1 |
Alan McKinnon <alan.mckinnon <at> gmail.com> writes: |
2 |
|
3 |
|
4 |
> > This is my (mis)conception, although, as you have suggest, |
5 |
> > there are (gentoo) cultural norms that do suggest |
6 |
> > certain boolean operations should not be used, |
7 |
> > in say for example, package.keywords? |
8 |
|
9 |
> That's more just a safeguard against forgetting you put it there than anything |
10 |
> else |
11 |
|
12 |
Good to know. |
13 |
|
14 |
> The vast majority of cases will use only the "=" operator or nothing. That's |
15 |
> so you unmask the one version you are interested in, not everything from here |
16 |
> on out, including every buggy, pre-release and just plain broken version that |
17 |
> happens to have an ebuild. |
18 |
|
19 |
So entries in package.keywords should just have the ~ in front of them? |
20 |
No point in using other boolean operations in the package.keywords file? |
21 |
|
22 |
|
23 |
|
24 |
|
25 |
> The use-case for no operator is mostly for the case where you run say a stable |
26 |
> box, and want the latest of a specific well-known package. You might want the |
27 |
> latest Qt for example. Another example is -svn ebuilds - enlightenment is a |
28 |
> case in point. The snapshots are always out of date, latest svn is pretty |
29 |
> stable, so one must unmask everything to get the -9999 versions |
30 |
|
31 |
|
32 |
Ok, now you just tossed my little (pee brain) around quite significantly... |
33 |
Your saying that not operator will get me the -9999 (SVN) version |
34 |
of a package?And that this is most likely the most stable because |
35 |
the devs/hacks work on it often? |
36 |
|
37 |
|
38 |
If so then lets put it to the test. |
39 |
Maybee app-arch/xz-utils ? |
40 |
so my entry in /etc/portage/package.keywords should look like this: |
41 |
|
42 |
app-arch/xz-utils |
43 |
Nothing I tried in either package.keywords or package.unmask |
44 |
make the app-arch/xz-utils-9999 (SVN) version available. |
45 |
|
46 |
|
47 |
So what did I miss? |
48 |
|
49 |
|
50 |
James |