cause a newline there. It's possible that the apparent
newline is an artefact of the cut and paste process in between the
build and your email.
You could try adding "set -x" at the beginning of the shell
function, inside the definition of the MAKEDIRTARGET variable in
src/share/mk/bsd.own.mk. THis might help you to figure out which
of the commands is failing.
--apb (Alan Barrett)
On Tue, 06 Aug 2013, Paul Goyette wrote:
udf.lo: In function `udf_makefs':
udf.c:(.text+0x1ded): undefined reference to `snprintb'
I am adding snprintb to src/tools/compat. That should fix this problem.
--apb (Alan Barrett)
On Wed, 07 Aug 2013, Andreas Gustafsson wrote:
Alan Barrett wrote:
I am adding snprintb to src/tools/compat. That should fix this problem.
Alas, the build now fails even sooner, with the error message:
checking lex output file root... configure: error: cannot find output from
/tmp/bracket
On Wed, 07 Aug 2013, Alan Barrett wrote:
More log output at:
http://releng.netbsd.org/b5reports/i386/build/2013.08.07.05.19.40/build.log.tail
I am unable to replicate this failure. Is it a clean build, or a
MKUPDATE build?
I am still unable to replicate the failure, but uwe helped me
e tmpfs /dev directory that will be lost on reboot.
Running "mount" by itself shows
tmpfs on /dev type tmpfs (union, local)
I replied to that first, because it'e the key to understanding
what went wrong.
--apb (Alan Barrett)
is no appropriate
/dev/dk* node?
Why not just make more /dev/dk* nodes?
--apb (Alan Barrett)
ng at the log file in
/var/run/rc.log.
--apb (Alan Barrett)
obably not run the fsck from one OS on a disk
from another OS.
Most valuable part of this USB-stick NetBSD installation is
subversion (svn), which I don't want to disable by high-risk
fixes to /dev.
I don't think the /dev fixes are high risk.
--apb (Alan Barrett)
On Sat, 19 Oct 2013, Alan Barrett wrote:
Now that the tmpfs file system has been removed, your /dev directory
should be on the underlying disk. To make device nodes there, use
commands like this:
cd /dev ./MAKEDEV all ./MAKEDEV dk16 dk17 dk18 dk19
That was supposed to be multiple
d;
mv /mnt/dev.new /mnt/dev;
Later, delete the /dev.old directory.
--apb (Alan Barrett)
On Fri, 01 Nov 2013, Thomas Klausner wrote:
Sometime between 20131023 and 20131031, /etc/release was broken.
Nearly all values are recorded as "(undefined)". It now looks like
this for me (with the same build command line as before):
That would be my fault. I'll fix it.
--apb (Alan Barrett)
On Fri, 01 Nov 2013, Alan Barrett wrote:
On Fri, 01 Nov 2013, Thomas Klausner wrote:
Sometime between 20131023 and 20131031, /etc/release was broken.
Nearly all values are recorded as "(undefined)". It now looks like
this for me (with the same build command line as before):
That w
ate, unless
somebody else does so first.
--apb (Alan Barrett)
esr/faqs/smart-questions.html>.
--apb (Alan Barrett)
out a clean source tree);
* no left over output files or temporary files from previous build
attempts, such as TOOLDIR, RELEASEDIR, DESTDIR, or obj dirs;
* no /etc/mk.conf or similar file;
* no -V options passed to build.sh;
* no environment variables that are intended to affect the build.
--apb (Alan Barrett)
TA1/netbsd-HEAD/usr/src/external/bsd/iscsi/lib
Is that "unable to rename temporary" message the very first error
message, or were there others earlier? (There will be other
messages relating to initiator.po earlier in your build log, in
the part that you did not paste.)
Is this probl
e NetBSD tree
performed a "partial update", perhaps by forgetting to commit
something, then please explain in more detail, or just update to
a source tree from a date and time when it was working. People
do make mistakes, but anything that breaks the build on common
platforms is usually noticed and fixed quickly.
--apb (Alan Barrett)
hould allow building NetBSD from
FreeBSD, or building NetBSD from another version of NetBSD, or
many other things.
It is theoretically possible to use a different compiler, but
that's usually done only by people who are porting NetBSD to a new
platform, porting a new compiler to NetBSD, or something similarly
tricky.
--apb (Alan Barrett)
gure a smart host then the
"relayhost" paremeter in /etc/postfix/main.cf is the place to do
it.
If you don't have a working mailer, then you can use the web
interface to send-pr, as kre mentioned.
--apb (Alan Barrett)
virtual terminal, no more
response to keyboard input, though most, but not all the time, unplugging the
mouse brings back the keyboard.
Please start a new thread with a new Subject for each bug.
--apb (Alan Barrett)
On Sun, 29 Dec 2013, Thomas Mueller wrote:
Basically, I'm lost with the debugger, and the best I can do is type "reboot".
You can write down what's on the screen.
Please read and follow the advice in "How to report bugs
effectively" .
--apb (Alan Barrett)
me other OS to copy
the crash dump from the raw dump partition to a file, but I don't
know of any suitable tool that already exists.
--apb (Alan Barrett)
':
mdoc_validate.c:(.text+0x256a): undefined reference to `strnlen'
strnlen needs to be added to src/tools/compat. I will do that.
--apb (Alan Barrett)
On Thu, 09 Jan 2014, Alan Barrett wrote:
On Wed, 08 Jan 2014, John D. Baker wrote:
# link mandoc/mandoc
[...]
mdoc_validate.lo: In function `post_hyph':
mdoc_validate.c:(.text+0x256a): undefined reference to `strnlen'
strnlen needs to be added to src/tools/compat. I will do
of src/tools/compat/compat_defs.h.
--apb (Alan Barrett)
until your problem is resolved.
Yes, if you know that sources of a particular date/time have problems,
then don't build sources from that date/time.
--apb (Alan Barrett)
be enough for the
chroot. Alternatively, ntp could be taught to fall back to 123 if
the name lookup fails.
--apb (Alan Barrett)
ct, at the
expense of detecting the error later than before.
--apb (Alan Barrett)
cript that logged all uses of itself, and where HOST_SH pointed
to a real shell under a name other than /bin/sh, and I fixed all
the problems that were exposed. It seems to be time to do that
again, to fix new problems that have been introduced since the
previous sweep.
--apb (Alan Barrett)
y put this somewhere on the wiki, everybody please feel free
to properly document it. Same place should document MKDEBUG=yes and
MKDEBUGLIB=yes for userland.
I see the wiki as extra documentation, possibly more like a
tutorial or how-to than a reference. The reference documentation
should be in-tree.
--apb (Alan Barrett)
s is a new
installation? Perhaps check whether the man db's timestamp is
newer than /etc/mtree/set.man ?
--apb (Alan Barrett)
On Mon, 09 Jun 2014, Thomas Mueller wrote:
Is there any way to run etcupdate for NetBSD on other than the current root?
No, there is not.
Adding DESTDIR support to etcupdate is on my to-do list.
--apb (Alan Barrett)
ut files).
Sorry if this sounds too much like a “CVS questions” or
“documentation concerns” email, but I’d like to save some
frustration and get some actual code written, not fight with the
version control. If I should split this mail up, I’ll be happy
to do so.
It's easier to deal with one question at a time, unless the questions
are closely related.
--apb (Alan Barrett)
On Mon, 09 Jun 2014, Thomas Mueller wrote:
Is there any way to run etcupdate for NetBSD on other than the current root?
Please try revision 1.56 of etcupdate, which now takes a "-d destdir"
option.
--apb (Alan Barrett)
ement. For example, a three-way
merge between the older distribution version of the file, the
newer distribution version, and your version, would be a lot
more useful than the current two-way merge between the newer
distribution version and your version.
--apb (Alan Barrett)
ions
from that file. You could try adding these three lines near the
top of src/external/bsd/pcc/dist/pcc/os/netbsd/ccconfig.h:
#if HAVE_NBTOOL_CONFIG_H
#include "nbtool_config.h"
#endif
but I don't know whether that will be enough.
--apb (Alan Barrett)
ing in kernel or standalone environment
417 #endif
You'd obviously need to add some code for "#elif
defined(__NEW_COMPILER__)" and "#define __RENAME(x)
whatever_the_new_compiler_needs". Similarly for many other
macros.
--apb (Alan Barrett)
]
Please could you document this somewhere. Some possible places
that come to mind are sys/arch/README, sys/arch/*arm*/???,
distrib/notes/*arm*/install.
--apb (Alan Barrett)
MACHINE_ARCH=earmv4 ALIAS=hpcearmv4 DEFAULT
The MACHINE_ARCH tagged "DEFAULT" is used if you run build.sh with
"-m MACHINE" and without "-a MACHINE_ARCH". The "ALIAS" is usable
as a shorthand, where running build.sh with "-m ALIAS" implies "-m
MACHINE -a MACHINE_ARCH".
--apb (Alan Barrett)
missing at the server end, or did something get
screwed at my end?
I had the same problem yesterday. The repository was changed in an unsafe
manner after a CVS import into the wrong directory.
I have added a note about this to src/UPDATING.
--apb (Alan Barrett)
... I'd keep the new ALIAS=evbearm* and remove the old ALIAS=evbarm*
lines, not provide both.
--apb (Alan Barrett)
thing.
If backward compatibility in alias names is important, then let's
at least invent a DEPRECATED keyword to mark the old aliases, or
use OALIAS= instead of ALIAS=. Adding a new keyword will not
require code changes, whereas adding OALIAS= will require a small
patch, such as the attach
the alias names were changed?
--apb (Alan Barrett)
section of /etc/X11/xorg.conf.
--apb (Alan Barrett)
my private git repository
branch foo, commit deadc0ffee, date 2014-08-03,
based on NetBSD-current as of 2014-08-03 10:00 UTC
--apb (Alan Barrett)
netbsd.org/ports/evbarm/
It would be nice if somebody documented all this.
--apb (Alan Barrett)
a kernel version change.
What are the exact build.sh commands that you use?
--apb (Alan Barrett)
n a colon instead of just end-of-word at that
point.
--apb (Alan Barrett)
On Wed, 06 Aug 2014, Martin Husemann wrote:
On Wed, Aug 06, 2014 at 03:08:07PM +0200, Alan Barrett wrote:
I don't know what's going wrong for you. I am pretty sure it
works for me, though it's been some time since I tried a MKUPDATE
build (a.k.a. build.sh -u) across a kernel
dea than
I can get with twm.
Please don't mix multiple unrelated issues into the same message.
--apb (Alan Barrett)
On Thu, 21 Aug 2014, Thomas Klausner wrote:
I have scripts that mount images via vnd + cgd. So far I have been
using fixed numbers for the vnd and cgd devices I use, making sure I
don't use them for other purposes.
Now this doesn't work any longer, because as soon as vnd is created, a
dk device
way. Since the shall has the values as
strings already, I usually prefer to use a string test.
--apb (Alan Barrett)
number may be updated for some
other reason.
See the cvs log messages for src/sys/sys/param.h for reasons used
in the past.
--apb (Alan Barrett)
ink that this helps both people using the interface and
people changing the implementation.
--apb (Alan Barrett)
c/external/**
and src/sys/external/** which can be identified by searching for
"external" in all the makefiles under src/lib/libc, and perhaps
more that I missed.
--apb (Alan Barrett)
efile.inc. In
this case, it's rather difficult to work out, by examining
makfiles and build logs, that the KLINK_MACHINE variable and
related rules in bsd.klink.mk are responsible for creating the
DESTDIR/include/machine symlink, and that machine/asm.h comes from
src/sys/arch/${KLINK_MACH
ln, mv, mkdir, echo,
etc.), with a compiler and host include files and host libraries
that allows "build.sh tools" to work. If you have a host that is
mostly unix-like, but where build.sh tools fails, then I think
NetBSD will be interested in fixing that.
--apb (Alan Barrett)
out actually trying it. Again, if it fails on a system where
it's reasonable to expect it to work (say a unix-like system that
is able to build lots of other software), then any problems can be
fixed, as has happened many times in the past.
--apb (Alan Barrett)
which I like better :)
That seems fine. I wish you luck.
I would like to understand whether build.sh fails on your system,
and if so what can be done to fix it, or whether build.sh works
but you simply prefer to do something else instead.
--apb (Alan Barrett)
s like the correct command.
However, the fact that "grep -R BPF_COP" didn't find the expected
line in src/sys/net/bpf.h suggests that it didn't work.
the development version is 7.99.1?
Yes.
--apb (Alan Barrett)
strib/sets/maketars changes it to "./bin/\133", and then the
"${PAX} -dOw ..." command in line 199 fails to add the file to
base.tgz, and it also doesn't print an error message about failure
to add the file.
--apb (Alan Barrett)
On Sun, 19 Oct 2014, Alan Barrett wrote:
On Sat, 18 Oct 2014, Markus Kilbinger wrote:
While looking for old / outdated binaries in userland I noticed
that './bin/[' is no longer updated / part of base.tgz (or any
other *.tgz file) in HEAD and netbsd-7, though it still seems
to be
you have join.awk revision 1.4 or 1.6, then you should be OK, but
if you have revision 1.5 then you will have this problem.
--apb (Alan Barrett)
IPv6, and you
should further qualify the rules with "family" clauses that match
the desired address family, or "from" or "to" clauses that imply
an address family.
--apb (Alan Barrett)
documentation for ip6.conf,
and very little documentation for ipf(8)'s "-6" option. Most of
our differences seem to come down to your experience being at odds
with my reading of what little documentation exists. I have no
experience using IPv6 with ipf.
--apb (Alan Barrett)
{DESTDIR}/stand/${MACHINE}-xen/${subdir}
in addition to removing ${DESTDIR}/stand/${MACHINE}/${subdir} ?
--apb (Alan Barrett)
#x27;t
see much interest to indicate it was worth my time.
I am sorry I am unable to assist, but I am very interested in having
a light desktop environment. At present, I use icewm as a light(ish)
window manager, but I do not have a GUI file manager or network
configurator, and I would like those.
--apb (Alan Barrett)
nevertheless
fails as you indocated, then we have a problem to investigate.
--apb (Alan Barrett)
On Mon, 09 Feb 2015, Kamil Rytarowski wrote:
How do you checkout / update current src?
anon...@anoncvs.netbsd.org:/cvsroot
cvs -q update -Pd over src/
Without "-A", it's possible that some parts of your tree are stuck in
a different branch, or with a sticky date.
--apb (Alan Barrett)
tello(), pread(), and pwrite(), should
be declared when _POSIX_C_SOURCE >= 200112L. I see that
NetBSD-current fails to do so. The obvious addition of "||
(_POSIX_C_SOURCE - 0) >= 200112L" in appropriate places in stdio.h
and unistd.h should fix it.
--apb (Alan Barrett)
On Mon, 23 Mar 2015, Thomas Klausner wrote:
Thanks. So if it survives a build, can I commit the attached patch?
Yes, that looks fine.
--apb (Alan Barrett)
userland from 2015-01-27 00:00 UTC. A kernel from 2015-01-27
00:00 UTC does not have this problem.
--apb (Alan Barrett)
72 matches
Mail list logo
