[prev in list] [next in list] [prev in thread] [next in thread] 

List:       gentoo-dev
Subject:    Re: [gentoo-dev] -* v.s. package.mask
From:       Chris Gianelloni <wolf31o2 () gentoo ! org>
Date:       2005-02-25 13:18:34
Message-ID: 1109337515.30253.48.camel () cgianelloni ! nuvox ! net
[Download RAW message or body]

On Fri, 2005-02-25 at 19:13 +0900, Georgi Georgiev wrote:
> I hope I am not really hijacking the thread, but:
> 
> maillog: 25/02/2005-10:18:12(+0100): Francesco Riosa types
> > Maybe after some time it has been in the tree keyworded "-*" (testing, 
> > testing, testing)?
> 
> What's the policy for "-*" v.s. "package.mask"? '-*' is pretty annoying
> in my opinion. Is "-*" for really, really bad ebuilds or something?

Like many things in Gentoo, there is no hard and fast policy.  The
general rule is that you use -* when you don't want *anyone*
accidentally installing the ebuild.  A good example would be a newer
version of glibc that could potentially do something very nasty if the
user didn't follow the directions.  In this case, you would p.mask it
*and* use -*, to keep people who happen to have "sys-libs/glibc" in
their p.unmask.  This forces the person want to test to manually
intervene to get the package, and ensures that people running the
package *want* to run the package.

You will also see -* in binary ebuilds, as it tells people that this
will *only* work on the platforms in KEYWORDS, like KEYWORDS="-* x86
amd64" would do.

-- 
Chris Gianelloni
Release Engineering - Strategic Lead/QA Manager
Games - Developer
Gentoo Linux

["signature.asc" (application/pgp-signature)]
--
gentoo-dev@gentoo.org mailing list


[prev in list] [next in list] [prev in thread] [next in thread]