maillog: 17/07/2005-23:57:05(-0700): Donnie Berkholz types
> On Sun, 2005-07-17 at 22:07 -0700, Anthony Gorecki wrote:
> > On Sunday, July 17, 2005 9:26 pm, Casey Allen Shobe wrote:
> > > I'm also a bit confused about the portdir_overlay thing - If there
> > > exists a -r15, do I then add a -r16 to make emerge realize an
> > > update is available.  What happens then when an -r16 hits the
> > > regular portage tree?
> > 
> > Nothing at all. If there's a conflict between the standard tree and your 
> > overlay, the ebuild in the overlay takes priority.
> 
> In other words, something very important: You miss all the potentially
> critical changes contained in the new "official" revision.

Since that's a common issue, maybe portage could warn the user sort of
like this:

$ emerge -pv app-foo/bar
...
Warning: app-foo/bar-0.1-r2: [1] overrides a newer ebuild in [0]
[ebuild  R] app-foo/bar-0.1-r2 0kB [1]
Portage overlays:
  [0] /usr/portage
  [1] /usr/portage-local

I know [0] exists in my head only, but that's for illustration only.

"newer" in this case refers to the mtime of the two ebuilds. The user
checks, merges changes (or simply touches the ebuild in their overlay)
and the warning goes away.

However, there is also the possibility to lose all the critical changes
from your overlay if you accidentally install the "official" newer
revision (I bet there are plenty of people who may forget a package that
they have in their overlays). That's already covered by
http://bugs.gentoo.org/show_bug.cgi?id=67072 though.


-- 
*)   Georgi Georgiev   *) Department chairmen never die, they just     *)
(*    [EMAIL PROTECTED]    (* lose their faculties.                        (*
*)  +81(90)2877-8845   *)                                              *)

Attachment: pgpE8IV9GaH4U.pgp
Description: PGP signature

Reply via email to