]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
26 years agoWe can now build libm on Alpha. There is very little MD alpha code.
jb [Sat, 10 Jan 1998 22:51:51 +0000 (22:51 +0000)]
We can now build libm on Alpha. There is very little MD alpha code.

26 years agoAvoid building x86 specific libraries on Alpha.
jb [Sat, 10 Jan 1998 22:50:00 +0000 (22:50 +0000)]
Avoid building x86 specific libraries on Alpha.

26 years agoPAL codes needed for asm.h
jb [Sat, 10 Jan 1998 22:42:21 +0000 (22:42 +0000)]
PAL codes needed for asm.h

26 years agoThis is the only alpha math source that NetBSD has.
jb [Sat, 10 Jan 1998 22:17:24 +0000 (22:17 +0000)]
This is the only alpha math source that NetBSD has.

26 years agoHeader files which are linked from /usr/include to /usr/include/machine.
jb [Sat, 10 Jan 1998 22:09:53 +0000 (22:09 +0000)]
Header files which are linked from /usr/include to /usr/include/machine.
NetBSD's endian.h needs sys/types.h for the typedefs that are used to
get htonl() correct when a long is 64 bits.

26 years agoPause according to the redial setting when dialing manually.
brian [Sat, 10 Jan 1998 21:51:31 +0000 (21:51 +0000)]
Pause according to the redial setting when dialing manually.
Pointed out by: Berend de Boer <berend@compuserve.com>

26 years agoA partial frontal lobotomy for find if using the NetBSD libc which
jb [Sat, 10 Jan 1998 21:36:34 +0000 (21:36 +0000)]
A partial frontal lobotomy for find if using the NetBSD libc which
doesn't know about getvfsbyname() and the vfsconf structure. This
disables the -fstype option if compiled with a pre-processor that
defines __NetBSD__. With the FreeBSD built pre-processor, find can only
be built with the FreeBSD libc. So when running with a NetBSD kernel,
FreeBSD's libc will have to return ENOSYS for things that NetBSD
doesn't support. That's life in a hybrid world.

26 years agoPass an actual empty environment to execle() as per POSIX rather than
alex [Sat, 10 Jan 1998 20:11:06 +0000 (20:11 +0000)]
Pass an actual empty environment to execle() as per POSIX rather than
rely on undocumented behavior.

The following fixes were obtained from OpenBSD:

  o -Wall fixes to tlist array initialization and assignment used
    as truth value.
  o Use a restricted environment.
  o Improved error message when shutdown fails to exec reboot or halt.

26 years agoPut back __libalias_version so ppp(8) build again.
steve [Sat, 10 Jan 1998 19:37:19 +0000 (19:37 +0000)]
Put back __libalias_version so ppp(8) build again.

26 years agoRename the pwd_mkdb(8) option '-c' to '-C' for better compatiblity
wosch [Sat, 10 Jan 1998 17:27:28 +0000 (17:27 +0000)]
Rename the pwd_mkdb(8) option '-c' to '-C' for better compatiblity
with BSD/OS.

26 years agoSync with ipfw interface change: fw_pts is now part of a union (a
alex [Sat, 10 Jan 1998 16:14:18 +0000 (16:14 +0000)]
Sync with ipfw interface change: fw_pts is now part of a union (a
necessary evil due to the 108 byte setsockopt() limit).

26 years agoMinor brace gymnastics so that /etc/wall_cmos_clock is really
steve [Sat, 10 Jan 1998 15:55:11 +0000 (15:55 +0000)]
Minor brace gymnastics so that /etc/wall_cmos_clock is really
created.

PR: 5473
Submitted by: MIHIRA Yoshiro <sanpei@yy.cs.keio.ac.jp>

26 years agoinclude <net/if.h> and restore this to sanity.
jkh [Sat, 10 Jan 1998 15:04:06 +0000 (15:04 +0000)]
include <net/if.h> and restore this to sanity.

26 years agoWhoops. softclock is called from doreti_swi as well. Abandon call from
phk [Sat, 10 Jan 1998 14:55:14 +0000 (14:55 +0000)]
Whoops.  softclock is called from doreti_swi as well.  Abandon call from
hardclock().

Forgot this:

Pointed hat sent by: bd

26 years agoWhoops. softclock is called from doreti_swi as well. Abandon call from
phk [Sat, 10 Jan 1998 14:54:05 +0000 (14:54 +0000)]
Whoops.  softclock is called from doreti_swi as well.  Abandon call from
hardclock().

26 years agoSync with sys/i386/isa/syscons.c revision 1.244.
kato [Sat, 10 Jan 1998 13:31:27 +0000 (13:31 +0000)]
Sync with sys/i386/isa/syscons.c revision 1.244.

26 years agoEffect the divorce of kern_clock.c and kern_timeout.c (which was
phk [Sat, 10 Jan 1998 13:16:26 +0000 (13:16 +0000)]
Effect the divorce of kern_clock.c and kern_timeout.c (which was
repository copied from kern_clock.c)

26 years agoThese are a few of the alpha machine dependent header files - the first
jb [Sat, 10 Jan 1998 10:13:16 +0000 (10:13 +0000)]
These are a few of the alpha machine dependent header files - the first
referenced by the build of user-space libraries. These files were
obtained from NetBSD (with ansi.h being modified to reflect the FreeBSD
off_t and pid_t implementation).

26 years agoBruce says that ${.CURDIR}/csu/${MACHINE} will prevent finding a
jb [Sat, 10 Jan 1998 09:09:24 +0000 (09:09 +0000)]
Bruce says that ${.CURDIR}/csu/${MACHINE} will prevent finding a
stale obj directory and we wouldn't want to do that! I trust he knows
what he's talking about. 8-)

Also avoid building libm at all until the NetBSD asm code is imported.
I wrongly commented this out last time. Oops.

26 years agoAdded PnP I/O port address of NEC PC-98.
kato [Sat, 10 Jan 1998 07:41:43 +0000 (07:41 +0000)]
Added PnP I/O port address of NEC PC-98.

Submitted by: Takahashi Yoshihiro <nyan@wyvern.cc.kogakuin.ac.jp>

26 years agoFixed change prerequisites for <net/if_arp.h>:
bde [Sat, 10 Jan 1998 07:29:10 +0000 (07:29 +0000)]
Fixed change prerequisites for <net/if_arp.h>:
- don't declare `struct arpcom' except in the kernel, so that there is no
  dependency on <net/if.h> except in the kernel.  This may break something
  else.
- spell ETHER_ADDR_LEN as 6 again, so that there is no dependency on
  <net/ethernet.h> even in the kernel.

26 years agoMake the ".set" directive copy the aux field when the expression
jdp [Sat, 10 Jan 1998 05:36:35 +0000 (05:36 +0000)]
Make the ".set" directive copy the aux field when the expression
reduces to a relocatable symbol plus an offset.  This preserves
the symbol type information (function vs. object).  It is important
for SVR4-style weak symbols, e.g., "#pragma weak foo=bar".  Without
this change, the linker complains that the jmpslot entry is not a
function.

26 years agoCompare return code from ipfw against 0 for success instead of == 1
alex [Sat, 10 Jan 1998 03:33:39 +0000 (03:33 +0000)]
Compare return code from ipfw against 0 for success instead of == 1
for error.

Pointed out by: Matthew Thyer <thyerm@camtech.net.au>

26 years agoImplement Reset{Req,Ack} properly, as per rfc 1962.
brian [Sat, 10 Jan 1998 01:55:11 +0000 (01:55 +0000)]
Implement Reset{Req,Ack} properly, as per rfc 1962.
(I completely mis-read the rfc last time 'round!)

This means:
  o Better CCP/WARN Reset diagnostics.
  o After we've sent a REQ and before we've received an ACK, we drop
    incoming compressed data and send another REQ.
  o Before sending an ACK, re-sequence all pending PRI_NORMAL data in
    the modem queue so that pending packets won't get to the peer
    *after* the ResetAck.
  o Send ACKs with the `identifier' from the REQ frame.
  o After we've received a correct ACK, duplicate ACKs are ok (and will
    reset our history).
  o Incorrect ACKs (not matching the last REQ) are moaned about and dropped.

Also,

  o Calculate the correct FCS after compressing a packet.  DEFLATE
    *may* produce an mbuf with more than a single link in the chain,
    but HdlcOutput didn't know how to calculate the FCS :-(
  o Make `struct fsm'::reqid a u_char, not an int.
    This fix will prevent us from sending id `255' 2,000,000,000 times
    before wrapping to `0' for another 2,000,000,000 sends :-/
  o Bump the version number a little.

The end result:  DEFLATE now works over an unreliable link layer.
                 I can txfr a 1.5Mb kernel over a (rather bad) null-modem
                 cable at an average of 21679 bytes per second using rcp.
Repeat after me: Don't test compression using a loopback ppp/tcp setup as
                 we never lose packets and therefore never have to reset!

26 years agoUse manpath(1) to determine the value of ${MANPATH} instead of
steve [Fri, 9 Jan 1998 23:56:45 +0000 (23:56 +0000)]
Use manpath(1) to determine the value of ${MANPATH} instead of
using one of two hard-coded values.

PR: 5365
Submitted by: Ruslan Shevchenko <rssh@cam.grad.kiev.ua>

26 years agoAllow this to compile with NetBSD tools.
jb [Fri, 9 Jan 1998 23:51:04 +0000 (23:51 +0000)]
Allow this to compile with NetBSD tools.

26 years agoTeach libalias to work with IPFW firewalls (controlled by a flag).
eivind [Fri, 9 Jan 1998 21:13:35 +0000 (21:13 +0000)]
Teach libalias to work with IPFW firewalls (controlled by a flag).

Obtained from: Yes development tree (+ 10 lines of patches from
Charles Mott, original libalias author)

26 years agoNetBSD has sys_errlist in errno.h, not stdio.h like FreeBSD.
jb [Fri, 9 Jan 1998 20:02:20 +0000 (20:02 +0000)]
NetBSD has sys_errlist in errno.h, not stdio.h like FreeBSD.

26 years agoDon't build libc, libc_r or libm on Alpha yet. We'll do the other
jb [Fri, 9 Jan 1998 19:46:30 +0000 (19:46 +0000)]
Don't build libc, libc_r or libm on Alpha yet. We'll do the other
libraries first and use NetBSD's libc/libm for now.

26 years agoFix an anomaly in the option descriptions.
steve [Fri, 9 Jan 1998 18:46:53 +0000 (18:46 +0000)]
Fix an anomaly in the option descriptions.

PR: 5469
Submitted by: NOKUBI Hirotaka <h-nokubi@yyy.or.jp>

26 years agoFix generation of Comtrol RocketPort devices.
steve [Fri, 9 Jan 1998 18:32:51 +0000 (18:32 +0000)]
Fix generation of Comtrol RocketPort devices.

PR: 5468
Submitted by: Steve Gericke <steveg@comtrol.com>

26 years agoMF22: Additional paranoia in cleanup.
jkh [Fri, 9 Jan 1998 14:52:18 +0000 (14:52 +0000)]
MF22: Additional paranoia in cleanup.

26 years agoRevert to previous behaviour, only generate NLST for explicit "nlist"
msmith [Fri, 9 Jan 1998 13:45:11 +0000 (13:45 +0000)]
Revert to previous behaviour, only generate NLST for explicit "nlist"
command.
Submitted by: wollman

26 years agoRevert the changes yet again, after some email from Bruce. Sorry.
sef [Fri, 9 Jan 1998 09:31:42 +0000 (09:31 +0000)]
Revert the changes yet again, after some email from Bruce.  Sorry.

26 years ago- Produce the accent letter if the user hits the accent key twice.
yokota [Fri, 9 Jan 1998 09:06:55 +0000 (09:06 +0000)]
- Produce the accent letter if the user hits the accent key twice.
  (accent_key + space does still print the accent letter too, as in
  the previous commit.)
  Requested by a couple of users.
- Clear the accent flag when the next_screen key is pressed.
- Added some comment lines regarding accent key processing.

26 years agoSync with sys/i386/conf/files.i386 revision 1.184.
kato [Fri, 9 Jan 1998 08:49:43 +0000 (08:49 +0000)]
Sync with sys/i386/conf/files.i386 revision 1.184.

26 years agoAdd __NetBSD__ in a couple of places to allow this program to
jb [Fri, 9 Jan 1998 06:32:54 +0000 (06:32 +0000)]
Add __NetBSD__ in a couple of places to allow this program to
compile under NetBSD, but behave like FreeBSD. For the time being,
look for NetBSD's cpp in /usr/bin.

26 years agoMake a couple of the stat flags dependent on the sys/stat.h header file
jb [Fri, 9 Jan 1998 06:14:59 +0000 (06:14 +0000)]
Make a couple of the stat flags dependent on the sys/stat.h header file
that this source is compiled against. This source is referenced by
install which is needed as a build tool and must be able to compile
against NetBSD headers and libraries if we have a hope of supporting
another architecture.

With this change, that's two working programs down and 3945 (?) to go.
The other one was make, but that didn't need any changes to work under
FreeBSD/Alpha. 8-)

26 years agoAllow install to build from FreeBSD source using NetBSD tools and
jb [Fri, 9 Jan 1998 06:05:13 +0000 (06:05 +0000)]
Allow install to build from FreeBSD source using NetBSD tools and
headers during the port of FreeBSD to a NetBSD supported architecture.
FreeBSD needs to be able to install itself from very early on.

26 years agoBuild lib/csu/${MACHINE} only if it exists so that when porting FreeBSD
jb [Fri, 9 Jan 1998 05:37:41 +0000 (05:37 +0000)]
Build lib/csu/${MACHINE} only if it exists so that when porting FreeBSD
to another architecture (in this case the Alpha) we can continue to use
the host csu objects (from NetBSD). This should be a non-function change
to FreeBSD/i386.

26 years agoMake the BOOTP family new-style options (in opt_bootp.h)
eivind [Fri, 9 Jan 1998 03:21:07 +0000 (03:21 +0000)]
Make the BOOTP family new-style options (in opt_bootp.h)

26 years agoMinor fixups after INET option change.
eivind [Fri, 9 Jan 1998 01:01:21 +0000 (01:01 +0000)]
Minor fixups after INET option change.

26 years agoNETATALK -> opt_atalk.h
eivind [Fri, 9 Jan 1998 00:51:57 +0000 (00:51 +0000)]
NETATALK -> opt_atalk.h

26 years agoProper way to do the previous mis-commit. Still not quite right, because
sef [Fri, 9 Jan 1998 00:39:10 +0000 (00:39 +0000)]
Proper way to do the previous mis-commit.  Still not quite right, because
some header files (e.g., <err.h>) include <machine/something.h>, and this
will not pick up the right header files, so it may be removed eventually
anyway.  But some people who are not willing to build the right way
apparantly want this, so this is for them.

26 years agoGet rid of the bogus include -- it is incomplete (as it doesn't handle
sef [Fri, 9 Jan 1998 00:22:50 +0000 (00:22 +0000)]
Get rid of the bogus include -- it is incomplete (as it doesn't handle
anything other than <sys/*.h>), and unnecessary in most cases.  (The
situations where it is necesary can be dealt with by manually-made symlinks,
which is acceptable since they should only occur during testing.  Remember:
the tree does not compile well if you do not have matching header files
installed.  Half-baked -I directives don't cover enough of the cases.)

26 years agoMake the LKMs compile with the INET option as a newstyle option in
eivind [Thu, 8 Jan 1998 23:50:27 +0000 (23:50 +0000)]
Make the LKMs compile with the INET option as a newstyle option in
opt_inet.h.

26 years agoZap any addresses on the network side of the `tun' as soon
brian [Thu, 8 Jan 1998 23:47:57 +0000 (23:47 +0000)]
Zap any addresses on the network side of the `tun' as soon
as we've successfully opened the /dev side.

26 years agoMake INET a proper option.
eivind [Thu, 8 Jan 1998 23:42:31 +0000 (23:42 +0000)]
Make INET a proper option.

This will not make any of object files that LINT create change; there
might be differences with INET disabled, but hardly anything compiled
before without INET anyway.  Now the 'obvious' things will give a
proper error if compiled without inet - ipx_ip, ipfw, tcp_debug.  The
only thing that _should_ work (but can't be made to compile reasonably
easily) is sppp :-(

This commit move struct arpcom from <netinet/if_ether.h> to
<net/if_arp.h>.

26 years agoAdd support for dead keys in the French ISO keymap.
roberto [Thu, 8 Jan 1998 23:15:53 +0000 (23:15 +0000)]
Add support for dead keys in the French ISO keymap.

Submitted by: Pierre Beyssac <pb@fasterix.freenix.org>

26 years agoupdate the AWE32 wave table driver to Iwai's 0.4.2c version. This also
jmg [Thu, 8 Jan 1998 23:13:22 +0000 (23:13 +0000)]
update the AWE32 wave table driver to Iwai's 0.4.2c version.  This also
includes the patches to make it work under -current from Randall Hopper.

Remove the old AWE driver.

26 years agoThis commit was generated by cvs2svn to compensate for changes in r32345,
jmg [Thu, 8 Jan 1998 22:17:17 +0000 (22:17 +0000)]
This commit was generated by cvs2svn to compensate for changes in r32345,
which included commits to RCS files with non-trunk default branches.

26 years agoimport of GPL'd AWE32 Sound Driver v0.4.2c
jmg [Thu, 8 Jan 1998 22:17:17 +0000 (22:17 +0000)]
import of GPL'd AWE32 Sound Driver v0.4.2c

Submitted by: Randall Hopper

26 years agoAdded environment parameter to execle calls.
alex [Thu, 8 Jan 1998 20:05:45 +0000 (20:05 +0000)]
Added environment parameter to execle calls.

Obtained from: NetBSD PR2737 (augustss@cs.chalmers.se) via OpenBSD

26 years agoReflect reality when documenting the link-level flags for 3c509 cards.
joerg [Thu, 8 Jan 1998 17:09:38 +0000 (17:09 +0000)]
Reflect reality when documenting the link-level flags for 3c509 cards.

26 years agoStart syslogd before savecore, so savecore's messages will be logged
joerg [Thu, 8 Jan 1998 17:08:22 +0000 (17:08 +0000)]
Start syslogd before savecore, so savecore's messages will be logged
properly.  Only users of 4 MB machines might not want it this way (to
prevent accidentally clobbering the coredump), but they are somewhat
in the minority these days.

26 years agoDon't assume install script is directly runnable - use sh -c.
jkh [Thu, 8 Jan 1998 11:27:06 +0000 (11:27 +0000)]
Don't assume install script is directly runnable - use sh -c.

26 years agoSync with sys/i386/isa/syscons.c revision 1.243.
kato [Thu, 8 Jan 1998 10:50:46 +0000 (10:50 +0000)]
Sync with sys/i386/isa/syscons.c revision 1.243.

26 years agoSync with sys/i386/isa/sio.c revision 1.194.
kato [Thu, 8 Jan 1998 10:50:06 +0000 (10:50 +0000)]
Sync with sys/i386/isa/sio.c revision 1.194.

26 years ago1. Supporting a bogus 16550A compatible PCMCIA CARD stuffs
amurai [Thu, 8 Jan 1998 04:53:43 +0000 (04:53 +0000)]
1. Supporting a bogus 16550A compatible PCMCIA CARD stuffs
    - IIR_TXRDY is never off even if reading a IIR register.
    - Know as PIAFS "Palido 321S", "DC-*S" oemed by Sharp corp.

 2. Omiting a restrict probing if it's already probed by pccardd.
    Note: Define a new id_flags as follows

           0x40000 - NO PROBE (Already probed as serial)
           0x80000 - Has a bogus IIR_TXRDY register

Sato Junichi <junichi@astec.co.jp>
Nrihiro Kumagai <kuma@slab.tnr.sharp.co.jp>
Hirao Tetsuya <ai.cs.fujitsu.co.jp>
Toshiharu Asai <asai@mbc.infoshere.or.jp>
Shin'ya Kumabuchi <kumabu@t3.rim.or.jp>
        Freebsd-users-jp@jp.freebsd.org
bsd-nomads@ai.cs.fujitsu.co.jp

26 years agoBump up packet and byte counters to 64-bit unsigned ints. As a
alex [Thu, 8 Jan 1998 03:03:54 +0000 (03:03 +0000)]
Bump up packet and byte counters to 64-bit unsigned ints.  As a
consequence, ipfw's list command now adjusts its output at runtime
based on the largest packet/byte counter values.

NOTE:
  o The ipfw struct has changed requiring a recompile of both kernel
    and userland ipfw utility.

  o This probably should not be brought into 2.2.

PR: 3738

26 years agoCorrect a comment from my last commit.
alex [Thu, 8 Jan 1998 02:23:59 +0000 (02:23 +0000)]
Correct a comment from my last commit.

26 years agoAllow 'shutdown datespec' to work into the next century. Handle dates
alex [Thu, 8 Jan 1998 02:21:30 +0000 (02:21 +0000)]
Allow 'shutdown datespec' to work into the next century.  Handle dates
in the 22nd century and beyond even though it's irrelevant with a 32-bit
time_t which expires in the year 2038.

26 years agoDefine integer types added by POSIX.1g. We already had int8_t, int16_t,
alex [Thu, 8 Jan 1998 00:27:34 +0000 (00:27 +0000)]
Define integer types added by POSIX.1g.  We already had int8_t, int16_t,
and int32_t -- this adds their unsigned counterparts uint8_t, uint16_t,
and uint32_t.  Bruce has noted that C9x also defines these types (in a
different file), so this may have to be revisited with some ifdefs at a
later date.

Reviewed by: bde

26 years agoFormat mismatch in error message.
alex [Thu, 8 Jan 1998 00:27:31 +0000 (00:27 +0000)]
Format mismatch in error message.

Submitted by: bde

26 years agoOops, back out rev 1.5, it wasn't my intention to also commit the
joerg [Wed, 7 Jan 1998 22:32:28 +0000 (22:32 +0000)]
Oops, back out rev 1.5, it wasn't my intention to also commit the
-g. ;)

26 years agoUse the correct value in the call to swab(3) with conv=swab. Previously,
joerg [Wed, 7 Jan 1998 22:31:19 +0000 (22:31 +0000)]
Use the correct value in the call to swab(3) with conv=swab.  Previously,
dd if=/dev/zero of=/dev=null obs=23520 conv=swab
coredumped.

Please somebody review it, i'm not 105 % sure i'm understanding all
this mess correctly.

Detected by: Holm Tiffe <holm@geophysik.tu-freiberg.de>

26 years agoImprove hardpps readability a bit:
phk [Wed, 7 Jan 1998 12:29:17 +0000 (12:29 +0000)]
Improve hardpps readability a bit:
* Rename usec to p_usec so you can search for it.
* Macroize the huge median_of_3_samples if statement.

26 years agoMention S/key and the skey /etc files.
obrien [Wed, 7 Jan 1998 10:42:57 +0000 (10:42 +0000)]
Mention S/key and the skey /etc files.

26 years agoDisable io optimizations again, minor bug found, and will be fixed in
dyson [Wed, 7 Jan 1998 09:26:29 +0000 (09:26 +0000)]
Disable io optimizations again, minor bug found, and will be fixed in
a few days.

26 years ago1: Fixed compile error when WB_CACH is defined.
kato [Wed, 7 Jan 1998 09:03:42 +0000 (09:03 +0000)]
1: Fixed compile error when WB_CACH is defined.
2: Fixed !M_EPSON_PC98 case.

Reviewed by: Chiharu Shibata <chi@rd.njk.co.jp>

26 years agoicelandic.iso.kbd
yokota [Wed, 7 Jan 1998 08:46:53 +0000 (08:46 +0000)]
icelandic.iso.kbd
- New, Icelandic keymap with accent (dead) key definitions.
  Based on the work done by totii@est.is.

spanish iso.kbd
- Added accent (dead) key definitions.
  Based on the work done by jmrueda@diatel.upm.es.

26 years agoAdded accent (dead) key support to syscons and kbdcontrol.
yokota [Wed, 7 Jan 1998 08:43:28 +0000 (08:43 +0000)]
Added accent (dead) key support to syscons and kbdcontrol.

With a keymap with accent key definitions loaded to syscons, you press
an accent key followed by a regular letter key to produce an accented
letter.  Press an accent key followed by the space bar to get the
accent letter itself.

Code is based on the ideas and work by jmrueda@diatel.upm.es and
totii@est.is.

PR: i386/4016

- Added keywords for accent (dead) keys: dgra, dacu, dcir, dtil...
- Recognize accent map definitions.
<accent_map_definition> ::= <accent_key_name> <accent_char> <accent_map>
<accent_key_name> ::= dgra | dacu | dcir | dtil | dmac | dbre | ddot |
                      duml | dsla | drin | dced | dapo | ddac | dogo |
                      dcar
<accent_map> ::= <map_entry>
               | <map_entry> <accent_map>
<map_entry> ::= ( <regular_letter_char> <accented_char> )
- Use ioctls PIO_DEADKEYMAP and GIO_DEADKEYMAP to set and get the accent
  key map table in syscons.
- Made the output for the -L option more intelligible and look like
  initializers in kbdtables.h.
- Reorganized print functions in order to print the accent key map.

26 years agoAdded accent (dead) key support to syscons and kbdcontrol.
yokota [Wed, 7 Jan 1998 08:40:34 +0000 (08:40 +0000)]
Added accent (dead) key support to syscons and kbdcontrol.

With a keymap with accent key definitions loaded to syscons, you press
an accent key followed by a regular letter key to produce an accented
letter.  Press an accent key followed by the space bar to get the
accent letter itself.

Code is based on the ideas and work by jmrueda@diatel.upm.es and
totii@est.is.

PR: i386/4016

console.h
- Defined structures and constants for accent (dead) keys.

syscons.c, kbdtables.h
- When an accent key is pressed, set the corresponding index to
  `accents'.  If the next key is the space key, produce the accent char
  itself.  Otherwise search the accent key map entry, indexed by
  `accents', for a matching pair of a regular char and an accented char.
- Added ioctl functions to set and get the accent key map (PIO_DEADKEYMAP
  and GIO_DEADKEYMAP).

26 years agoAdd missing dot.
charnier [Wed, 7 Jan 1998 07:56:15 +0000 (07:56 +0000)]
Add missing dot.

26 years agoRemove more unused #includes.
charnier [Wed, 7 Jan 1998 07:55:26 +0000 (07:55 +0000)]
Remove more unused #includes.
Obtained from: Bruce.

26 years agoRemove unused #includes and make it compile again with -DDEBUG. Use syslog()
charnier [Wed, 7 Jan 1998 07:54:18 +0000 (07:54 +0000)]
Remove unused #includes and make it compile again with -DDEBUG. Use syslog()
instead of errx() when being a daemon.

26 years agoRemove unused #include. Display if inetd started us in case of failure.
charnier [Wed, 7 Jan 1998 07:51:00 +0000 (07:51 +0000)]
Remove unused #include. Display if inetd started us in case of failure.

26 years agoSync usage and synopsis. Add rcsid.
charnier [Wed, 7 Jan 1998 07:46:33 +0000 (07:46 +0000)]
Sync usage and synopsis. Add rcsid.

26 years agoTypos.
charnier [Wed, 7 Jan 1998 07:44:01 +0000 (07:44 +0000)]
Typos.

26 years agoTypo.
charnier [Wed, 7 Jan 1998 07:43:04 +0000 (07:43 +0000)]
Typo.

26 years agoSomething important happened last year. And this file was last updated
charnier [Wed, 7 Jan 1998 07:40:54 +0000 (07:40 +0000)]
Something important happened last year. And this file was last updated
in 82.
PR: misc/5416

26 years agoinclude sys so it builds on 2.2.x
jmg [Wed, 7 Jan 1998 06:19:50 +0000 (06:19 +0000)]
include sys so it builds on 2.2.x
also, fix misspelling of -1 (as EOF for getopt)

26 years agoTurn off the VTEXT flag when an object is no longer referenced, so
dyson [Wed, 7 Jan 1998 03:12:19 +0000 (03:12 +0000)]
Turn off the VTEXT flag when an object is no longer referenced, so
that an executable that is no longer running can be written to.  Also,
clear the OBJ_OPT flag more often, when appropriate.

26 years agoSupport listing/showing specific rules supplied on the command line.
alex [Wed, 7 Jan 1998 02:23:04 +0000 (02:23 +0000)]
Support listing/showing specific rules supplied on the command line.

Use error codes from <sysexits.h>.

26 years agoRevert previous commit and SEE ALSO yp(4) instead.
steve [Wed, 7 Jan 1998 00:57:31 +0000 (00:57 +0000)]
Revert previous commit and SEE ALSO yp(4) instead.

26 years agohandle long usernames more carefully
imp [Wed, 7 Jan 1998 00:28:36 +0000 (00:28 +0000)]
handle long usernames more carefully
Reviewed by: guido
Obtained from: OpenBSD (Theo de Raadt)

26 years agoAvoid fclose(NULL).
imp [Wed, 7 Jan 1998 00:12:04 +0000 (00:12 +0000)]
Avoid fclose(NULL).
Submitted by:  OpenBSD (Theo de Raadt)

26 years agoWait for job to finish and exit if exec fails.
imp [Wed, 7 Jan 1998 00:05:20 +0000 (00:05 +0000)]
Wait for job to finish and exit if exec fails.
Obtained from: OpenBSD (Theo de Raadt)

26 years ago1. Don't reject 0145 if started at 22XX.
helbig [Tue, 6 Jan 1998 17:27:09 +0000 (17:27 +0000)]
1. Don't reject 0145 if started at 22XX.
   Applied suggested fix from Andrew Andrew <andrew@ugh.net.au> with
   some stylistic changes. Thanks.
2. #include <sys/time.h> -> #include <time.h>
3. Removed #include <sys/param.h>
4. Use setlocale(3) and strftime(3) instead of ctime(3).
5. Clean up -Wall warnings.
6. Make sure, time to leave are integral minutes if the argument
   is absolute. (i. e. without "+"). If started at 10:10:55 with
   argument "1020" it computed time to leave as 10:20:55 instead of
   10:20:00.

PR: 5395

26 years agoSync with sys/i386/conf/files.i386 revision 1.183.
kato [Tue, 6 Jan 1998 13:47:54 +0000 (13:47 +0000)]
Sync with sys/i386/conf/files.i386 revision 1.183.

26 years agoDelete a stray semicolon at the end. This only breaks gcc -pedantic
wosch [Tue, 6 Jan 1998 11:44:55 +0000 (11:44 +0000)]
Delete a stray semicolon at the end. This only breaks gcc -pedantic
because COPYRIGHT_HPFPLIB is appended last.
Submitted by: Bruce

26 years agodocumentation changes.
jamil [Tue, 6 Jan 1998 09:39:07 +0000 (09:39 +0000)]
documentation changes.

26 years agodocumentation changes.
jamil [Tue, 6 Jan 1998 09:27:37 +0000 (09:27 +0000)]
documentation changes.

26 years agothe pt driver should make one DEVFS entry for each one found rather than
julian [Tue, 6 Jan 1998 08:28:46 +0000 (08:28 +0000)]
the pt driver should make one DEVFS entry for each one found rather than
just one for the whole system..

26 years agoMake pstat.c compile, since in the new kernel code, VVMIO doesn't
dyson [Tue, 6 Jan 1998 05:33:28 +0000 (05:33 +0000)]
Make pstat.c compile, since in the new kernel code, VVMIO doesn't
exist anymore.

26 years agoMake our v_usecount vnode reference count work identically to the
dyson [Tue, 6 Jan 1998 05:26:17 +0000 (05:26 +0000)]
Make our v_usecount vnode reference count work identically to the
original BSD code.  The association between the vnode and the vm_object
no longer includes reference counts.  The major difference is that
vm_object's are no longer freed gratuitiously from the vnode, and so
once an object is created for the vnode, it will last as long as the
vnode does.

When a vnode object reference count is incremented, then the underlying
vnode reference count is incremented also.  The two "objects" are now
more intimately related, and so the interactions are now much less
complex.

When vnodes are now normally placed onto the free queue with an object still
attached.  The rundown of the object happens at vnode rundown time, and
happens with exactly the same filesystem semantics of the original VFS
code.  There is absolutely no need for vnode_pager_uncache and other
travesties like that anymore.

A side-effect of these changes is that SMP locking should be much simpler,
the I/O copyin/copyout optimizations work, NFS should be more ponderable,
and further work on layered filesystems should be less frustrating, because
of the totally coherent management of the vnode objects and vnodes.

Please be careful with your system while running this code, but I would
greatly appreciate feedback as soon a reasonably possible.

26 years agoUse CHECKIO in procfs_ioctl() to ensure that any changes in UID/GID result
sef [Tue, 6 Jan 1998 01:37:12 +0000 (01:37 +0000)]
Use CHECKIO in procfs_ioctl() to ensure that any changes in UID/GID result
in the expected failure.

26 years agoCorrect the return value of ReportProtStatus() ("show proto").
brian [Tue, 6 Jan 1998 00:58:31 +0000 (00:58 +0000)]
Correct the return value of ReportProtStatus() ("show proto").
Submitted by: Dave Bodenstab <imdave@mcs.net>

26 years agoAdd Latin LC_TIME file.
helbig [Tue, 6 Jan 1998 00:54:33 +0000 (00:54 +0000)]
Add Latin LC_TIME file.

26 years ago1. Don't be confused if the first day of a month is dropped due to
helbig [Tue, 6 Jan 1998 00:46:11 +0000 (00:46 +0000)]
1. Don't be confused if the first day of a month is dropped due to
   Gregorian Reformation. E. G. in Albania December 1, 1912 doesn't exist.
2. Add some countries in the switch table (BE, FI, LN, LU)
3. Correct some entries from the new found source of calendar knowledge:
   http://www.pip.dknet.dk/~c-t/calendar.html,
   That is Claus Tondering's Calendar FAQ.