Subject: Re: CVS commit: basesrc
To: None <lukem@cs.rmit.edu.au>
From: Johnny C. Lam <lamj@stat.cmu.edu>
List: source-changes
Date: 09/06/2000 17:52:10
Luke Mewburn <lukem@cs.rmit.edu.au> writes:
> Jason R Thorpe writes:
> > On Wed, Sep 06, 2000 at 05:03:49PM +1100, Luke Mewburn wrote:
> > 
> >  > I don't want to encourage an `abuse' of `BEFORE' entries; the `BEFORE'
> >  > support was primarily added for third party scripts which may be
> >  > installed into the system where you don't want to have to modify the
> >  > existing scripts to change the dependency ordering.
> > 
> > While that may be true, it seems like a perfect thing to use for this
> > type of situation.  Everything that requires securelevel 0 should run
> > "BEFORE securelevel" is changed...
> 
> I feel that this may end up in with far more complex dependancy rules;
> there was some stuff using BEFORE when the target should have REQUIREd
> it instead. It was starting to get messy, and IIRC, there was the
> potential for making things more difficult to insert the
> afore-mentioned third party scripts (although I can't currently recall
> the scenario I ran into then).
> 
> Whilst BEFORE may stylistically make more sense here, I don't want to
> set the precedent where people will abuse the BEFORE stuff
> unnecessarily, which may then result in the aforementioned difficulty
> for end-users to insert their own scripts into the dependancy tree...

Well, this is quite arguably an "abuse" of REQUIRE, as I expect these
statements to translate to real requirements.

> Flippant comment: it was also easier to modify one script rather than
> hacking all of the scripts that wanted to run before it... :)

Perhaps you should at least insert a comment into the script to note
that it's kind of a hack.

     -- Johnny C. Lam <lamj@stat.cmu.edu>
        Department of Statistics, Carnegie Mellon University
        http://www.stat.cmu.edu/~lamj/