cvs commit: src/include pwd.h src/lib/libc/gen getpwent.c src/usr.sbin/pwd_mkdb pwd_mkdb.c

[Jacques A. Vidrine]

On Mon, Apr 21, 2003 at 01:15:32AM -0700, Doug Barton wrote:
> Heh... well, I suppose what I meant was, for example, compiling a standard
> desktop system from scratch on a system running this new code.

Right, yes.  Three of my own machines have been running this here,
where my own user comes from the Hesiod source.  My desktop could be
called `standard', I suppose: dual-CPU, dual-headed machine running
GNOME 2.  Most everything was freshly rebuilt.

> > It happens that I don't use BIND8, so I missed this issue.  It
[...]
> Not using BIND? Heresy! :) 

Not using _BIND8_ :-)   BIND9 from ports apparently doesn't use its own
getpwent(3) implementations as built.

> Seriously though, what I'm trying to get across
> is that while what named seems to be doing is definitely wonky, it's
> probably not the only code doing it. Binary compat is excellent of course,
> but things that you could compile and run prior to the nss import should
> still compile and run after.

Not necessarily.  I mostly agree - that's why I reverted the _PWF_*
macros.  But, some things may break simply because we now have POSIX.1
thread-safe versions of getpwent(3) and friends, when we did not
before.

Cheers,
-- 
Jacques A. Vidrine <nectar at celabo.org>          http://www.celabo.org/
NTT/Verio SME          .     FreeBSD UNIX     .       Heimdal Kerberos
jvidrine at verio.net     .  nectar at FreeBSD.org  .          nectar at kth.se