Switch device database to cdb(5). Rework ttyname(3) and ttyname_r(3) to
depend on new devname_r(3) as heart. Add /dev/pts magic directly to
devname(3). While it can lead to returning non-existing pat

Switch device database to cdb(5). Rework ttyname(3) and ttyname_r(3) to
depend on new devname_r(3) as heart. Add /dev/pts magic directly to
devname(3). While it can lead to returning non-existing paths, the
behavior is more consistent that way. Drop caching layer in devname(3),
it doesn't buy anything for the common case of having access to the
database. Teach devname(3) proper fallback behavior of scanning /dev.
Create both old-style and new-style database for now in /etc/rc.d/sysdb.

show more ...

static openinfo. dead, not unused usage.

Fix -Wsign-compare issue

fix suseconds_t format.

Remove the \n and tabs from the __COPYRIGHT() strings.
Tweak to use a consistent format.

convert __attribute__s to applicable cdefs.h macros

The latest fts fix that removes double-slashes from paths exposed a bug
in the code. Don't assume that the extra slash is going to be there.

Don't assume that fts(3) will add a slash to the name. From Steve Woodford.
KNF.

Kill __P(), use ANSI function declarations; WARNS=3

Move UCB-licensed code from 4-clause to 3-clause licence.

Patches provided by Joel Baker in PR 22366, verified by myself.

use strlcpy. don't hardcode sizes, use sizeof.

Set explicit database layout. Default is filesystem block size
and hashdb doesn't work for bucket size >= 65536.
Approved by: Jason R. Thorpe <thorpej@netbsd.org>

Removed two debug messages commited by mistake. Sorry about this.

errno is not tested any more if dbopen succeeds (to quote errno(2), "It
should only be examined after an error")
Added an additionnal chdir so that -o works with a file in the current
directory.

Moved the errno=0 in the loop for opening the temp file. The problem was that
successful syscalls never set errno, and hence we could have had an inifinite
loop if the opening failed once and succede

Moved the errno=0 in the loop for opening the temp file. The problem was that
successful syscalls never set errno, and hence we could have had an inifinite
loop if the opening failed once and succeded on second attempt.

show more ...

FIxed a problem with string length (a leading / was appearing in w, who and others)
We were using rename() to create the target file from the temp file. Now the temp file is created in the same direc

FIxed a problem with string length (a leading / was appearing in w, who and others)
We were using rename() to create the target file from the temp file. Now the temp file is created in the same directory of the targer file so that rename() will always work (it needs to have the files in the same filesystem)
Using the -o file, the output file may be on a world writable directory, we hence open the temporary file using O_EXCL, and we loop trying different names until it works.

show more ...

We use rename() to produce the dev.db file, and rename() is not able to move
files across filesystems. Hence we use /var/run for dev.tmp, since dev.db
will be in /var/run.

- Moved the temp file to /tmp instead of /var/tmp
- Added a O_EXCL flag so that we cannot open an existing file for temp file

Added flags to choose the location of the input device directory and the
output device database. Goal is to be consistent with other *_mkdb
utilities such as pwd_mkdb or kvm_mkdb.

Descent into subdirs so that pstat -v prints files on fdescfs symbolically.

- Cosmetic changes like:
* Put space after keyword `switch'.
* Explicitly dereference pointer to function.
- Remove unnecessary cast.

deprecate register, use memset/memmove instead of bzero/bcopy

WARNSify, getopt returns -1 not EOF

RCSid police

Merge with 4.4-Lite, and use libc err(3) and warn(3).