1 |
On Wed, 2 Dec 2015 13:04:19 +0100 |
2 |
Ulrich Mueller <ulm@g.o> wrote: |
3 |
|
4 |
> >>>>> On Wed, 2 Dec 2015, Alexis Ballier wrote: |
5 |
> |
6 |
> > What's the point, need or advantage in moving this to |
7 |
> > all-ebuild-scope? |
8 |
> |
9 |
> > Usually eclass refactor/api cleanup are done in a -r{n+1} while |
10 |
> > deprecating -rn. This would have the advantage that you can quickly |
11 |
> > post a complete implementation and get wider reviews. |
12 |
> |
13 |
> A proof-of-concept implementation for the two version manipulation |
14 |
> functions is here: |
15 |
> https://482170.bugs.gentoo.org/attachment.cgi?id=418072 |
16 |
> Add some comments and you'll have a working eclass. :) |
17 |
|
18 |
Yes, but I don't see the point of defining bash code in PMS and |
19 |
copy/pasting it in every compliant PM, when said code used to be in |
20 |
only one place. Sounds like a "regression" to me. |
21 |
|
22 |
|
23 |
[...] |
24 |
> Also version_test is missing, but the idea there was to avoid |
25 |
> redundancy and use the implementation that already exists in the |
26 |
> package manager (which does version comparison all the time). This is |
27 |
> one of the reasons for moving it to the package manager. |
28 |
|
29 |
|
30 |
well, then maybe version_test is the only one that makes sense to be |
31 |
added to PM. Though it is unclear to me how to interface it: spawning a |
32 |
python interp. every time the function is called doesn't seem to be a |
33 |
terrible idea wrt performance. |