Ralf Corsepius wrote:
It likely is something worth looking into, but based on my experiences
with Fedora on my netbook, I am having doubts compiler optimizations
alone are worth a "secondary arch".
At least on my netbook, neither "speed" or "space" (mine has a disk) are
actual problems.
Is for me! :-) But I don't think -Os is the solution; the problems tend
to be long dep-chains and data (and especially, the intersection of
both*). data tends to be much larger than code. Better would be more
breaking up of huge packages (e.g. KDE) especially where it can reduce
dependencies, or even conflicting package builds to provide different
sets of optional features.
(* As a concrete example, I finally gave up on having kcalc installed,
which allowed me to pull out foomatic. That was quite some space in that
dep-chain. As another example, why does system-config-date require
xulrunner? Something in F10 updates changed there and wanted to pull in
quite some dependencies until I tracked down and nuked s-c-d as the
responsible package.)
--
Matthew
ENOWIT: .sig file for this machine not set up yet
--
fedora-devel-list mailing list
fedora-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/fedora-devel-list
