[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Init : someone could comment this ?



On Wed, 2008-01-09 at 08:39 -0500, Yaakov Nemoy wrote:
> On Jan 9, 2008 7:14 AM, Ralf Corsepius <rc040203 freenet de> wrote:
> >
> > On Sun, 2008-01-06 at 14:34 -0500, Casey Dahlin wrote:
> > > Ralf Corsepius wrote:
> > > > I could not disagree more - To me any init-script system requiring
> > > > anything outside of what POSIX requires is a mis-conception and flawed
> > > > design.
> > > >
> > > >
> > > Is there a reason for this? It still sounds like a whole lot of "PURGE
> > > THE IMPURE!!!" to me.
> > No, it's "slim down a system to what is inevitably necessary" and don't
> > try to fall into the trap of trying to replace one interpreter with
> > another one, which in reality only means to _add_ another one.
> 
> Having spent enough time with religious people, that sounds a lot like
> the same argument for the same reasons.  Well, many religions are
> different, so YMMV.
This has nothing to do with religion. We are talking about making a
system to compliant to official standards.

> Don't forget about the other advantages to replacing one interpreter
> with another.  Sh and POSIX give you a certain kind of flexibility
> which I think is falling out of vogue in many circles because many
> people don't like embedding several DSL (domain specific languages)
> into their production code.  Switching interpreters lets you pick a
> language that is more suited to the task, or writing one where there
> isn't.
> 
> Also, due to implementation details, it might be easier to optimize
> one interpreter over another.  Knowing very little about the CPython
> implementation, this argument is probably a stretch.

I really don't care which python or other interpreter you want to add to
the inevitable infrastructure and how well accepted and how stable or
not it might be.

OS history is filled with people having tried to do the same and all of
them having gone through quite steep and bumpy learning curves.

> I think it's quite possible to pick a balance between what we need and
> don't need, and the fun part is going to be making up those
> requirements, trimming a system down to meet those requirements, and
> then coding in the startup routines.

I think you are vastly underestimating the amount of resistance and
accusations you can expect from breaking backward compatibility.

You might want to check which amount of resistance SuSE devs had been
facing on their (comparatively moderate) changes to init-scripts they
had introduced several years ago. They already had dynamic sorting of
init-scripts deps many years, then.

Ralf



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]