(${KERNEL_SUFFIXES}), put the list in a shell variable and use that.
Both bash (before 2.05a) and Solaris /usr/xpg4/bin/sh report a syntax
error otherwise.
This occurs before the first load_rc_config() so that it may be
overridden by the user, and appears in single quotes so the
variables don't get evaluated until the eval in run_rc_command().
Problem noted by Patrick Welche <prlw1@cam.ac.uk> in [bin/15912].
Replace $critical_filesystems with $critical_filesystems_remote .
The new names are now consistent with the type argument that
mount_critical_filesystems() is called with, and allows for other types to
be easily supported by that function.
For backwards compatibility purposes, if the now obsolete variable is defined
(even empty), it takes precedence over the new form, and you will be warned.
If you want to stop the warnings, update your rc.conf(5) settings!
NETWORKING, and SERVERS) by specifying that certain things should
come BEFORE a given barrier, rather than having the barrier REQUIRE
a service. This allows scripts to be removed without having to
edit the barrier dependencies.
As discussed on tech-userlevel, and approved by Luke.
make them "externally" available:
Previous Current Purpose
-------- ------- -------
_arg rc_arg Argument to command, after fast/force
processing performed (and prefix
removed)
_flags rc_flags Flags to start the default command
with. Defaults to ${name}_flags,
unless overridden by $flags from the
environment. This variable may be
changed by the precmd method.
_pid rc_pid PID of command (if appropriate).
_rc_run_fast rc_fast Not empty is "fast" was provided.
_rc_run_force rc_force Not empty is "force" was provided.
- Use rc_flags instead of _flags or ${name}_flags in various rc.d scripts,
so that $flags from the environment overrides ${name}_flags from rc.conf(5).
Fixes [bin/15800].
If set to yes, block-type swap partitions will be deleted upon shutdown.
This can be useful if swapping onto a RAIDframe device, but may cause
unnecessary delays during shutdown for the general case, so it's
disabled by default.
Should resolve [bin/14433] and [kern/14769].
If set, the first line of $command is read, the leading #! is
removed, whitespace is normalised, the first word is checked
against $command_interpreter, and the result is used as the
string to match in the ps(1) output.
This code isn't infallable, but works for common cases such as
#!/bin/sh
#! /usr/pkg/bin/perl -w
This helps solve the problem described in [bin/15563], and by
Ed Ravin on tech-userlevel.
- Move the common elements of check_process() and check_pidfile() into
_find_processes(), and call appropriately.
- add libc/citrus directory.
this directory contains the common stuffs not only for locale
but also other facilities relating to i18n (e.g. iconv).
- To keep the binary compatibility for the ctype modules easily,
the interface between libc and the modules is simplified.
- For the future integrated extension, module pool is renamed
from "/usr/lib/runemodule" to "/usr/lib/i18n".
In the future, this directory will contain the modules for "iconv",
"collation", etc.
- some cosmetic changes.
- Bug fix for runetype.h; __attribute__((__packed__)) is placed at
the wrong position and it is invalid unintentionally.
But, the all members of the structures seem well-aligned. Thus,
this bug causes no problem, hopefully.
HEADS UP:
- /usr/lib/rumemodule is obsoleted. If you use the multibyte locales,
you need to install /usr/lib/i18n/* from sys/lib/i18n_module .
- The binary compatibility of /usr/share/locale/*/LC_CTYPE is probably kept.
Perhaps, the bug fix about __packed__ mentioned above breaks the
compatibility... Be careful especially on 64bit platforms.
become ippp (ISDN ppp) and irip (ISDN raw IP). The character device now
are called: /dev/isdn (isdnd <-> kernel communication), /dev/isdnctl (dialing
and other control), /dev/isdntrc* (tracing), /dev/isdnbchan* (raw B channel
access, i.e. for user land PPP) and /dev/isdntel* (telephone devices, i.e.
for answering machines).
* use "ln -fs foo bar" instead of "rm -f bar; ln -s foo bar"
* remove unnecessary chown root or chgrp wheel directives; older code here
didn't bother doing this, so why be inconsistent
* when making directories, don't use -p and don't redirect stdout to /dev/null
* clean up whitespace
* wrap the main parser code in a function makedev(), and call makedev
instead of "sh $0". (this is a bit faster)
* don't barf when we have fdesc mounted; just skip the devices fdesc provides
(these files should really be automagically generated...)
- Don't su(1) to ${name}_user if running as that user.
- Support $procname as the name of the process to search for.
Defaults to the value of $command.
I can't remember the rationale for why I made that change in the first place,
and there's been some good arguments presented to me as to why the rcvar
should be checked before status is run.
- always unset the variables used by run_rc_command() before sourcing an
rc.d script, rather than just when rc_fast_and_loose is in effect.
this allows run_rc_command() to be used in .sh scripts (with care).
- remove comment describing now-former limitation of .sh scripts.
- make the rc_fast_and_loose check only after the `` not *.sh / not scratch
/ script-is-executable '' checks. rc_fast_and_loose now behaves a lot more
like the normal operation, in that it skips scratch files and
non-executable files
you'll see the following message:
# /etc/rc.d/ipfs stop
/etc/rc.d/ipfs: WARNING: $ipfs is not set properly.
This horrible change is needed because of the "shutdown" keyword.
keep state to be locked (modification prevented) and then saved to disk,
allowing for the system to experience a reboot, followed by the restoration
of that information, resulting in connections not being interrupted.
To activate this feature, set ipfs=YES in /etc/rc.conf
(easier to read that large nested .if's
- use INSTALL_FILE, INSTALL_LINK and INSTALL_SYMLINK instead of INSTALL
- when creating release directories, don't bother setting owner and group
- wrap the main parser code in a function makedev(), and call makedev
instead of 'sh $0'. (this is a bit faster)
- remove unnecessary chown root or chgrp wheel directives; older code here
didn't bother doing this, so why be inconsistent
- when making directories, don't use -p and don't redirect stdout to /dev/null
"pty0 pty1 pty2 pty3" before, for a total of 64 device nodes, we now want
just "pty0" for a total of 62 -- the original commit had "pty0 pty1" for a
total of 124, which ate too many inodes and made "ls /dev" a bit messy.
Changes taken from the i386 MAKEDEV. Note that the unusual use of 'dd'
to index an array was chosen because 'dd' is on all ports' install media.
This form of this change comes from Andrew Brown; the original idea to
bump the number of ptys this way comes from Brian Marcotte at Panix.
The openpty() routine has been modified to use the extra pty names
before the "traditional" names, so that programs too dumb to use
openpty() will, hopefully, find one of the "traditional" ptys free
even if many others are in use. The modifications to MAKEDEV are
courtesy Andrew Brown, and are pretty clever: the unit numbers used
by the "traditional" names stay the same, to avoid trouble when
upgrading existing systems. The unusual use of "dd" to index an
array in MAKEDEV is because no other simple method seemed feasible
using only the programs on the install media for all ports.
install-etc-files iso_image_mi iso-image_md_pre iso-image_md_post snap_kern
- add the following targets to the RELEASEDIR=="" check:
iso_image_mi iso-image_md_pre iso-image_md_post snap_kern
- improve comments around make .flow control statements (including adding
some helper # { ... # } comments around large .if statements for (ab)use
with vi showmatch mode)
- clean up whitespace
at the end, and as wscons (actually ttyE0) is required to login on the console,
it is probably better that an out of space MAKEDEV fail on some other device.
and links exist:
${ntpd_chrootdir}/dev/clockctl
/var/db/ntp.drift -> ${ntpd_chrootdir}/var/db/ntp.drift
and then start ntpd with the appropriate options to run chroot(2)ed
under $ntpd_chrootdir as user ntpd group ntpd.
to take advantage of this, set ntpd_chrootdir in /etc/rc.conf.
[this is based on similar work i did for rc.d/named]
when building the kernel sets and placing gzip'd kernels in binary/kernels.
For example, if KERNEL_SUFFIXES were set to "ecoff srec", then the kernel
set would include:
netbsd
netbsd.ecoff (only if it exists in the kernel compile directory)
netbsd.srec (only if it exists in the kernel compile directory)
This is useful for packaging kernel sets for platforms which have
extra special requirements for loading the kernel.
more consistent. To quote the comment in etc/Makefile
that describes how it's done:
# This target builds the kernels specified by each port. A port may
# specify the following kernels:
#
# KERNEL_SETS The list of kernels that will be
# packaged into sets, named
# kern-${kernel}.tgz. These kernels
# are also placed in the binary/kernels
# area of the release package as
# netbsd-${kernel}.gz.
#
# EXTRA_KERNELS Additional kernels to place in the
# binary/kernels area of the release
# package as netbsd-${kernel}.gz, but
# which are not placed into sets. This
# allows a port to provide e.g. a netbootable
# installation kernel containing a ramdisk.
#
# BUILD_KERNELS Additional kernels to build which are
# not placed into sets nor into the
# binary/kernels area of the release
# package. These are typically kernels
# that are built for inclusion only in
# installation disk/CD-ROM/tape images.
#
the latter being called by the "distribution" target. This allows the
various /etc/... files to be installed manually in a convenient way, if
desired.
NOTE: It is INTENTIONAL that this target is not named "install".
or distribution.
Also, clean up the check and include <sys/endian.h> instead -- some platforms'
<machine/endian_machdep.h> pull in the definitions of _BIG_ENDIAN and
_LITTLE_ENDIAN, invalidating the test; this makes the check values uniformly
"4321" and "1234" respectively.
Shouldn't be needed, but install has no other good way to deal with
this.
Pointed out by Rob Windsor in PR 14394 -- I committed his patch plus
one for something he didn't hit yet.
the etc Makefile override that by putting USETOOLS into $.MAKEOVERRIDES
This way the default for kernel compiles is still to use the installed
toolchain instead of depending on $TOOLDIR. $TOOLDIR can be used by
simply adding USETOOLS=yes to the command line as usual.
Adjust each ports template to set the default no setting and also pull in
bsd.own.mk if they weren't already to ensure they'll build correctly
with the new toolchain setup.
csh: Permission denied
csh: Trying to start from "/var/log"
message.
This was caused by the
su -m uucp -c "uustat -a"
line being executed in a directory not readable by uucp. The login
shell implied by -m is of course root's shell, /bin/csh, which doesn't
like not being able to read the dir it is in, and thus the errors. By
temporarily cd'ing to /tmp the problem is fixed.
What is really needed, of course, is a way to tell su what shell you
want to use explicitly, especially for use in scripts where the
vagaries of which shell the login executing the script uses should not
be depended on. No such method exists. One should be added.
Indeed, it might also be nice to have a way of telling su to directly
execute a command with -c rather than using a shell to interpret the
command.
I cannot find any standards documents that specify su at the moment,
though. SuSv2 is silent on su(8).
- Resurrect /etc/changelist, even if it's an "empty" file by default,
because it's easier to use than /etc/mtree/special.local for adding
a couple of simple files. Back by popular demand (hi @@@! :-)
- Add /etc/rc.d/* to the list of "dynamic" files; this notices changes
in user-added scripts
- Only calculate the mtree -I nomail list once, and re-use
- Use "cat foo | while read file" instead of "for file in `cat foo`" ;
handles whitespace better...
Features:
- Add a bunch of stuff to /etc/mtree/special to enable removal of
/etc/changelist:
- files which we want to monitor for changes but don't want to
see the diffs of (master.passwd, ssh_host_key, ...) are
tagged with "nomail"
- files which we don't want to monitor are tagged with "exclude"
(such as netgroup.db, kvm.db, ...)
- monitor /etc/mtree/special.local, /root/.ssh/*
- remove /etc/changelist, and a bunch of XXX comments
- use mtree(8)'s -D, -I, and -E to generate lists of files to
actually do the changelist stuff on.
- support /etc/mtree/special.local as an optional user-provided
version of /etc/mtree/special (effectively, an enhanced
/etc/changelist)
- Add code to monitor: /etc/ifconfig.* /etc/raid*.conf /etc/rc.conf.d/*
including support for these files being added and removed at will.
- If /sbin/fdisk exists, backup the output of "fdisk $disk" for all
the active disk drives as part of $check_disklabels
- Check permissions on: ~/.ssh/* ~/.shosts
Details:
- Reorder initialisation of defaults
- Remove special case for /etc/master.passwd "monitor but don't email diffs"
with general case for other similar files.
- Keep all `autogenerated' files (such as disklabel.*, setuid.current, ...)
in "$backup_dir/work", to minimise name clashes.
- Add migrate_file(old, new) to do the hard work of migrating files
from the old `top level' /var/backups mechanism to the `full path'
mechanism recently added. Use this appropriately.
- Add backup_and_diff(file, printdiffs), to the hard work of backing-up
and diff-ing files.
- Cleanup use of shell redirects
- /bin/sh supports ~root globbing, so use it.
- Improve umask checking; use awk regex rather than awk math
