an environment as possible for a buildworld you
> > really have to do more than that:
> >
> > + Remove everything under /usr/obj.
> >
> >   # chflags -R noschg /usr/obj
> >   # rm -fr /usr/obj/*
> 
> I did this step a couple of days ago, just before my first post to this
> list.  To be on the safe side, I did it again before the most recent
> attempt.
> 
> > + CVSup your sources.
> 
> I did this step yesterday.
> 
> > + Start a new buildworld.  Avoid using -DNOCLEAN or similar flags.
> 
> I've done this now with NO_OPENSSH=true in make.conf, but I *still* get
> errors!

I had the exact same results when I tried NO_OPENSSH=true. Kernel
wouldn't compile. Seems like kernel is broken with NO_OPENSSH=true.
Compile with OPENSSH and use sshd_program="/usr/local/sbin/sshd" at ur
rc.conf if you are trying to use the new sshd. U will also need to
change the PATH so that /usr/local/. comes first. Its not the proper
way but still, its a workaround.


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Chris Pressey]

On Wed, 30 Oct 2002 23:41:49 +0200
Giorgos Keramidas <[EMAIL PROTECTED]> wrote:

> On 2002-10-29 19:26, Chris Pressey <[EMAIL PROTECTED]> wrote:
> > On Tue, 29 Oct 2002 14:28:39 +
> > Matthew Seaman <[EMAIL PROTECTED]> wrote:
> > > I think what has happened is that one or more of the Makefiles in
> > > the src tree hasn't been updated properly, possibly because cvsup(1)
> > > doesn't think it owns the file.  Try doing what the cvsup FAQ says,
> > > and running an update with the 'list=cvs:RELENG_4' added to the
> > > supfile.
> > >
> > > With any luck you should see updates to files matching '*.mk' and
> > > 'Makefile*', and then the build will succeed.
> >
> > Unfortunately, no luck there.  No Makefile-type files were updated.
> > So, I toasted my *entire* source tree (including the 'checkouts'
> > files) and grabbed it all anew, tried rebuilding the world... and
> > got the same error.
> 
> To start with as clean an environment as possible for a buildworld you
> really have to do more than that:
> 
> + Remove everything under /usr/obj.
> 
>   # chflags -R noschg /usr/obj
>   # rm -fr /usr/obj/*

I did this step a couple of days ago, just before my first post to this
list.  To be on the safe side, I did it again before the most recent
attempt.

> + CVSup your sources.

I did this step yesterday.

> + Start a new buildworld.  Avoid using -DNOCLEAN or similar flags.

I've done this now with NO_OPENSSH=true in make.conf, but I *still* get
errors!

cc -O -pipe  -o ppp acf.o arp.o async.o auth.o bundle.o cbcp.o ccp.o
chap.o chat.o command.o datalink.o deflate.o defs.o exec.o filter.o fsm.o
hdlc.o iface.o ip.o ipcp.o ipv6cp.o iplist.o lcp.o link.o log.o lqr.o
main.o mbuf.o mp.o ncp.o ncpaddr.o pap.o physical.o pred.o probe.o
prompt.o proto.o route.o server.o sig.o slcompress.o sync.o systems.o
tcp.o tcpmss.o throughput.o timer.o tty.o tun.o udp.o vjcomp.o nat_cmd.o
atm.o id.o chap_ms.o mppe.o radius.o i4b.o ether.o -lcrypt -lmd -lutil -lz
-lalias -lcrypto -lradius -lnetgraph
radius.o: In function `demangle':
radius.o(.text+0xd1): undefined reference to `rad_request_authenticator'
radius.o(.text+0x102): undefined reference to `rad_server_secret'
radius.o: In function `radius_Process':
radius.o(.text+0x84d): undefined reference to `rad_get_vendor_attr'
radius.o: In function `radius_Authenticate':
radius.o(.text+0x1217): undefined reference to `rad_put_vendor_attr'
radius.o(.text+0x12a3): undefined reference to `rad_put_vendor_attr'
radius.o(.text+0x1323): undefined reference to `rad_put_vendor_attr'
*** Error code 1

Stop in /usr/src/usr.sbin/ppp.
*** Error code 1

Stop in /usr/src/usr.sbin.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.


Is there any way to tell make to not compile PPP?  I don't use it.  But I
don't see anything like NO_PPP in make.conf.

-Chris

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Giorgos Keramidas]

On 2002-10-29 19:26, Chris Pressey <[EMAIL PROTECTED]> wrote:
> On Tue, 29 Oct 2002 14:28:39 +
> Matthew Seaman <[EMAIL PROTECTED]> wrote:
> > I think what has happened is that one or more of the Makefiles in the
> > src tree hasn't been updated properly, possibly because cvsup(1)
> > doesn't think it owns the file.  Try doing what the cvsup FAQ says,
> > and running an update with the 'list=cvs:RELENG_4' added to the supfile.
> >
> > With any luck you should see updates to files matching '*.mk' and
> > 'Makefile*', and then the build will succeed.
>
> Unfortunately, no luck there.  No Makefile-type files were updated.
> So, I toasted my *entire* source tree (including the 'checkouts'
> files) and grabbed it all anew, tried rebuilding the world... and
> got the same error.

To start with as clean an environment as possible for a buildworld you
really have to do more than that:

+ Remove everything under /usr/obj.

# chflags -R noschg /usr/obj
# rm -fr /usr/obj/*

+ CVSup your sources.

+ Start a new buildworld.  Avoid using -DNOCLEAN or similar flags.

Giorgos.

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Matthew Seaman]

On Tue, Oct 29, 2002 at 07:26:15PM -0600, Chris Pressey wrote:
> Unfortunately, no luck there.  No Makefile-type files were updated.  So, I
> toasted my *entire* source tree (including the 'checkouts' files) and
> grabbed it all anew, tried rebuilding the world... and got the same error.

Well, I'm fresh out of ideas about what's wrong.

> So I added /usr/src/include to the start of my C_INCLUDE_PATH and tried it
> again.  It seemed to work better - it got over the place where the glob.h
> error occurred - but now it stops with:

The C_INCLUDE_PATH environment variable is not something that I've
ever seen mentioned in the context of 'make buildworld'.  It should
all work perfectly well with that variable unset.  In fact, looking at
the 'Environment Variables' node in the gcc.info pages, other than the
locale settings LC_* and TMPDIR I think that all of those
env. variables should be left unset for best results.

However, that's just the theory.  It may well work in practice.

> cc -O -pipe  -I/usr/src/secure/libexec/sftp-server/../../../crypto/openssh
> -DNO_IDEA -o sftp-server sftp-server.o sftp-common.o -lssh -lcrypto
> /usr/obj/usr/src/i386/usr/lib/libssh.so: undefined reference to
> `readpassphrase'
> *** Error code 1

I think it's worth trying NO_OPENSSH=true in /etc/make.conf together
with your C_INCLUDE_PATH thing to see if you can get to the end of the
buildworld process.  If you can get the rest of the system up to
4.7-STABLE, you should have no trouble compiling normally.  Means two
buildworld cycles to get everything up to date, I'm afraid.

Cheers,

Matthew

-- 
Dr Matthew J Seaman MA, D.Phil.   26 The Paddocks
  Savill Way
  Marlow
Tel: +44 1628 476614  Bucks., SL7 1TH UK

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Chris Pressey]

On Tue, 29 Oct 2002 14:28:39 +
Matthew Seaman <[EMAIL PROTECTED]> wrote:

> On Tue, Oct 29, 2002 at 06:47:30AM -0600, Chris Pressey wrote:
> 
> > I do indeed have /usr/src/include/glob.h, and it is 4480 bytes and it
> > defines the macros in question.  However, the copy in /usr/include is
> > only 4055 bytes and is older (Aug 21 2001) and it does NOT define the
> > macros in question.  This is likely the problem - if
> > /usr/include/glob.h is being used in the build instead of
> > /usr/src/include/glob.h - but also, that sounds wrong - to build the
> > world, shouldn't it be using the latest and greatest header files,
> > i.e., /usr/src/include ?  If not, would it be safe to just copy the
> > newer glob.h into /usr/include (or explicitly add/usr/src/include to
> > the start of my C_INCLUDE_PATH, at least for building the world?)
> 
> Yes.  I would expect that building in /usr/src would use the headers
> from /usr/src/include.  You could try installing the latest headers
> into /usr/include before starting your build, but that's a hack and
> not guarranteed to work.
> 
> I think what has happened is that one or more of the Makefiles in the
> src tree hasn't been updated properly, possibly because cvsup(1)
> doesn't think it owns the file.  Try doing what the cvsup FAQ says,
> and running an update with the 'list=cvs:RELENG_4' added to the supfile.
> 
> With any luck you should see updates to files matching '*.mk' and
> 'Makefile*', and then the build will succeed.

Unfortunately, no luck there.  No Makefile-type files were updated.  So, I
toasted my *entire* source tree (including the 'checkouts' files) and
grabbed it all anew, tried rebuilding the world... and got the same error.

So I added /usr/src/include to the start of my C_INCLUDE_PATH and tried it
again.  It seemed to work better - it got over the place where the glob.h
error occurred - but now it stops with:

cc -O -pipe  -I/usr/src/secure/libexec/sftp-server/../../../crypto/openssh
-DNO_IDEA -o sftp-server sftp-server.o sftp-common.o -lssh -lcrypto
/usr/obj/usr/src/i386/usr/lib/libssh.so: undefined reference to
`readpassphrase'
*** Error code 1

Stop in /usr/src/secure/libexec/sftp-server.
*** Error code 1

Stop in /usr/src/secure/libexec.
*** Error code 1

Stop in /usr/src/secure.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.


The libshh.so being referred to is about an hour old, so I assume it was
freshly created during the build.  But still - maybe I have to adjust the
LIBRARY_PATH ?  I don't know.

Thanks again,
-Chris

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Matthew Seaman]

On Tue, Oct 29, 2002 at 06:47:30AM -0600, Chris Pressey wrote:

> I do indeed have /usr/src/include/glob.h, and it is 4480 bytes and it
> defines the macros in question.  However, the copy in /usr/include is only
> 4055 bytes and is older (Aug 21 2001) and it does NOT define the macros in
> question.  This is likely the problem - if /usr/include/glob.h is being
> used in the build instead of /usr/src/include/glob.h - but also, that
> sounds wrong - to build the world, shouldn't it be using the latest and
> greatest header files, i.e., /usr/src/include ?  If not, would it be safe
> to just copy the newer glob.h into /usr/include (or explicitly add
> /usr/src/include to the start of my C_INCLUDE_PATH, at least for building
> the world?)

Yes.  I would expect that building in /usr/src would use the headers
from /usr/src/include.  You could try installing the latest headers
into /usr/include before starting your build, but that's a hack and
not guarranteed to work.

I think what has happened is that one or more of the Makefiles in the
src tree hasn't been updated properly, possibly because cvsup(1)
doesn't think it owns the file.  Try doing what the cvsup FAQ says,
and running an update with the 'list=cvs:RELENG_4' added to the supfile.

With any luck you should see updates to files matching '*.mk' and
'Makefile*', and then the build will succeed.

Cheers,

Matthew

-- 
Dr Matthew J Seaman MA, D.Phil.   26 The Paddocks
  Savill Way
  Marlow
Tel: +44 1628 476614  Bucks., SL7 1TH UK

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Re: make buildworld fails

[Chris Pressey]

On Tue, 29 Oct 2002 11:29:07 +
Matthew Seaman <[EMAIL PROTECTED]> wrote:

> On Tue, Oct 29, 2002 at 03:31:43AM -0600, Chris Pressey wrote:
> 
> > I've been following the -STABLE development tree for a while now
> > without any problems, using cvsup from 'make update'.  But sometime
> > over the summer, a problem crept in somewhere and 'make buildworld'
> > now fails.
> 
> Curious.  It's not unknown for there to be problems compiling when
> trying to upgrade over a big version jump.  Not that 4.5 -- 4.7 is
> that big of a jump.  However, such problems are almost never
> experienced by one person in isolation and will usually generate
> substantial amounts of traffic on the [EMAIL PROTECTED] list.
> 
> My guess is that the problem is specific to your system: something has
> become corrupted or you've made a modification that didn't turn out to
> be quite so cunning after all.
> 
> Are you cvsup'ing the whole src-all collection, or have you tried to
> pick and choose amongst the the individual collections?  Make sure you
> grab the whole source tree.

I'm grabbing the whole thing, yes.

> When you first started using cvsup on your src tree, did you either
> start with a completely empty /usr/src or did you follow the
> instructions at http://www.cvsup.org/faq.html#caniadopt (that and the
> following Questions 12 and 13)?  Of course you probably want to use
> the RELENG_4 tag nowadays.  Not doing this isn't a disaster: all it
> means is that there are odd files scattered through your source tree
> that cvsup doesn't think it owns.  That's usually fairly harmless, but
> can lead to occasional compilation problems.

I installed with sources, and I simply started using cvsup a couple of
months later and it worked, so I didn't look into cleaning out the source
tree.  I will do that though, if the problem persists.

> > uname -a says my system is 4.5-STABLE #18: Tue Mar 19 14:22:13 CST
> > 2002(i386).
> > 
> > make buildworld says:
> > 
> > cc -O -pipe  -DLIBC_RCS -DSYSLIBC_RCS -I/usr/src/lib/libc/include
> > -D__DBINTERFACE_PRIVATE -DINET6 -DPOSIX_MISTAKE
> > -I/usr/src/lib/libc/../libc/locale -DBROKEN_DES -DYP  -c
> > /usr/src/lib/libc/../libc/gen/glob.c -o glob.o
> > /usr/src/lib/libc/../libc/gen/glob.c: In function `glob':
> > /usr/src/lib/libc/../libc/gen/glob.c:187: `GLOB_NOESCAPE' undeclared
> > (first use in this function)
> > /usr/src/lib/libc/../libc/gen/glob.c:187: (Each undeclared identifier
> > is reported only once
> > /usr/src/lib/libc/../libc/gen/glob.c:187: for each function it appears
> > in.)
> > /usr/src/lib/libc/../libc/gen/glob.c: In function `glob0':
> > /usr/src/lib/libc/../libc/gen/glob.c:503: `GLOB_NOMATCH' undeclared
> > (first use in this function)
> > /usr/src/lib/libc/../libc/gen/glob.c: In function `glob2':
> > /usr/src/lib/libc/../libc/gen/glob.c:564: `GLOB_ABORTED' undeclared
> > (first use in this function)
> > /usr/src/lib/libc/../libc/gen/glob.c: In function `glob3':
> > /usr/src/lib/libc/../libc/gen/glob.c:618: `GLOB_ABORTED' undeclared
> > (first use in this function)
> > *** Error code 1
> 
> You should have a file /usr/src/include/glob.h that contains the
> definitions for those macros.  On a recent 4.7-STABLE system it's 4480
> bytes.  There should be a similar, if not identical, copy in
> /usr/include/glob.h:
> 
> happy-idiot-talk:/usr/src:% fgrep '$FreeBSD' /usr/src/include/glob.h
>  * $FreeBSD: src/include/glob.h,v 1.3.6.3 2002/09/18 14:13:30 mikeh
>  Exp $
> happy-idiot-talk:/usr/src:% md5 /usr/src/include/glob.h
> MD5 (/usr/src/include/glob.h) = ac1c36ec9e7ae6e713ae0d04416bc618

I do indeed have /usr/src/include/glob.h, and it is 4480 bytes and it
defines the macros in question.  However, the copy in /usr/include is only
4055 bytes and is older (Aug 21 2001) and it does NOT define the macros in
question.  This is likely the problem - if /usr/include/glob.h is being
used in the build instead of /usr/src/include/glob.h - but also, that
sounds wrong - to build the world, shouldn't it be using the latest and
greatest header files, i.e., /usr/src/include ?  If not, would it be safe
to just copy the newer glob.h into /usr/include (or explicitly add
/usr/src/include to the start of my C_INCLUDE_PATH, at least for building
the world?)

> If that file is corrupt, you might try deleting the whole of
> /usr/src/include, and re-cvsup'ing:
> 
> cd /usr/src/include
> rm -rf *
> cd ..
> make update
> 
> Then see if the build goes any better.

I've now done this to make absolutely sure I have the latest headers.

> > FWIW, this does not seem to be the same error that I first encountered
> > a few months ago (if memory serves that error was somewhere in ppp,
> > not the C library.  But it was of similar nature, an undeclared
> > identifier.)
> 
> Seems that either you've got some scrambled C header files or your C
> compiler tool chain is completely befuddled.  Have you made any
> modifications to the stock system compiler? Have you installed, say

Re: make buildworld fails

[Matthew Seaman]

On Tue, Oct 29, 2002 at 03:31:43AM -0600, Chris Pressey wrote:

> I've been following the -STABLE development tree for a while now without
> any problems, using cvsup from 'make update'.  But sometime over the
> summer, a problem crept in somewhere and 'make buildworld' now fails.

Curious.  It's not unknown for there to be problems compiling when
trying to upgrade over a big version jump.  Not that 4.5 -- 4.7 is
that big of a jump.  However, such problems are almost never
experienced by one person in isolation and will usually generate
substantial amounts of traffic on the [EMAIL PROTECTED] list.

My guess is that the problem is specific to your system: something has
become corrupted or you've made a modification that didn't turn out to
be quite so cunning after all.

Are you cvsup'ing the whole src-all collection, or have you tried to
pick and choose amongst the the individual collections?  Make sure you
grab the whole source tree.

When you first started using cvsup on your src tree, did you either
start with a completely empty /usr/src or did you follow the
instructions at http://www.cvsup.org/faq.html#caniadopt (that and the
following Questions 12 and 13)?  Of course you probably want to use
the RELENG_4 tag nowadays.  Not doing this isn't a disaster: all it
means is that there are odd files scattered through your source tree
that cvsup doesn't think it owns.  That's usually fairly harmless, but
can lead to occasional compilation problems.

> uname -a says my system is 4.5-STABLE #18: Tue Mar 19 14:22:13 CST 2002
> (i386).
> 
> make buildworld says:
> 
> cc -O -pipe  -DLIBC_RCS -DSYSLIBC_RCS -I/usr/src/lib/libc/include
> -D__DBINTERFACE_PRIVATE -DINET6 -DPOSIX_MISTAKE
> -I/usr/src/lib/libc/../libc/locale -DBROKEN_DES -DYP  -c
> /usr/src/lib/libc/../libc/gen/glob.c -o glob.o
> /usr/src/lib/libc/../libc/gen/glob.c: In function `glob':
> /usr/src/lib/libc/../libc/gen/glob.c:187: `GLOB_NOESCAPE' undeclared
> (first use in this function)
> /usr/src/lib/libc/../libc/gen/glob.c:187: (Each undeclared identifier is
> reported only once
> /usr/src/lib/libc/../libc/gen/glob.c:187: for each function it appears
> in.)
> /usr/src/lib/libc/../libc/gen/glob.c: In function `glob0':
> /usr/src/lib/libc/../libc/gen/glob.c:503: `GLOB_NOMATCH' undeclared (first
> use in this function)
> /usr/src/lib/libc/../libc/gen/glob.c: In function `glob2':
> /usr/src/lib/libc/../libc/gen/glob.c:564: `GLOB_ABORTED' undeclared (first
> use in this function)
> /usr/src/lib/libc/../libc/gen/glob.c: In function `glob3':
> /usr/src/lib/libc/../libc/gen/glob.c:618: `GLOB_ABORTED' undeclared (first
> use in this function)
> *** Error code 1

You should have a file /usr/src/include/glob.h that contains the
definitions for those macros.  On a recent 4.7-STABLE system it's 4480
bytes.  There should be a similar, if not identical, copy in
/usr/include/glob.h:

happy-idiot-talk:/usr/src:% fgrep '$FreeBSD' /usr/src/include/glob.h
 * $FreeBSD: src/include/glob.h,v 1.3.6.3 2002/09/18 14:13:30 mikeh Exp $
happy-idiot-talk:/usr/src:% md5 /usr/src/include/glob.h
MD5 (/usr/src/include/glob.h) = ac1c36ec9e7ae6e713ae0d04416bc618

If that file is corrupt, you might try deleting the whole of
/usr/src/include, and re-cvsup'ing:

cd /usr/src/include
rm -rf *
cd ..
make update

Then see if the build goes any better.

> FWIW, this does not seem to be the same error that I first encountered a
> few months ago (if memory serves that error was somewhere in ppp, not the
> C library.  But it was of similar nature, an undeclared identifier.)

Seems that either you've got some scrambled C header files or your C
compiler tool chain is completely befuddled.  Have you made any
modifications to the stock system compiler? Have you installed, say, a
more recent version of gcc into somewhere that puts an alternate 'cc'
command on root's path and which the /usr/src Makefiles might be
picking up? Hmmm... actually if that were the case I think the
compilation would have blown up long before it got to
/usr/src/lib/libc/gen/glob.c

Cheers,

Matthew

-- 
Dr Matthew J Seaman MA, D.Phil.   26 The Paddocks
  Savill Way
  Marlow
Tel: +44 1628 476614  Bucks., SL7 1TH UK

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message