1 |
On Thu, 29 Mar 2012 11:33:50 +0200, Willie WY Wong wrote about Re: |
2 |
[gentoo-user] Anyone Else "Ping-Ponging" with fltk?: |
3 |
|
4 |
> On Wed, Mar 28, 2012 at 05:43:00PM +0100, Penguin Lover David W Noon |
5 |
> squawked: |
6 |
> > In that case, the source of the breakage is almost certainly |
7 |
> > Portage. |
8 |
> > |
9 |
> > If a slotted package is in the world file without a slot |
10 |
> > specification, Portage should really take that to mean "all |
11 |
> > installed slots are required" rather than "any slot will do" -- or, |
12 |
> > worse still, ignore the world entry and fall back to package |
13 |
> > dependencies. |
14 |
> |
15 |
> I disagree. Portage has always been very clear about this: atoms |
16 |
> without slot or version specification means precisely **any |
17 |
> slot/version will do**. |
18 |
|
19 |
In that case, it is a design flaw in Portage. |
20 |
|
21 |
> The behaviour is entirely consistent between |
22 |
> the command line, ebuilds, the world and set files, as well as other |
23 |
> things in the profile (per package use flag and keyword |
24 |
> specifications). |
25 |
|
26 |
When I set a flag in package.use without a version specification, it |
27 |
applies to *all* versions of that package that support that use flag. |
28 |
I have been doing this for quite some years for several slotted |
29 |
packages, e.g. wxWidgets. |
30 |
|
31 |
When I manually stabilize a package in package.accept_keywords without |
32 |
a version specification, *all* unmasked versions of that package become |
33 |
stable. I do this for only one package: paludis. |
34 |
-- |
35 |
Regards, |
36 |
|
37 |
Dave [RLU #314465] |
38 |
====================================================================== |
39 |
dwnoon@××××××××.com (David W Noon) |
40 |
====================================================================== |