The following reply was made to PR misc/35366; it has been noted by GNATS.

From: Hubert Feyrer <hubert@feyrer.de>
To: he@NetBSD.org
Cc: "Dominik Joe Pantu*c(ek" <joe@joe.cz>, gnats-bugs@NetBSD.org
Subject: Re: misc/35366
Date: Sat, 20 Jan 2007 02:35:51 +0100 (CET)

 Havard,
 
 can you please sanity-check the patch I mailed, before I commit it?
 
 
   - Hubert
 
 On Fri, 19 Jan 2007, Dominik Joe Pantu*c(ek wrote:
 
 > Hello,
 >
 >   it works and handles the actual amavisd case correctly. The only case
 > it doesn't handle is the test case I showed you on irc. If the daemon
 > forks itself without adding ':' between interpreter and script. I don't
 > know of any real program that behaves this way, so it might not be
 > neccessary. If this is true, then your patch would be sufficient.
 >
 > Joe
 > Hubert Feyrer napsal(a):
 >>
 >> [CC:ing he@ for rev. 1.66 of rc.subr & PR 31932]
 >>
 >>
 >> Hi,
 >>
 >> following up our chat on IRC tonight, I'm sending you this mail to
 >> document the latest state of things, also in the PR (hopefully...).
 >>
 >> Citing you from IRC:
 >>
 >> <dzoe> But it is the same as the cas 1.66 addresses
 >> <dzoe> But not for interpreter, but for the interpreted program itselft.
 >> <dzoe>  /path/to/interp /path/to/program
 >> <dzoe>  /path/to/interp program
 >> <dzoe>  interp /path/to/program
 >> <dzoe>  interp program
 >> <dzoe> 1.66 handles first 3 cases
 >> <dzoe> Why not the fourth, especially when the 4th is only a variation
 >> on 3rd.
 >> <dzoe> 1.66 catches the "perl: bla" case as long as bla=/path/to/program
 >>
 >> rev. 1.66 was in response to PR 31932, it documents postgrey running as
 >>
 >>    476 ?     IWs   0:41.25 perl: /usr/pkg/sbin/postgrey -i 2525 -d
 >> --pidfile=/var/run/postgrey.pid (perl)
 >>
 >> while for amavis this is something like:
 >>
 >>  16541 ?     S<s    0:06.44 perl: amavisd (master)
 >>
 >> so the difference is to either match a full or short pathname in the
 >> _proccheck variable, after the interpreter and the ": ". The patch
 >> below is also available from
 >> http://www.feyrer.de/Misc/pr-35366-patch4, please test and let me know
 >> if it improves the situation for you.
 >>
 >>
 >>  - Hubert
 >>
 >>
 >> --- /usr/cvs/src-current/etc/rc.subr    2007-01-18 22:24:25.000000000
 >> +0100
 >> +++ /etc/rc.subr    2007-01-19 01:07:44.000000000 +0100
 >> @@ -179,6 +179,7 @@
 >>      _psargs=$3
 >>
 >>      _pref=
 >> +    _procnamebn=${_procname##*/}
 >>      if [ $_interpreter != "." ]; then    # an interpreted script
 >>          read _interp < ${_chroot:-}/$_procname    # read interpreter
 >> name
 >>          _interp=${_interp#\#!}        # strip #!
 >> @@ -190,9 +191,8 @@
 >>          _interpbn=${1##*/}
 >>          _fp_args='_argv'
 >>          _fp_match='case "$_argv" in
 >> -            ${_interp}|"${_interp} "*|"${_interpbn}: ${_procname}"*)'
 >> +            ${_interp}|"${_interp} "*|"${_interpbn}: "*${_procnamebn}*)'
 >>      else                    # a normal daemon
 >> -        _procnamebn=${_procname##*/}
 >>          _fp_args='_arg0 _argv'
 >>          _fp_match='case "$_arg0" in
 >>              $_procname|$_procnamebn|${_procnamebn}:|"(${_procnamebn})")'
 >
 >
 >