1 |
When it'll be possible to start to use it? |
2 |
|
3 |
On Sat, May 20, 2017 at 8:30 PM, Michał Górny <mgorny@g.o> wrote: |
4 |
|
5 |
> Hi, everyone. |
6 |
> |
7 |
> Here's a set of patches inspired by the recent Sphinx dependency |
8 |
> discussion. They make python-r1 (and therefore distutils-r1) capable |
9 |
> of any-of dependency logic similar to the one used in python-any-r1. |
10 |
> |
11 |
> The basic goal is relatively simple -- to improve handling of pure |
12 |
> build-time dependencies in the eclass. It solves two common problems: |
13 |
> |
14 |
> a. dependencies on packages that support only a subset of PYTHON_COMPAT, |
15 |
> |
16 |
> b. dependencies that need to be implementation-bound between themselves |
17 |
> (e.g. Sphinx plugins). |
18 |
> |
19 |
> The new API improves both of those cases significantly. For the former, |
20 |
> we no longer force user to select additional targets via REQUIRED_USE -- |
21 |
> instead, we just any-of dependencies + python_check_deps() to select |
22 |
> implementation independently of whether it is enabled or not. |
23 |
> |
24 |
> For the latter, we no longer have to force all targets of the package |
25 |
> on all the involved dependencies. Again, using any-of dep |
26 |
> and appropriate python_check_deps() we can enforce a single (any) |
27 |
> target throughout all the packages and use it. |
28 |
> |
29 |
> The first three patches do some code refactoring that makes the change |
30 |
> easier and possibly improves maintainability of the code. The next two |
31 |
> patches add support for python_check_deps() and python_gen_any_dep() |
32 |
> respectively. The last two patches provide examples for both use cases |
33 |
> mentioned. |
34 |
> |
35 |
> Please review. |
36 |
> |
37 |
> -- |
38 |
> Best regards, |
39 |
> Michał Górny |
40 |
> |
41 |
> |
42 |
> |