1 |
On Wed, 14 Aug 2013 20:28:02 +0100 |
2 |
Ciaran McCreesh <ciaran.mccreesh@××××××××××.com> wrote: |
3 |
|
4 |
> [.. SNIP ..] |
5 |
|
6 |
Thank you. |
7 |
|
8 |
> > > In order for sets to be added to the tree, we need a spec, we need |
9 |
> > > to decide where sets are allowed (package.mask?), and we need an |
10 |
> > > implementation. |
11 |
> > |
12 |
> > Sets in package.mask sounds unreliable as that would prohibit the |
13 |
> > set from being changed as long as it masked; unless of course, the |
14 |
> > specification would allow for a concept like "mask sets" to exist |
15 |
> > where a particular set is denoted as masked. Otherwise, you would |
16 |
> > have people add / remove things from a normal set and break the mask |
17 |
> > intent. |
18 |
> |
19 |
> Using the conventional view of what a "set" is, |
20 |
|
21 |
But what kind of view would that be, a mathematical set, a set from a |
22 |
prior discussion or a completely different set? I assume the first one. |
23 |
|
24 |
> the point is to allow you to mask, say, kde7 using a single line, and |
25 |
> then define what kde7 is using a set. Then the user can unmask kde7 |
26 |
> without having to copy a big, potentially changing list of packages |
27 |
> out of package.mask. |
28 |
|
29 |
Thank you for clarifying this possibility, that's a good feature; |
30 |
though my wonder here is whether that set will unintentionally change, |
31 |
this could happen if the set is used for other purposes. (emerge @set) |
32 |
|
33 |
See my previous mail for that concern. |
34 |
|
35 |
> Now, if you're viewing a set as being a metapackage rather than a list |
36 |
> of specs, this doesn't apply. But then why have sets at all if they're |
37 |
> just a metapackage? |
38 |
|
39 |
You can also ask the opposite: |
40 |
|
41 |
But why have meta packages if they can be sets? |
42 |
|
43 |
Sets are placed together; which as a result, give you a better overview |
44 |
of the existing sets. How does one for example query all the meta |
45 |
packages that are in the tree? |
46 |
|
47 |
It also keeps you from defining package specific matters that a meta |
48 |
package doesn't really need (EAPI, KEYWORDS, S, ...). |
49 |
|
50 |
On the other hand, meta packages have IUSE; that might not fit in sets. |
51 |
|
52 |
(I'm not choosing for a particular side, just enumerating thoughts.) |
53 |
|
54 |
-- |
55 |
With kind regards, |
56 |
|
57 |
Tom Wijsman (TomWij) |
58 |
Gentoo Developer |
59 |
|
60 |
E-mail address : TomWij@g.o |
61 |
GPG Public Key : 6D34E57D |
62 |
GPG Fingerprint : C165 AF18 AB4C 400B C3D2 ABF0 95B2 1FCD 6D34 E57D |