conf/163508: [rc.subr] [patch] Add " enable" and " disable" commands to rc.subr

[Warner Losh]

On Dec 24, 2011, at 6:15 AM, Chris Rees wrote:

> On 24 December 2011 12:30, Maxim Ignatenko <gelraen.ua at gmail.com> wrote:
>> The following reply was made to PR conf/163508; it has been noted by GNATS.
>> 
>> From: Maxim Ignatenko <gelraen.ua at gmail.com>
>> To: Doug Barton <dougb at freebsd.org>
>> Cc: bug-followup at freebsd.org
>> Subject: Re: conf/163508: [rc.subr] [patch] Add &quot;enable&quot; and
>>  &quot;disable&quot; commands to rc.subr
>> Date: Sat, 24 Dec 2011 14:20:19 +0200
>> 
>>  On 24 December 2011 04:15, Doug Barton <dougb at freebsd.org> wrote:
>>  > This idea has been considered before and rejected because it's too
>>  > difficult to catch all the corner cases, and actually editing a config
>>  > file is not really all that hard of a thing to do.
>>  >
>> 
>>  The idea was to make enabling/disabling services less error-prone. It
>>  don't need to catch _all_ corner cases, because if administrator do
>>  something unusual with startup configuration he should be able to
>>  manipulate it in proper way, or even have tools that do something
>>  similar.
>>  Proposed patch handles /etc/rc.conf, /etc/rc.conf.local and
>>  /etc/rc.conf.d/* properly (I hope), so it should fit nicely in 95% of
>>  cases.
>>  Doing `service someserive enable` is much faster and less error-prone
>>  that `service someservice rcvar ; echo someservicercvar_enable=YES >>
>>  /etc/rc.conf`
> 
> Disagree, sorry.
> 
> If we're going to implement these ideas we should do it properly, not
> for 95% of cases.

A lot depends on what those 5% of the cases are.  Absent an implementation to throw stones at, such criticism is premature.  If the 5% of cases are when someone has done something complicated to the rc.conf file, then I don't care: they won't use this interface and we can detect this case and do nothing.  If the 5% of the cases are when someone has enabled ntpd, then that would be a non-starter.

Warner