if i install a jvm into a $ROOT where there is none selected, the eclass checks /, not $ROOT, and so my $ROOT remains unconfigured.
Created attachment 312073 [details, diff] java-vm-2.eclass.patch java-config and such needs quite a bit of work, but this fixes the eclass at least
Created attachment 312075 [details, diff] initial java-config-2 ROOT support this at least lets me do things like -L and --set-system-vm with ROOT and as non-root user (since there's no reason to restrict that based on UID) these are the only operations i really care about right now to get a working system. if other stuff comes up, we can add more then :).
(In reply to comment #1) > Created attachment 312073 [details, diff] [details, diff] > java-vm-2.eclass.patch > > java-config and such needs quite a bit of work, but this fixes the eclass at > least Thanks, committed to tree.
(In reply to comment #2) > Created attachment 312075 [details, diff] [details, diff] > initial java-config-2 ROOT support > > this at least lets me do things like -L and --set-system-vm with ROOT and as > non-root user (since there's no reason to restrict that based on UID) Agreed, write access to the config file should be sufficient, Prefix has a similar issue here. > these are the only operations i really care about right now to get a working > system. if other stuff comes up, we can add more then :). There is more :) Fixing this is a one of the goals for java-config-2.2, so I ask for some more patience.
(In reply to comment #4) do we have to wait for an all-or-nothing approach ? this patch fixes a bunch of obvious shortcomings.
(In reply to SpanKY from comment #5) > (In reply to comment #4) > > do we have to wait for an all-or-nothing approach ? this patch fixes a > bunch of obvious shortcomings. Is the patch still valid? spanky/vapier?
ping. vapier?