random from within the range at file read time. Very useful if you
want to avoid having a fleet of machines melt a server by all trying to
contact it at a precise time every morning. See docs for details.
Reviewed by: christos, apb, vixie, others.
XXX apb suggests, quite reasonably, that ?10-16/2 should mean
something like 10,12,14,16 or 11,13,15. I'm too lazy to do it right
now, but it should be done.
XXX vixie suggests, quite reasonably, that if you're using "?" one
should delay randomly by 0-59 seconds. In the modern NTP world, you
could imagine that with a million well synchronized machines the
second just at the minute would be hit quite hard. I'm too lazy to do
it right now, but it should be done.
XXX cron needs to be updated to Vixie's cron 4.1 code.
As the input is read from a specfile into a tree of linked lists,
keep each linked list sorted. The sort order is the same as that
already used by "mtree -c": directories sort after non-directories, but
otherwise names are sorted in the order used by strcmp().
RT_ prefix and use them appropriately, instead of making copies. Make
pppd use the RT_ROUNDUP macro; fixes proxyarp setting on 64 bit hosts.
XXX: All this should be pulled up to 5.0
so that it can return 128KB, 256KB and 512KB properly instead of
truncated 0MB or rounded 1MB.
Problem reported by nisimura@ on port-amd64 and port-i386.
ddb.onpanic to 1, change it back to 0 in sysctl.conf and make sure
postinstall installs this setting.
This avoids us trying to dump while booting from install CD, but keeps
the default the same once we are far enough through /etc/rc.d. Failing
earlier is unlikely to be recovered by an automatic reboot.
OK: core.
ddb running on crash dumps, but with two notable changes:
- Breakpoints, watches, etc are obviously never going to work so they
are not handled.
- You can pipe output to the shell, e.g. ps | grep foo
Items remaining to be done:
- Port it to architectures other than i386. This isn't difficult, just
a case of making db_disasm.c/db_trace.c or their equivalent compile
and work.
- Make more of the "show" commands work, e.g "show uvmexp".
NetBSD conventional bswap16, bswap32, and bswap64 functions or macros.
In a non-tools build, include <sys/types.h> and <machine/bswap.h> to
get definitions of these functions. In a tools build, rely on the
functions or macros being provided by nbtool_config.h, but don't incluce
<machine/bswap.h> (which might not exist in a cross build).
This should address a problem building on OpenBSD, which has swap16,
swap32, and swap64 macros that conflicted with the local definitions.
The problem was reported by Alexander Bluhm, but this patch is quite
different from the one he suggested.
return something other than just 0.
caveat: statvfs is done for the mountpoint path, so might not give
the truth about a directory inside the mountpoint.
made to fail. Specifically, change
.ifdef(SYMBOL) -> .ifdef SYMBOL or .if defined(SYMBOL),
and corresponding for .ifndef.
Also correct one error in lib/libm/Makefile (.ifdef (${MKCOMPLEX} != "no")?!?).
This is done by adding the general ability to stuff non-SUSP data into
the end of the System Use field of a Directory Record, which required
some amount of rewriting of the SUSP support. I think the result is
at least as good as what came before, and I've fixed at least one bug
along the way. Tested against RISC OS 3.70 on my Risc PC.
Oh, why do we want it? It should allow us to make acorn{26,32} CDs that
can be booted directly from RISC OS without mucking around copying the
bootloader to a native filing system.
the System Use field with fewer then 28 bytes to spare, we were
remembering the wrong length for the System Use field and hence
emitting a corrupt directory entry. This could be triggered by trying
to build a filesystem containing a regular file with a 120-byte name.
Now we're a little more careful.
Continuation Area, be sure to insert the "CE" record at the correct
point in the chain of SUSP records. Without this, makefs emits
an invalid image if any file needs a Continuation Area. The bug
seems to have been introduced with the conversion to TAILQs.
without supplying a value, complain at them instead of segfaulting.
Or, to put it another way, check that the value is non-NULL before calling
set_option().
ok'ed by core and releng.
(thanks for agc@, snj@ and i'm sorry for long time patience).
[libc]
- localeio.[ch] and lc*.[ch] in src/lib/libc/locale was replaced by
new locale-db implementation using citrus_db backend,
see src/lib/libc/citrus/citrus_lc_*.[ch].
- add citrus_bcs_strtou?l.c. don't use strtou?l locale implementation
internally, because they're locale-aware function.
- add some stubs for multi-locale issue, see {current,global}_locale.c.
- remove some obsolete file, setrunelocale.c, ___runetype_mb.c.
- remove __savectype() from ctypeio.[ch].
[tools]
- mklocale(1): add new option ``-t'' that generates new style
LC_{MONETARY,NUMERIC,TIME,MESSAGES} locale-db format.
- chrtbl(1): added ctypeio.[ch] for __savectype().
[locale-db]
- added en_US.US-ASCII locale.
- removed some shareable locale definition file:
en_US.US-ASCII -> en_US.ISO8859-1, en_US.UTF-8
zh_CN.eucCN -> zh_CN.GB18030
and more...see src/share/locale/*/Makefile.
- remove obsoleted locale sr_YU, added new locale sr_ME, sr_RS.
- change locale name ja_JP.ISO2022-JP* -> ja_JP.ISO-2022-JP*
for X11's locale.alias file alignments.
- fix regression test, wrong wcs?width(3), NAN/INF usage.
i tested release-build following arch:
i386, amd64, hpc{mips,arm,sh}, sparc64, vax.
citrus_lc_*.[ch] also can read old-plain-text style locale-db.
so that backward compatibility is keeped, but lc*.[ch] can't read
new citrus_db'ed locale-db and localeio.c never check sanity,
so forward compatibility is broken ;-<
old mklocale(1) doesn't know -t option, so you have to rebuild toolchain.
- Make MD poffd(8) retire, and use MI powerd(8) instead of it.
- Make /dev/pow1 retire, because nobody holds /dev/pow0 any longer.
Use /dev/pow0 for pow(4) ioctl.
- POWIOCSSIGNAL ioctl which is for poffd(8) is also obsoleted.
specific set of user-derived variables, to be used in SOURCEMODE.
- In SOURCEMODE, generate the rc.d scripts xdm and xfs.
- Auto-detect if X11 sets are used (either through the value of MKX11 in
SOURCEMODE, or by finding an xetc-xpecific file in sets mode).
- Ignore X11-specific rc.d scripts if X11 is not used.
Syslogd does not properly handle:
1) the ADDDATE flag which is set with -T invocation and when messages
come from the kernel. Other cases where it is set it is ignored
as timestamping is always done (e.g. logmsg_async())
2) the variable found_ts in check_timestamp(). It would determine
whether or not the message had a (possibly valid) timestamp, set
found_ts to true, then ignore that in most cases. If we can't find
a timestamp return.
3) messages without a parsable timestamp should get one when outputting
the BSD syslog format so that a syslog-protocol timestamp isn't
injected (chopped off with BSD syslog length) giving something like:
"2008-11-27T15:0 cisco -: 1790:"
^ time might have been 2008-11-27T15:02:35.296497+11:00
4) syslog protocol version checking only checked for a leading numeral
one (1) then skipped two places (presuming a space). Messages sent
from some sources (e.g. my cisco) may be
"1795: Nov 27 04:12:52: %LINEPROTO-5-..."
which would be chopped to
"95: Nov 27 04:12:52: %LINEPROTO-5-..."
- new syslog protocol api syslogp(3) that supports structured data
and draft-rfc timestamps
- reliable tcp connections with queueing
- encrypted connections
Initialise memory before writing to only part of it,
making sure that the string is NUL-terminated
Allocate the maximum allowed length (including the
terminating NUL byte) for the temporary name string, make sure
the final output is NUL terminated, and copy back the NUL when
overwriting the previously created name (since maxlen is always
at least 1 less than ISO_FILENAME_MAXLENGTH_WITH_PADDING, this
is no problem)
Patches from Thorsten Glaser through freenode.net.
This may need more work to prevent warning messages during
"make cleandir" when the commands in "!=" assignments are executed
even though tools may not have been built.
lukem