Re: CVS commit: src/etc/rc.d

[Tetsuya Isaki <isaki%pastel-flower.jp@localhost>]

At Tue, 12 Jun 2012 16:42:59 +0200,
Joerg Sonnenberger wrote:
> > > > > > Modified Files:
> > > > > >     src/etc/rc.d: makemandb
> > > > > > 
> > > > > > Log Message:
> > > > > > Allows you to disable it if makemandb=NO in rc.conf.
> > > > > > Thanks tsutsui@ and Yasushi Oshima.
> > > > > 
> > > > > Is the manual checkyesno really needed?
> > > > 
> > > > It's too slow and annoying on non modern x86 machines as fccache.
> > > > Isn't it the reason why you added makemandb=YES in 
> > > > src/etc/defaults/rc.conf?
> > > 
> > > It should just run stat on all entries in /usr/share/man etc and be
> > > done. So it shouldn't be that bad even on slow hardware after the first
> > > run.

BTW, though it is an off topic, it takes about half an hour
at every startup on my 68030/30MHz machine.

> > >  The rcvar assignment was missing, but shouldn't the normal logic
> > > skip calling the start command if the rcvar is not YES?
> > 
> > Ah, you mean it's enough to add
> > >> rcvar=${name}
> > line into rc.d/makemandb instead of checkyesno?
> 
> Exactly.

Is it true?
I could not disable it with this patch (diff from rev1.1):
Please show the precise patch which you suggest.

--- makemandb.ORG       2012-03-11 13:27:21.000000000 +0900
+++ makemandb   2012-06-13 10:35:22.000000000 +0900
@@ -10,6 +10,7 @@
 $_rc_subr_loaded . /etc/rc.subr
 
 name="makemandb"
+rcvar=$name
 required_files="/etc/man.conf /usr/sbin/makemandb"
 start_cmd="makemandb_start"
 stop_cmd=":"

---
Tetsuya Isaki <isaki%pastel-flower.jp@localhost / isaki%NetBSD.org@localhost>