1 |
On Mon, 2018-12-10 at 11:59 -0500, Ian Stakenvicius wrote: |
2 |
> On 2018-12-10 11:07 a.m., Michał Górny wrote: |
3 |
> > What can we do to solve it? |
4 |
> > =========================== |
5 |
> > |
6 |
> > a) Do nothing and hope upstreams solve it at some point. I don't think |
7 |
> > this is acceptable because Portage's slowdown is going to be |
8 |
> > significant. |
9 |
> > |
10 |
> > b) Start dropping py2 from packages. Sadly, this is hard because we're |
11 |
> > talking about huge reverse dependency graph, and I'm pretty sure some of |
12 |
> > our users need those packages w/ py2 support. |
13 |
> > |
14 |
> > c) Slot IPython? That's probably the least intrusive option, though |
15 |
> > ugly as hell, and I'm not sure if it's not going to impact dep |
16 |
> > calculation severely anyway. We'd have a :5 slot that supports py2 |
17 |
> > only, and :0 slot that supports py3 only. Dependencies will be tricky, |
18 |
> > life's going to be hard but maybe things will work. |
19 |
> > |
20 |
> > Any other options? |
21 |
> |
22 |
> This'll be uglier for maintainers but cleaner for the repo -- what |
23 |
> about a single IPython "package" that installs :5 for py2 and :0 for |
24 |
> py3 ? Since python multi-builds anyways it shouldn't be too difficult |
25 |
> to have it install from separate distfiles. Versioning would be a |
26 |
> royal pain, likely some sort of combination of versions into a |
27 |
> snapshot version number of sorts would need to be done... |
28 |
> |
29 |
|
30 |
We can add a virtual. We don't have to do horrible stuff. |
31 |
|
32 |
-- |
33 |
Best regards, |
34 |
Michał Górny |