I agree that cave should be smart enough to do that, I was just trying to think of an effective temporary workaround. As a real solution, could something like this be done:
"$cave foo (-z|-c|-e)" overrides "CAVE_RESOLVE_OPTIONS = (-z|-c|-e)", which overrides "cave foo" automatically passing "-z|-c|-e" to "cave resolve". In other words, could the options specified in CAVE_RESOLVE_OPTIONS be given a higher precedence than the options automatically passed by a "cave foo"? Sort of like "global" options in git have a higher precedence than "system" ones. Regards, Tom -----Original Message----- From: Ciaran McCreesh [mailto:[email protected]] Sent: Tuesday, September 07, 2010 3:09 PM To: Tom Cc: [email protected] Subject: Re: [paludis-user] CAVE_RESOLVE_OPTIONS and fix-linkage On Tue, 7 Sep 2010 15:01:46 -0400 "Tom" <[email protected]> wrote: > To cover all bases, should "--no-complete" and "--no-everything" also > be put at the beginning of CAVE_RESOLVE_OPTIONS? That would guarantee > that only one of "-z", "-c", or "-e" would be passed to "cave > resolve". In this case, CAVE_RESOLVE_OPTIONS would be: > > "--no-lazy --no-complete --no-everything --complete" > > Or, since "--complete overrides the previous --no-complete": > > "--no-lazy --no-everything --complete" To cover all bases, cave should be smart enough to ignore complete / everything / lazy if it's in an _OPTIONS and if something else is specified not in _OPTIONS. Or maybe we shouldn't be paying attention to CAVE_RESOLVE_OPTIONS at all for fix-linkage... -- Ciaran McCreesh _______________________________________________ paludis-user mailing list [email protected] http://lists.pioto.org/mailman/listinfo/paludis-user
