Re: --hosts=... semantics ?

On Sun, 18 Mar 2001, Will Partain wrote:

> I am inclined to have *two* options --hosts=... and
> --constraint-hosts=...  --hosts=... would specify the hosts
> where the specified action ('reveal' in the example just
> past...) is to happen; --constraint-hosts=... would be other
> hosts that could be used (only) to satisfy constraints.

That seems reasonable - I bumped into the same problem the other day. But
how should this interact with proxyable methods? eg. if I were to do:

arkpkg compile --hosts=slimy foo--1.2.3

and my compile method is only supposed to run on the build host 'slicker',
what should happen? Clearly it would be incorrect to actually run the
compile on slimy, but should in what circumstances should this cause a
build on slicker to occur? There is no explicit constraint to say that
compile-slimy depends on compile-slicker.

Similarly if I were to do:

arkpkg reveal --hosts=slimy logrotate--3.3

Now to build logrotate, I need to have "popt" revealed on the *build*
host, so I have a constraint such as:

  <constraint><dependency type="normal" name="popt" /></constraint>

on the *compile* method. So while the command does not cause logrotate to
be revealed on any host other than slimy, it is quite possible that other
packages (in this case popt) could be revealed on other hosts.

Sorry if I'm drivelling on about nothing, I'm just slightly confused how
these two mechanisms (constraints and proxyable methods) interact. Feel
free to flame me if I'm talking gibberish ;-)

matt