[Toybox] toybox initd

[Rob Landley]

On 01/29/2013 12:43:38 AM, Ilya Kuzmich wrote:
> Hi Rob.
> 
> Thank you for your repy.
> 
> > Mind if I forward this response to the toybox mailing list?
> Sure, why not.

Sorry this is taking so long. I started a new job a couple weeks back,  
and it's away from home so there was travel and apartment hunting and  
visiting nearby relatives before that.

Finally catching up a bit, and then I was sick all last night. (I have  
four nieces and nephews. There are four active strains of flu. I expect  
each of them to give me one, but this was probably just something I  
ate.)

> > Finally, I need to study upstart and systemd and see what that's  
> all about.
> > But I haven't even started looking at that yet.
> 
> I didn't digged very deep yet, but it's appears to me, that both of  
> them eat
> dynamic hw events and apply them to the dependency tree, if it brings  
> boot
> closer to the "goal" (i.e. runlevel). Like - "Oh, we got net device.  
> You want
> apache running, right? Let's then start net first and let's see where  
> it will
> bring us". Something like that, plus systemd reimplemented xinetd in  
> own way.
> But there is a great possibility, that i'm misinterpreted things  
> somehow.

I know systemd is a horrible hairball that sucks in a dozen different  
daemons into one big interdependent mess. I'm hoping it dies with  
desktop Linux and Android does something completely different.

I'm currently cutting a bugfix release because cp -Rn checks the  
initial target directory, notices it's there, and doesn't overwrite  
it... or recurse into it. (Oops. This means aboriginal can't collate  
the simple-root-filesystem and native-compiler, the result is just  
simple-root-filesystem with a couple files removed by the prep script.  
And _this_ is why I do native test builds).

While I'm there I have a few other pending things that are nearly done  
anyway: new time command, some id fixups that let me do groups as a  
wrapper around id -Gn which oddly enough is what LSB says to do... I'm  
actually partway through implementing "test" but dunno if that should  
wait until next time.

Then I really need to finish mount and umount, and then I can take a  
look at init.

Thanks,

Rob
 1359658317.0