svn commit: r197790 - head/etc

[Doug Barton]

Hiroki Sato wrote:
> Hi Doug, sorry again for the delayed response.  I have been distracted
> by a busy few weeks due to my day job...

No problem, I know the feeling. :)

> Doug Barton <dougb at FreeBSD.org> wrote
>   in <4AD7B8FA.7020703 at FreeBSD.org>:
> 
> do> >  I noticed there was something wrong about ${name}_program but it
> do> >  seems I mistakenly changed it (sorry...).  Then I received a report
> do> >  "it does not work" so I just reverted it.
> do>
> do> Understood. I am sure you realize that it's always Ok to ask for help
> do> here on -rc. The rc.d system is not life-threateningly complex but it
> do> does have a lot of "behind the scenes" interactions that are not
> do> always obvious. I certainly don't hesitate to ask for review on
> do> changes myself and I encourage others to do so (as you have done in
> do> the past).
> do>
> do> FWIW, what I do object to about your changes in r197144 and r197790
> do> are that in the first case you neglected to mention that you were
> do> changing that part of the code, and in the second you neglected to
> do> mention that you were changing it back to what it was before you
> do> changed it. That made debugging this problem more difficult for me
> do> than (I think) it should have been. You also did not mention that you
> do> were removing $command in your changes to route[6]d, which made
> do> debugging Mark's original complaint harder, but only for about 30
> do> seconds or so. :)
> 
>  Yes, indeed.  I should have been more careful and will keep it in
>  mind.  Sorry about that.

These things happen, I'm not really that concerned about it. The whole
"keep in mind how people will read this commit message years from now"
thing is something that I have to constantly remind myself of as well.

> do> I'm sort of ambivalent about whether we need to continue encouraging
> do> people to use command in the script or not. As long as what's in the
> do> script matches what's in /etc/defaults/rc.conf we're not hurting
> do> anything, although we are duplicating effort.
> do>
> do> My preference at this point is to let the change that I just made
> do> settle for a while, mostly to see if it has any negative interactions
> do> with scripts from ports, then MFC it after 8.0-RELEASE along with the
> do> changes you've made to the IPv6 stuff. After that we can start talking
> do> about ripping command= out of the individual rc.d scripts if people
> do> think that's a good idea.
> 
>  Sounds reasonable to me.

Given that no one has reported any problems with this, and that it
restores the code to what was happening prior to yar's command= fixes
I decided to ask for MFC permission but given the impending RC2 the
request is on hold. My feelings won't be hurt if it doesn't get
approved for 8.0-release though, so I'm happy either way.

FYI, I'm currently in the process of cleaning things up locally so
that I'll have time to give more review to your changes over the last
couple months. I'm seeing some weirdness with my wireless interface
that I think we can probably nail down without too much effort. At
this point I'm not actually sure that it is related to your changes in
any case.


hth,

Doug

-- 

	Improve the effectiveness of your Internet presence with
	a domain name makeover!    http://SupersetSolutions.com/