David Wolfskill | 19 Mar 2002 18:34

nl_langinfo is MFCed, but what about compat/libc.so.4?

About a week ago, src/lib/libc/locale/nl_langinfo.c in RELENG_4 received
an update from 1.13.2.2 to 1.13.2.3, which added nl_langinfo to
RELENG_4's libc.

I have no problem with that at all.  :-)

However, the symbol in question does not appear to exist in the version
of libc.so.4 in -CURRENT's /usr/lib/compat -- and that can cause
programs which had been compiled under -STABLE and which might well
otherwise work Just Fine under -CURRENT to fail, thus:

freebeast(5.0-C)[8] screen
/usr/libexec/ld-elf.so.1: screen: Undefined symbol "nl_langinfo"
freebeast(5.0-C)[9] 

In this particular case, I've been working around the nuisance by merely
recompiling a "-CURRENT-private" version of screen (on each of my build
machine & laptop).  I'm rather less eager to do this for, say, mozilla
(for example)....

What is the procedure for getting /usr/lib/compat/libc.so.4 updated to
include nl_langinfo (and, perhaps, other symbols that might be in
-STABLE's /usr/lib/libc.so.4)?

[Were this strictly a -STABLE issue, I'd file a PR.  But for -CURRENT,
things seem to be sufficiently fluid that the additional overhead of
dealing with a PR appears (to me) to be just that -- additional
overhead.  And it's not a problem that manifests itself as a "problem
with/running -STABLE" -- it bites someone running -CURRENT, who is
supposed to have some minimal level of clue....]
(Continue reading)

David O'Brien | 19 Mar 2002 18:42
Picon
Favicon

Re: nl_langinfo is MFCed, but what about compat/libc.so.4?

On Tue, Mar 19, 2002 at 09:34:19AM -0800, David Wolfskill wrote:
> However, the symbol in question does not appear to exist in the version
> of libc.so.4 in -CURRENT's /usr/lib/compat -- and that can cause
> programs which had been compiled under -STABLE and which might well
> otherwise work Just Fine under -CURRENT to fail, thus:

This just means we need to update the libc.so.4 that is in -CURRENT's
/usr/lib/compat

    *****  H O W E V E R  *****

before someone does this, I want the very last libc.so.4 from a RELENG_4
snapshot.  Can someone help pin-point what snapshot that would be?

> Sorry about the lack of attached patch; I'm not (yet) clear on how
> compat/libc.so.4 get built....  :-(

poke me. :-)

To Unsubscribe: send mail to majordomo <at> FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message

Christoph Kukulies | 19 Mar 2002 21:11
Picon
Picon
Picon

cc1 warning as errors - can one switch this off?

I cvsuped again. Got the vm_map.c now compiled but now
kernel compilation stops somehwere in ipv6 files.

Can one switch off compiler pickyness?

--

-- 
Chris Christoph P. U. Kukulies kuku <at> gil.physik.rwth-aachen.de

To Unsubscribe: send mail to majordomo <at> FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message

Steve Kargl | 19 Mar 2002 21:11
Picon

ipfilter breaks world, again.

cc -O -pipe  -DUSE_INET6 -DIPL_NAME=\"/dev/ipl\" -DSTATETOP
-I/usr/src/sbin/ipfstat/../../sys/contrib/ipfilter/netinet
-I/usr/src/sbin/ipfstat/../../sys/contrib/ipfilter
-I/usr/src/sbin/ipfstat/../../contrib/ipfilter     -static -o ipfstat fils.o parse.o opt.o kmem.o
facpri.o common.o  -lncurses -lkvm
fils.o: In function `showipstates':
fils.o(.text+0x1071): undefined reference to `printstate'
*** Error code 1

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

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

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

--

-- 
Steve

To Unsubscribe: send mail to majordomo <at> FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message

Brooks Davis | 19 Mar 2002 21:13

Re: cc1 warning as errors - can one switch this off?

On Tue, Mar 19, 2002 at 09:11:05PM +0100, Christoph Kukulies wrote:
> I cvsuped again. Got the vm_map.c now compiled but now
> kernel compilation stops somehwere in ipv6 files.
> 
> Can one switch off compiler pickyness?

From /usr/src/UPDATING:

20020225:
        Warnings are now errors in the kernel.  Unless you are a developer,
	you should add -DNO_WERROR to your make line.

-- Brooks

--

-- 
Any statement of the form "X is the one, true Y" is FALSE.
PGP fingerprint 655D 519C 26A7 82E7 2529  9BF0 5D8E 8BE9 F238 1AD4
Kris Kennaway | 19 Mar 2002 21:16

Re: cc1 warning as errors - can one switch this off?

On Tue, Mar 19, 2002 at 09:11:05PM +0100, Christoph Kukulies wrote:
> I cvsuped again. Got the vm_map.c now compiled but now
> kernel compilation stops somehwere in ipv6 files.
> 
> Can one switch off compiler pickyness?

Go back and read what config(8) is telling you, it's right there in
front of you!

Kris
Alfred Perlstein | 19 Mar 2002 21:19

fla LINT breakage.

Fixes format warnings.  Since there was so much... bitching about my
last commit to something contrib/* I'm posting the fix here.

Index: fla.c
===================================================================
RCS file: /home/ncvs/src/sys/contrib/dev/fla/fla.c,v
retrieving revision 1.29
diff -u -r1.29 fla.c
--- fla.c	12 Sep 2001 08:36:59 -0000	1.29
+++ fla.c	19 Mar 2002 20:22:51 -0000
 <at>  <at>  -181,8 +181,9  <at>  <at> 
 	enum doc2k_work what;

 	if (fla_debug > 1)
-		printf("flastrategy(%p) %s %x, %d, %ld, %p)\n",
-		    bp, devtoname(bp->bio_dev), bp->bio_flags, bp->bio_blkno, 
+		printf("flastrategy(%p) %s %x, %lld, %ld, %p)\n",
+		    bp, devtoname(bp->bio_dev), bp->bio_flags,
+		    (long long)bp->bio_blkno, 
 		    bp->bio_bcount / DEV_BSIZE, bp->bio_data);

 	sc = bp->bio_dev->si_drv1;
 <at>  <at>  -225,8 +226,9  <at>  <at> 
 		ENTER();

 		if (fla_debug > 1 || error) {
-			printf("fla%d: %d = rwe(%p, %d, %d, %d, %ld, %p)\n",
-			    unit, error, bp, unit, what, bp->bio_pblkno, 
+			printf("fla%d: %d = rwe(%p, %d, %d, %lld, %ld, %p)\n",
+			    unit, error, bp, unit, what,
(Continue reading)

Sheldon Hearn | 19 Mar 2002 21:36
Gravatar

Re: cc1 warning as errors - can one switch this off?


On Tue, 19 Mar 2002 12:16:05 PST, Kris Kennaway wrote:

> Go back and read what config(8) is telling you, it's right there in
> front of you!

You shouldn't be too hard on people about missing config(8)'s output,
given that it's buried if you use the buildkernel target, as
recommended.

Ciao,
Sheldon.

To Unsubscribe: send mail to majordomo <at> FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message

Poul-Henning Kamp | 19 Mar 2002 21:39
Picon

Re: fla LINT breakage.


You're welcome to commit it :-)

Poul-Henning

In message <20020319201901.GJ455 <at> elvis.mu.org>, Alfred Perlstein writes:
>Fixes format warnings.  Since there was so much... bitching about my
>last commit to something contrib/* I'm posting the fix here.
>
>Index: fla.c
>===================================================================
>RCS file: /home/ncvs/src/sys/contrib/dev/fla/fla.c,v
>retrieving revision 1.29
>diff -u -r1.29 fla.c
>--- fla.c	12 Sep 2001 08:36:59 -0000	1.29
>+++ fla.c	19 Mar 2002 20:22:51 -0000
> <at>  <at>  -181,8 +181,9  <at>  <at> 
> 	enum doc2k_work what;
> 
> 	if (fla_debug > 1)
>-		printf("flastrategy(%p) %s %x, %d, %ld, %p)\n",
>-		    bp, devtoname(bp->bio_dev), bp->bio_flags, bp->bio_blkno, 
>+		printf("flastrategy(%p) %s %x, %lld, %ld, %p)\n",
>+		    bp, devtoname(bp->bio_dev), bp->bio_flags,
>+		    (long long)bp->bio_blkno, 
> 		    bp->bio_bcount / DEV_BSIZE, bp->bio_data);
> 
> 	sc = bp->bio_dev->si_drv1;
> <at>  <at>  -225,8 +226,9  <at>  <at> 
> 		ENTER();
(Continue reading)

Alfred Perlstein | 19 Mar 2002 21:45

Re: fla LINT breakage.

* Poul-Henning Kamp <phk <at> critter.freebsd.dk> [020319 12:40] wrote:
> 
> You're welcome to commit it :-)

Thank you.

> 
> Poul-Henning
> 
> In message <20020319201901.GJ455 <at> elvis.mu.org>, Alfred Perlstein writes:
> >Fixes format warnings.  Since there was so much... bitching about my
> >last commit to something contrib/* I'm posting the fix here.
> >
> >Index: fla.c

To Unsubscribe: send mail to majordomo <at> FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message


Gmane