not get a SEGV; Richard Kettlewell <rjk@greenend.org.uk>

pw_error() is prefectly legal but now we have to check for it explicity
(it was handled implicitly before since warn(name) would become warn(NULL)).

this is consistent.

belongs. Closes PR/1554 from bk@rt.fm.

This lives in libutil because it uses pw_lock(3) and friends.  Needs
a man page (soon!).

Crank the shlib major number due to the interface change.

From ianm@kashmir.cit.nepean.uws.edu.au.

NL_TEXTMAX characters.

breaks and getting rid of short lines, making these files easier to work with.

be passed to the command interpreter.

argument to .Dt, and move the BUGS section to the end like most man pages.
This is my 1000th commit since joining the OpenBSD dev team.

onto separate lines in the NAME section.

combines binary search and table lookup.
tested on i386, sparc, ppc, hppa

constant).  These are not security holes but it is worth fixing
them anyway both for robustness and so folks looking for examples
in the tree are not misled into doing something potentially dangerous.
Furthermore, it is a bad idea to assume that pathnames will not
include '%' in them and that error routines don't return strings
with '%' in them (especially in light of the possibility of locales).

Clarify point about always NUL-terminating.

device is opened by default).
Don't open the file w/o adding a /dev/ prefix unless the file has
a '/' in it.  This prevents problems where someone says "disklabel
sd0" with a file called "sd0" in the cwd.
The OPENDEV_DRCT flag has been deprecated as it is the the default
behavior and always has been.
Add checks for >= MAXPATHLEN and return ENAMETOOLONG in that case.

errors specified by the library routine malloc(3)).

- s/MESSAGES/DIAGNOSTICS/, and put message descriptions in a nice list.
- Talk more about how using malloc() in signal handlers is bad.