Init Patch for review and evaluation

[Denys Vlasenko]

On Monday 24 December 2007 17:27, Paul Fox wrote:
>  > On Sunday 23 December 2007 01:36, Eugene Bordenkircher wrote:
>  > > On Sun, 2007-12-23 at 00:39 +0000, Denys Vlasenko wrote:
>  > > > Sorry. I don't think that this is worth doing.
>  > > >
>  > > > The patch itself is not bad, the idea of "fixing" init is.
>
>  ...
>
>  > > Your arguments against init and for runit seem fairly sound to me.  To
>  > > be honest, I've never had the opportunity to play with runit.  I will
>  > > try it in our device and see if it fits our needs better.
>  >
>  > If it does not, can I ask you to sacrifice five minutes of your time
>  > and explain what do you need to achieve, and why runit utilities
>  > cannot help.
>
> why this pushback?  i've not read the patch closely, but it's
> certainly a feature i've missed.  i'm sure runit is a fine init
> replacement (we've used minit in the past on some products, and
> i've used daemontools elsewhere), but we also sometimes use regular
> busybox init because it's easy, and often serves our purposes.
> and whether one uses the reconfig-on-HUP feature of init in
> production or not (i probably wouldn't), it's certainly sometimes
> convenenient during product development.
>
> to be clear -- i don't really care if it goes in or not, but
> given all the other special-purpose options and commands we've
> let into busybox over the years (colored ls output?  wget status
> bar? length?), i'd think we'd let in a prominent historic feature of
> traditional init.

I didn't know that traditional init did this.
Does it really do this?
--
vda