]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years ago- Revert previous change of atm storage pools -> uma_zones until a solution
Andrew R. Reiter [Mon, 22 Apr 2002 18:26:05 +0000 (18:26 +0000)]
- Revert previous change of atm storage pools -> uma_zones until a solution
  to atm_free() is written.

22 years agoBring in changes from smbfs-1.4.4.
Sheldon Hearn [Mon, 22 Apr 2002 16:18:36 +0000 (16:18 +0000)]
Bring in changes from smbfs-1.4.4.

22 years agoThis commit was generated by cvs2svn to compensate for changes in r95267,
Sheldon Hearn [Mon, 22 Apr 2002 16:15:20 +0000 (16:15 +0000)]
This commit was generated by cvs2svn to compensate for changes in r95267,
which included commits to RCS files with non-trunk default branches.

22 years agoImport smbfs-1.4.4.
Sheldon Hearn [Mon, 22 Apr 2002 16:15:20 +0000 (16:15 +0000)]
Import smbfs-1.4.4.

22 years agoAdd missing splx calls in bge_tick. These don't make any functional
John Polstra [Mon, 22 Apr 2002 16:15:16 +0000 (16:15 +0000)]
Add missing splx calls in bge_tick.  These don't make any functional
difference in -current, but they are important for -stable where
they are missing too.

MFC after: 1 day

22 years agoBack out remnants of revision 1.97: we don't need TARGET_ARCH here.
Ruslan Ermilov [Mon, 22 Apr 2002 15:53:04 +0000 (15:53 +0000)]
Back out remnants of revision 1.97: we don't need TARGET_ARCH here.

22 years agoUse standard bsd.init.mk prologue.
Ruslan Ermilov [Mon, 22 Apr 2002 15:47:11 +0000 (15:47 +0000)]
Use standard bsd.init.mk prologue.

22 years agoRemove my change to the synopsis field
Tom Rhodes [Mon, 22 Apr 2002 15:18:51 +0000 (15:18 +0000)]
Remove my change to the synopsis field

Requested by: bde
pointy hat to: myself ;)

22 years agoUsage style sweep: spell "usage" with a small 'u'.
Dag-Erling Smørgrav [Mon, 22 Apr 2002 13:44:47 +0000 (13:44 +0000)]
Usage style sweep: spell "usage" with a small 'u'.
Also change one case of blatant __progname abuse (several more remain)
This commit does not touch anything in src/{contrib,crypto,gnu}/.

22 years agoDon't needlessly redefine the afterinstall target.
Ruslan Ermilov [Mon, 22 Apr 2002 10:46:12 +0000 (10:46 +0000)]
Don't needlessly redefine the afterinstall target.

22 years agoProhibit the direct inclusion of bsd.dep.mk.
Ruslan Ermilov [Mon, 22 Apr 2002 10:04:41 +0000 (10:04 +0000)]
Prohibit the direct inclusion of bsd.dep.mk.

22 years agoFixed _SUBDIR.
Ruslan Ermilov [Mon, 22 Apr 2002 09:12:18 +0000 (09:12 +0000)]
Fixed _SUBDIR.

Obtained from: bsd.subdir.mk

22 years agoAlign for one true tab size of 8 in revision 1.2.
Ruslan Ermilov [Mon, 22 Apr 2002 09:06:27 +0000 (09:06 +0000)]
Align for one true tab size of 8 in revision 1.2.

22 years agoMake .asm transformation rules synonyms to the .s rules.
Ruslan Ermilov [Mon, 22 Apr 2002 08:56:34 +0000 (08:56 +0000)]
Make .asm transformation rules synonyms to the .s rules.

22 years agoImplement shared secret mode for hostap mode.
Warner Losh [Mon, 22 Apr 2002 07:09:13 +0000 (07:09 +0000)]
Implement shared secret mode for hostap mode.

Submitted by: Thomas Skibo <skibo@pacbell.net>

22 years agoComment out Kirks io-request priority hack until we can do this in a
Poul-Henning Kamp [Mon, 22 Apr 2002 06:53:20 +0000 (06:53 +0000)]
Comment out Kirks io-request priority hack until we can do this in a
civilized way which doesn't cause grief.

The problem is that it is not generally safe to cast a "struct bio
*" to a "struct buf *".  Things like ccd, vinum, ata-raid and GEOM
constructs bio's which are not entrails of a struct buf.

Also, curthread may or may not have anything to do with the I/O request
at hand.

The correct solution can either be to tag struct bio's with a
priority derived from the requesting threads nice and have disksort
act on this field, this wouldn't address the "silly-seek syndrome"
where two equal processes bang the diskheads from one edge to the
other of the disk repeatedly.

Alternatively, and probably better: a sleep should be introduced
either at the time the I/O is requested or at the time it is completed
where we can be sure to sleep in the right thread.

The sleep also needs to be in constant timeunits, 1/hz can be practicaly
any sub-second size, at high HZ the current code practically doesn't
do anything.

22 years agoAdd ia64_sal_init_state(). This function will initialize the machine
Marcel Moolenaar [Mon, 22 Apr 2002 06:43:31 +0000 (06:43 +0000)]
Add ia64_sal_init_state(). This function will initialize the machine
check handling. In its current form, it only determines the largest
amount of state information it can get from SAL and allocates a region
7 memory block for it.

The next steps involve:
o  get and log any unconsumed (NVM stored) error records across
   reboots,
o  register an OS_MCA handler and enable machine checks.

22 years agoAdd state information types.
Marcel Moolenaar [Mon, 22 Apr 2002 06:35:40 +0000 (06:35 +0000)]
Add state information types.

22 years agoDon't use PAM_SILENT unless hushlogin is set (perforce change 10123)
Dag-Erling Smørgrav [Mon, 22 Apr 2002 06:27:16 +0000 (06:27 +0000)]
Don't use PAM_SILENT unless hushlogin is set (perforce change 10123)

Sponsored by: DARPA, NAI Labs

22 years agoDon't report last login time in PAM case. (perforce change 10057)
Dag-Erling Smørgrav [Mon, 22 Apr 2002 06:26:29 +0000 (06:26 +0000)]
Don't report last login time in PAM case. (perforce change 10057)

Sponsored by: DARPA, NAI Labs

22 years agoFix warnings + wait for child so it doesn't go zombie (perforce change 10122)
Dag-Erling Smørgrav [Mon, 22 Apr 2002 06:25:13 +0000 (06:25 +0000)]
Fix warnings + wait for child so it doesn't go zombie (perforce change 10122)

22 years ago- Change the atm_attributes_pool and atm_stackq_pool to be uma_zone's
Andrew R. Reiter [Mon, 22 Apr 2002 05:11:44 +0000 (05:11 +0000)]
- Change the atm_attributes_pool and atm_stackq_pool to be uma_zone's
  (with the appropriate set_max) rather than using the HARP storage pool
  code.

22 years agoAdd a Makefile for this.
Jake Burkholder [Mon, 22 Apr 2002 04:08:01 +0000 (04:08 +0000)]
Add a Makefile for this.

22 years agoUnderline the default value for rp=, the same way all the other string-type
Garance A Drosehn [Mon, 22 Apr 2002 01:09:24 +0000 (01:09 +0000)]
Underline the default value for rp=, the same way all the other string-type
default values are underlined (instead of using fake double-quotes).

MFC after: 4 days

22 years agoAdd description of `mc' (max copies), add short-form to long-form mapping
Garance A Drosehn [Mon, 22 Apr 2002 01:04:02 +0000 (01:04 +0000)]
Add description of `mc' (max copies), add short-form to long-form mapping
for `tf' (troff filter), and add a cross-reference to chkprintcap in some
lpr-related man pages.

Submitted by: dwmalone
MFC after: 4 days

22 years agoAvoid using pmap_kenter "early", since it may need to dink with vm_page
Jake Burkholder [Sun, 21 Apr 2002 22:57:42 +0000 (22:57 +0000)]
Avoid using pmap_kenter "early", since it may need to dink with vm_page
structures, which may not be setup yet.  Minor cleanups.

22 years agoFix WAW dependency violation on r17 (line 198) that only exists for
Marcel Moolenaar [Sun, 21 Apr 2002 22:43:50 +0000 (22:43 +0000)]
Fix WAW dependency violation on r17 (line 198) that only exists for
the SMP case. While on the subject, remove unnecessary stops. I don't
know if this resolves the memory corruption I'm seeing, but it does
have the potential. We'll see...

22 years agoImplement elf_reloc(). The RT specification says that we can expect
Marcel Moolenaar [Sun, 21 Apr 2002 21:27:57 +0000 (21:27 +0000)]
Implement elf_reloc(). The RT specification says that we can expect
both Elf_Rel and Elf_Rela types of relocation, so handle them both
even though we only have Rel_Rela ATM. We don't handle 32-bit and
big-endian variants yet. Support for that is not trivial enough to
implement it without any evidence that we ever need it in the near
future.

For the FPTR relocations, we currently use the fptr_storage used by
_reloc() is locore.s. This is in no way a real solution, but for now
provides the service we need to get the basics going.

A static recursive function lookup_fdesc() is used to find the address
of a function in a way that keeps track of the load module so that
we can get the correct GP value if we need to construct an OPD (ie
there's no OPD yet for the function.

For simplicity, we create an OPD for the IPLT relocations as well and
simply fill the user provided function descriptor from the OPD. Since
the the official descriptors are unique, this has no bad side effects.
Note that we ignore the addend for FPTR relocations, but use the
addend for IPLT relocations as an offset to the function address.

This commit allows us to load and relocate modules and modules appear
to work correctly, although we probably need to make sure that we set
GP correctly in all cases when we have inter-module calls. This
especially applies to assembly coded functions that have cross module
calls.

22 years agoAdd function link_elf_get_gp(), specific to ia64 for now, to get
Marcel Moolenaar [Sun, 21 Apr 2002 21:08:30 +0000 (21:08 +0000)]
Add function link_elf_get_gp(), specific to ia64 for now, to get
the DT_PLTGOT value. On ia64 this is the value of GP. We need this
to construct function descriptors, but the elf file structure is
not exported to MD code.

Note that the name of the function is based on the meaning that
DT_PLTGOT has on ia64. This may differ on other architectures. As
such, link_elf_get_gp() has a high level of MD to it. Renaming the
function to describe what DT_* value is returned makes it generic,
but also makes the MD code less clear and if we only need this on
ia64, then a general name for a specific function doesn't help.

In short: I don't know what is "right" at this time, so I'll go
with what I have.

22 years agoNon-sendmail users use the FreeBSD sendmail startup functionality to start
Gregory Neil Shapiro [Sun, 21 Apr 2002 20:32:28 +0000 (20:32 +0000)]
Non-sendmail users use the FreeBSD sendmail startup functionality to start
alternative MTAs.  Therefore, always install rc.sendmail, regardless of
NO_SENDMAIL make.conf setting.  Users can still set mta_start_script to a
different script.

This commit is after a repo-copy of src/etc/sendmail/rc.sendmail to
src/etc/rc.sendmail.

Noticed by: Calvin NG <calvinng@brel.com>
MFC after: 3 days

22 years agoUse fseeko and uintptr_t to make sure that we get a sensible offset
David Malone [Sun, 21 Apr 2002 19:04:26 +0000 (19:04 +0000)]
Use fseeko and uintptr_t to make sure that we get a sensible offset
when trying to read from the stack.

PR: 37104
Submitted by: Thomas Quinot <thomas@cuivre.fr.eu.org>
MFC after: 3 weeks

22 years agoMFbed: translation updates
Udo Erdelhoff [Sun, 21 Apr 2002 18:45:07 +0000 (18:45 +0000)]
MFbed: translation updates

22 years agoTurn blank lines into .Pp . This does not significantly change the
Bill Fenner [Sun, 21 Apr 2002 18:41:29 +0000 (18:41 +0000)]
Turn blank lines into .Pp .  This does not significantly change the
rendering of the man pages (turns some sequences of two blank lines
into a single blank line), and eliminates 306 errors generated while
formatting named.conf.5 .

22 years agomdoc police: Use Em ("emphasis") to get italics/underlining, instead of
Bill Fenner [Sun, 21 Apr 2002 18:33:57 +0000 (18:33 +0000)]
mdoc police: Use Em ("emphasis") to get italics/underlining, instead of
 Sm Pa (a syntax error, and an assertion that the arguments are path
 names, which is a little far from true).

22 years agoMFen 1.318-1.327
Udo Erdelhoff [Sun, 21 Apr 2002 18:28:49 +0000 (18:28 +0000)]
MFen 1.318-1.327

22 years agoAdd mount_udf and mount_smbfs to crossrefs.
David Malone [Sun, 21 Apr 2002 18:23:58 +0000 (18:23 +0000)]
Add mount_udf and mount_smbfs to crossrefs.

PR: 37200
Submitted by: Cyrille Lefevre <cyrille.lefevre@laposte.net>

22 years agoStyle. Fix long lines and a <tab> indent that should be 4 spaces.
Mark Murray [Sun, 21 Apr 2002 17:26:16 +0000 (17:26 +0000)]
Style. Fix long lines and a <tab> indent that should be 4 spaces.

22 years agoLimit to the maximum allowed reply size the amount of data that
Ian Dowse [Sun, 21 Apr 2002 16:14:54 +0000 (16:14 +0000)]
Limit to the maximum allowed reply size the amount of data that
nfsrv_readdir and nfsrv_readdirplus can return. A client request
containing an over-large `count' field could trigger the "Bad nfs
svc reply" panic in nfs_syscalls.c.

Spotted while trying to reproduce kern/37304, which turned out to
be fixed in FreeBSD a long time ago.

MFC after: 1 week

22 years agoAdd 'rl' to the list of supported drivers.
Luigi Rizzo [Sun, 21 Apr 2002 15:47:59 +0000 (15:47 +0000)]
Add 'rl' to the list of supported drivers.

22 years agoDon't attempt to decvlare M_DEVFS whern MALLOC_DECLARE is not defined.
Bruce Evans [Sun, 21 Apr 2002 15:47:03 +0000 (15:47 +0000)]
Don't attempt to decvlare M_DEVFS whern MALLOC_DECLARE is not defined.
This fixes warnings that should be errors in fstat.

Reminded by: alpha tinderbox

Fixed some style bugs (ones near BOF and EOF; there are many more).

22 years agoInclude <sys/systm.h> for (at least) the definition of atomic functions
Bruce Evans [Sun, 21 Apr 2002 15:35:54 +0000 (15:35 +0000)]
Include <sys/systm.h> for (at least) the definition of atomic functions
which are sometimes used by the macros in <sys/mutex.h>; don't depend
on not-quite-necessary namespace pollution in <sys/mutex.h>.

22 years agoUse newly-added USB manpage entity references where appropriate.
Bruce A. Mah [Sun, 21 Apr 2002 15:35:52 +0000 (15:35 +0000)]
Use newly-added USB manpage entity references where appropriate.

22 years agoRemove duplicated text.
Giorgos Keramidas [Sun, 21 Apr 2002 14:48:47 +0000 (14:48 +0000)]
Remove duplicated text.

PR: docs/37287
Submitted by: Tony Finch <dot@dotat.at>
MFC after: 3 days

22 years agoMove LOGIN_CAP calls before all file descriptors are closed hard, since some
Andrey A. Chernov [Sun, 21 Apr 2002 13:31:56 +0000 (13:31 +0000)]
Move LOGIN_CAP calls before all file descriptors are closed hard, since some
descriptors may be used by LOGIN_CAP internally, add login_close().

Use "nocheckmail" LOGIN_CAP capability too like our login does.

22 years agoDot terminate a sentence. Rearrange to not use .Nm at the beginning
Philippe Charnier [Sun, 21 Apr 2002 13:02:20 +0000 (13:02 +0000)]
Dot terminate a sentence. Rearrange to not use .Nm at the beginning
of a sentence.

22 years agoUse `The .Nm utility'
Philippe Charnier [Sun, 21 Apr 2002 12:53:08 +0000 (12:53 +0000)]
Use `The .Nm utility'

22 years agoFixed some style bugs ("From:" in vendor id line, disordered MAN line, and
Bruce Evans [Sun, 21 Apr 2002 12:43:14 +0000 (12:43 +0000)]
Fixed some style bugs ("From:" in vendor id line, disordered MAN line, and
blank lines).

Not unapproved of by: markm

22 years agoSetup the child's return values correctly when forking an IA-32 process.
Doug Rabson [Sun, 21 Apr 2002 12:34:58 +0000 (12:34 +0000)]
Setup the child's return values correctly when forking an IA-32 process.

22 years agoUse protected names (_foo) to cutdown on boatloads of lint warnings.
Mark Murray [Sun, 21 Apr 2002 11:16:10 +0000 (11:16 +0000)]
Use protected names (_foo) to cutdown on boatloads of lint warnings.

22 years agoWrap GCC-specific stuff and provide alternative for lint.
Mark Murray [Sun, 21 Apr 2002 11:11:02 +0000 (11:11 +0000)]
Wrap GCC-specific stuff and provide alternative for lint.

22 years agoParenthesise macro arguments to reduce lint warnings.
Mark Murray [Sun, 21 Apr 2002 11:08:52 +0000 (11:08 +0000)]
Parenthesise macro arguments to reduce lint warnings.

22 years agoFix really dumb braino of mine; cast a sizeof() to an int, which it is
Mark Murray [Sun, 21 Apr 2002 11:02:36 +0000 (11:02 +0000)]
Fix really dumb braino of mine; cast a sizeof() to an int, which it is
being compared to, not size_t, which it already is.

22 years agoUsed protected names (_foo) for parameter names. This helps clean up
Mark Murray [Sun, 21 Apr 2002 10:57:43 +0000 (10:57 +0000)]
Used protected names (_foo) for parameter names. This helps clean up
a boatload of lint warnings.

22 years agoStylify (mainly line up macro EOL-continuation \'s), and add a dummy
Mark Murray [Sun, 21 Apr 2002 10:49:00 +0000 (10:49 +0000)]
Stylify (mainly line up macro EOL-continuation \'s), and add a dummy
alternative for lint.

22 years agoParenthesise macro arguments to reduce lint warnings.
Mark Murray [Sun, 21 Apr 2002 10:43:24 +0000 (10:43 +0000)]
Parenthesise macro arguments to reduce lint warnings.

22 years agoStylify, fix tabs.
Mark Murray [Sun, 21 Apr 2002 10:38:35 +0000 (10:38 +0000)]
Stylify, fix tabs.

Liked by: bde (a couple of months back)

22 years agoRemove macros that are defined elsewhere.
Mark Murray [Sun, 21 Apr 2002 10:32:48 +0000 (10:32 +0000)]
Remove macros that are defined elsewhere.

22 years agoImprove self-relocation and fix ABI misinterpretation. The changes
Marcel Moolenaar [Sun, 21 Apr 2002 08:56:17 +0000 (08:56 +0000)]
Improve self-relocation and fix ABI misinterpretation. The changes
here mostly mirror the changes made in
boot/efi/libefi/arch/ia64/start.S rev 1.5

Significant difference: We don't handle the IPLT relocation here.
For barebones KLD support, we make the fptr_storage global.

22 years agoImprove self-relocation:
Marcel Moolenaar [Sun, 21 Apr 2002 08:49:47 +0000 (08:49 +0000)]
Improve self-relocation:
o  We don't expect the PLT relocations to follow the .rela section
   anymore. We still assume that PLT relocations are long formed,
o  Document register usage,
o  Improve ILP,
o  Fix the FPTR relocation by creating unique OPDs per function.
   Comparing functions is valid now,
o  The IPLT relocation naturally handles the addend. Deal with it.
   We ignore the addend for FPTR relocations for now. It's not at
   all clear what it means anyway.

Fix ABI misinterpretation:
o  For Elf_Rela relocations, the addend is explicit and should not
   be loaded from the memory address we're relocating. Only do that
   for Elf_Rel relocations (ie the short form).
o  DIR64LSB is not the same as REL64LSB. DIR64LSB applies to a
   symbol (S+A), whereas REL64LSB applies to the base address (BD+A),

22 years agoConsistently use full pathnames for files, especially executables.
Crist J. Clark [Sun, 21 Apr 2002 08:32:35 +0000 (08:32 +0000)]
Consistently use full pathnames for files, especially executables.

PR: conf/37292
Submitted by: Helge Oldach <send-pr@oldach.net>
MFC after: 3 days

22 years agoRemove my MAINTAINER= from here, as it is in src/MAINTAINER
Mark Murray [Sun, 21 Apr 2002 08:03:25 +0000 (08:03 +0000)]
Remove my MAINTAINER= from here, as it is in src/MAINTAINER

22 years agoTidy up some of the configs; s/gcc/cc/ to make the choice of C compiler
Mark Murray [Sun, 21 Apr 2002 08:01:31 +0000 (08:01 +0000)]
Tidy up some of the configs; s/gcc/cc/ to make the choice of C compiler
more generic. Fix paths for things like scripts (fix perldoc). Tweak
externally passed macros.

22 years agoAdd FBSDID. Start errx(3) messages with lowercase. Remove a ')' with no '('.
Philippe Charnier [Sun, 21 Apr 2002 07:18:16 +0000 (07:18 +0000)]
Add FBSDID. Start errx(3) messages with lowercase. Remove a ')' with no '('.
In usage(), use fprintf(3) instead of errx(3), to make the printed string really
start with `usage: ...'.
Reviewed by:  marcel

22 years agoUn-do note of MFC of inetd change.
Bruce A. Mah [Sun, 21 Apr 2002 06:33:20 +0000 (06:33 +0000)]
Un-do note of MFC of inetd change.

22 years agoNew release note: mergemaster(8) -p/-C.
Bruce A. Mah [Sun, 21 Apr 2002 04:58:24 +0000 (04:58 +0000)]
New release note:  mergemaster(8) -p/-C.

MFC noted:  inetd(8) off by default.

22 years agoo Use our own elf2aout now.
David E. O'Brien [Sun, 21 Apr 2002 02:37:55 +0000 (02:37 +0000)]
o Use our own elf2aout now.
o Generalize a little.

22 years agoDon't load the netgraph line discipline. When it's in use, TIOCMGET
Brian Somers [Sun, 21 Apr 2002 02:00:29 +0000 (02:00 +0000)]
Don't load the netgraph line discipline.  When it's in use, TIOCMGET
fails and on loss of carrier, the device doesn't become selectable with
0 bytes to read.

Problem reported by: ache

22 years ago- Nuke some more not needed #ifdef cruft.
Andrew R. Reiter [Sun, 21 Apr 2002 01:41:05 +0000 (01:41 +0000)]
- Nuke some more not needed #ifdef cruft.

22 years agoFor the P function, the newline must be considered a part of the pattern
Juli Mallett [Sun, 21 Apr 2002 01:27:05 +0000 (01:27 +0000)]
For the P function, the newline must be considered a part of the pattern
space, so when printing it, don't let the pattern space and the output run
into eachother by omitting the \n when printing.

22 years agoAdd sed(1) tests for proper behaviour of the G option as fixed in PR 26152,
Juli Mallett [Sun, 21 Apr 2002 01:04:03 +0000 (01:04 +0000)]
Add sed(1) tests for proper behaviour of the G option as fixed in PR 26152,
and for proper behaviour of some sed functions given a nil pattern space,
as fixed in PR 34813.

The test for G was based on the test in the PR.  The nil pattern space test
is slightly different as we need to get *some* output, as the core dump will
also produce no output (old behaviour) and turn up falsely that the utility
is working fine.

22 years agoCheck to see if make(1)'s handling of doubly-defined targets is busted or
Juli Mallett [Sun, 21 Apr 2002 00:41:37 +0000 (00:41 +0000)]
Check to see if make(1)'s handling of doubly-defined targets is busted or
not.  A lot relies on this.

22 years agoOnly exit at the end of a test, exit with the last value of $? that was not 0.
Juli Mallett [Sun, 21 Apr 2002 00:18:29 +0000 (00:18 +0000)]
Only exit at the end of a test, exit with the last value of $? that was not 0.

This way the entire regression test is looked at as a pass/fail, not just one
sub-test.

22 years agoGCC 3.x WARNS: Add a break to the default case.
Marcel Moolenaar [Sat, 20 Apr 2002 21:56:42 +0000 (21:56 +0000)]
GCC 3.x WARNS: Add a break to the default case.

22 years agoUse %zu to print a size_t, not %u and definitely not %d.
Juli Mallett [Sat, 20 Apr 2002 21:37:26 +0000 (21:37 +0000)]
Use %zu to print a size_t, not %u and definitely not %d.

Remove WARNS?=2, as GCC will usually throw a fit right now, and I'm not going
to mix WARNS and NO_WERROR.

Submitted by: fenner

22 years ago.Nm is not required at the begining of this sentence.
Philippe Charnier [Sat, 20 Apr 2002 21:33:10 +0000 (21:33 +0000)]
.Nm is not required at the begining of this sentence.

22 years agoUse `The .Nm utility'. Add section number to .Xr.
Philippe Charnier [Sat, 20 Apr 2002 21:28:47 +0000 (21:28 +0000)]
Use `The .Nm utility'. Add section number to .Xr.

22 years agoImprove the Bzip2 handling.
David E. O'Brien [Sat, 20 Apr 2002 21:20:58 +0000 (21:20 +0000)]
Improve the Bzip2 handling.

Sponsored by: FreeBSD Mall, Inc.

22 years agoNew release notes: UserConfig gone (release note long overdue!), kget(8)
Bruce A. Mah [Sat, 20 Apr 2002 21:19:57 +0000 (21:19 +0000)]
New release notes:  UserConfig gone (release note long overdue!), kget(8)
gone, UDF support added, inetd(8) not run by default, inbound SMTP not
enabled by default.

22 years agoPrint a size_t as %u not %d.
Juli Mallett [Sat, 20 Apr 2002 21:13:00 +0000 (21:13 +0000)]
Print a size_t as %u not %d.

Pointed out by: mike, des's tinderbox

22 years agoAdd a test of variable evaluation and substitution for make(1), as a start
Juli Mallett [Sat, 20 Apr 2002 20:57:00 +0000 (20:57 +0000)]
Add a test of variable evaluation and substitution for make(1), as a start
of tests for it.

22 years agoResolve conflicts from import of post-8.12.3 bug fixes.
Gregory Neil Shapiro [Sat, 20 Apr 2002 20:34:16 +0000 (20:34 +0000)]
Resolve conflicts from import of post-8.12.3 bug fixes.

22 years agoThis commit was generated by cvs2svn to compensate for changes in r95154,
Gregory Neil Shapiro [Sat, 20 Apr 2002 20:31:51 +0000 (20:31 +0000)]
This commit was generated by cvs2svn to compensate for changes in r95154,
which included commits to RCS files with non-trunk default branches.

22 years agoImport of post-8.12.3 bug fixes from vendor repository.
Gregory Neil Shapiro [Sat, 20 Apr 2002 20:31:51 +0000 (20:31 +0000)]
Import of post-8.12.3 bug fixes from vendor repository.

These are being imported for the upcoming FreeBSD 4.6 release.

22 years agoInclude <arpa/inet.h> for prototype of ntohl() used in the N_BADMAG()
Mike Barcroft [Sat, 20 Apr 2002 20:09:29 +0000 (20:09 +0000)]
Include <arpa/inet.h> for prototype of ntohl() used in the N_BADMAG()
macro.

22 years agoRemove files no longer in vendor release from vendor branch.
Gregory Neil Shapiro [Sat, 20 Apr 2002 19:51:37 +0000 (19:51 +0000)]
Remove files no longer in vendor release from vendor branch.

22 years agoReenable the newly unbroken hfa device.
Mike Barcroft [Sat, 20 Apr 2002 19:44:38 +0000 (19:44 +0000)]
Reenable the newly unbroken hfa device.

22 years agoAdd a test for jot(1).
Juli Mallett [Sat, 20 Apr 2002 19:41:30 +0000 (19:41 +0000)]
Add a test for jot(1).

22 years agoTwo more files to keep track of.
Gregory Neil Shapiro [Sat, 20 Apr 2002 19:34:31 +0000 (19:34 +0000)]
Two more files to keep track of.

22 years agoA simple test for join(1) based on tjr's test case for -o 0.
Juli Mallett [Sat, 20 Apr 2002 19:33:23 +0000 (19:33 +0000)]
A simple test for join(1) based on tjr's test case for -o 0.

Partially based on: PR standards/36072

22 years agoChange two KM_COPY()s to bcopy(). This should have been done when the
Mike Barcroft [Sat, 20 Apr 2002 19:27:34 +0000 (19:27 +0000)]
Change two KM_COPY()s to bcopy().  This should have been done when the
KM_* macros were removed.

22 years agoDespite the fact that it is documented in the handbook, the release notes,
Gregory Neil Shapiro [Sat, 20 Apr 2002 19:04:59 +0000 (19:04 +0000)]
Despite the fact that it is documented in the handbook, the release notes,
and UPDATING and has been posted to both freebsd-current and
freebsd-stable, users are still not adding the required smmsp user and
group before doing an installworld.  Therefore, don't let users do an
installworld unless they have followed directions.

Add a new installcheck Makefile target which installworld runs before
actually starting the installation.  This target can be used by other parts
of userland as well.  The first addition to the target is to check for the
smmsp user and group if NO_SENDMAIL isn't defined.

Others may add checks to this target as they see fit.

MFC after: 1 week

22 years agoAdd an exceedingly simple regression test for the file2c program. One would
Juli Mallett [Sat, 20 Apr 2002 19:01:21 +0000 (19:01 +0000)]
Add an exceedingly simple regression test for the file2c program.  One would
hope that it would never regress, but this is a straightforward test, so add
it.

22 years agoIn my continuing crusade to make life better for non-sendmail users, avoid
Gregory Neil Shapiro [Sat, 20 Apr 2002 19:00:11 +0000 (19:00 +0000)]
In my continuing crusade to make life better for non-sendmail users, avoid
the creation of /var/spool/clientmqueue and therefore the need for the
smmsp user and group if NO_SENDMAIL is defined.  This required breaking out
the creation of the directory into a new BSD.sendmail.dist mtree file.

MFC after: 1 week

22 years agoAdd regression tests for parts of xargs(1).
Juli Mallett [Sat, 20 Apr 2002 18:46:33 +0000 (18:46 +0000)]
Add regression tests for parts of xargs(1).

22 years agoAdd a comment regarding the file header, and the mode that the file is created
Juli Mallett [Sat, 20 Apr 2002 18:44:54 +0000 (18:44 +0000)]
Add a comment regarding the file header, and the mode that the file is created
with.

This should be fixed shortly by adding the (desirable) option to set the file
creation mode on the command line.

22 years agoAdd uuencode(1) and uudecode(1) regression test input and output.
Juli Mallett [Sat, 20 Apr 2002 17:58:13 +0000 (17:58 +0000)]
Add uuencode(1) and uudecode(1) regression test input and output.

22 years agoMFi386 1.222. Remove vm_map_growstack and acquisition and release of Giant
Jake Burkholder [Sat, 20 Apr 2002 17:28:08 +0000 (17:28 +0000)]
MFi386 1.222.  Remove vm_map_growstack and acquisition and release of Giant
from trap_pfault.

22 years agoStart adding framework for userland regression tests, and tests for
Juli Mallett [Sat, 20 Apr 2002 17:20:21 +0000 (17:20 +0000)]
Start adding framework for userland regression tests, and tests for
uuencode and uudecode.

22 years agoImplement several of the c99 updates to scanf(3):
Bill Fenner [Sat, 20 Apr 2002 17:00:56 +0000 (17:00 +0000)]
Implement several of the c99 updates to scanf(3):
 - New length modifiers: hh, j, ll, t, z.

Still to do:
 - %C, %S, %lc, %ls (wide character support)
 - %a/%A (exact hex representation of floating-point numbers)

Removed old compatability equivalents:
 - %D for %ld, %O for %lo, %X for %lx, %E and %F for %le & %lf (these
   were buggy anyway, since they should have represented %Le & %Lf).
 - %[unknown uppercase char] for %ld, %[unknown lowercase char] for %d

22 years agoOops, fix an inverted if test.
Dag-Erling Smørgrav [Sat, 20 Apr 2002 16:52:41 +0000 (16:52 +0000)]
Oops, fix an inverted if test.