changing the install location of hal-system-power-pmu to libexec

[Richard Hughes]

On Fri, 2006-02-17 at 14:54 +0100, Sjoerd Simons wrote:
> On Fri, Feb 17, 2006 at 12:06:34PM +0000, Richard Hughes wrote:
> > On Fri, 2006-02-17 at 11:38 +0000, Matthew Garrett wrote:
> > > On Thu, Feb 16, 2006 at 03:56:32PM -0800, Artem Kachitchkine wrote:
> > > > Why can't we just rename directory from an implementation attribute to
> > > > something that answers the question "What's in this directory?" Because
> > > > the answer "scripts" only sparks one's curiosity: "What kinda scripts?
> > > > Can I put any script in it?" Calling it "method" or something is less
> > > > ambiguous.
> > > 
> > > Just renaming it isn't an option unless it also moves out of /usr/share 
> > > - compiled C code has to go under /usr/lib (or libexec).
> > 
> > I've been playing with re-writing the hal-power-* scripts in python,
> > thus removing the problem (as python can do ioctls). Would this be
> > acceptable as a solution?
> 
> Personally i would rather not have hald depending on python.

A valid concern.

> The callouts
> should be rather trivial programs, so i don't see the value of using python
> instead of C. 

Well, I think they are complicated! Issuing ioctl's and processing user
input (like David's mount scripts) is probably easier done in python
than bash. I've always felt the shell scripts were really hacky, but
then I guess that makes half of Linux hacky!

Richard.