1 |
On Sun, 17 May 2009 18:58:58 +0200 |
2 |
Arfrever Frehtes Taifersar Arahesis <arfrever.fta@×××××.com> wrote: |
3 |
> > No good, for two reasons. |
4 |
> > |
5 |
> > First, this is a global scope change |
6 |
> |
7 |
> Why do you think that it is a global scope change? |
8 |
|
9 |
Package managers still need to be able to get the EAPI, even if they |
10 |
don't support newer EAPIs, which means you're restricted to using syntax |
11 |
that bash-3 can parse. Although you can sneak some bash-4 features |
12 |
through bash-3's parser, it gets extremely confusing. |
13 |
|
14 |
> > and we can't make global scope |
15 |
> > changes to EAPIs using current mechanisms. EAPIs have to carry on |
16 |
> > using bash 3 until the EAPI mechanism is changed. |
17 |
> |
18 |
> IMHO ebuilds are allowed to set DEPEND=">=app-shells/bash-4.0" and use |
19 |
> bash-4.0 features anyway, but it would be easier to just set |
20 |
> appropriate EAPI in ebuilds. |
21 |
|
22 |
Er, no. An ebuild's deps aren't met when the package manager generates |
23 |
metadata from the ebuild. |
24 |
|
25 |
> > Second, by order of the Council, EAPI 3's feature list was locked |
26 |
> > several weeks ago. If we ignore that for one thing, it just means |
27 |
> > everyone else who had features that came along too late will start |
28 |
> > demanding we reconsider those too... |
29 |
> |
30 |
> IMHO addition of this feature would be acceptable. |
31 |
|
32 |
You could say that about any feature, but the Council chose to just go |
33 |
with an absolute cutoff. |
34 |
|
35 |
-- |
36 |
Ciaran McCreesh |