Sjoerd Simons wrote:
On Sat, Apr 26, 2008 at 09:24:40PM +0200, Harald Dunkel wrote:Michael Biebl wrote:We will certainly get a flood of messages from (desktop)users, where suspend is not working anymore (klicking on gnome-power-manager fails, boo). Imho we should optimise for the common case, not the special case.IMHO hibernate is the special case. In a network environment it is pretty optimistic to expect that a suspended Unix machine could
On a networked server, sure. But on a laptop (over 50% of newly sold computers are laptops nowadays), I expect suspend/hibernate to work.
resume only 5 minutes later and find all external services unchanged (on application level). But anyway, pm-utils can be kept optional, as I have shown, so why make it mandatory for a "minority" of server systems that never will be hibernated? Or maybe I'm unreasonable trying to install hal on a server?We could potentially make it a Recommend instead of a Depend. Most package management tools should handle that correctly by now.
If we consider that, then earliest post-lenny, so we can rely on a apt version, which installs recommends by default. We also should first make sure, that d-i adds pm-utils to the laptop-task.
Unless that has happened, we shouldn't change the dependency.
It's not unreasonable to install hal on a server, just rather odd. Hal is mostly used in very (hardwarewise) dynamic environments, which a server usually isn't. Just curious what on your server uses hal?
Wondering as well, especially as you complained about the size of console-tools (which isn't 5MB btw, but 913k for console-tools + 463k for libconsole).
Honestly, I also found the term "dependency hell", mildly exagerated :-O Michael -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth?
signature.asc
Description: OpenPGP digital signature
