| 1 |
Dnia 2014-08-18, o godz. 09:22:46 |
| 2 |
Chris Reffett <creffett@g.o> napisał(a): |
| 3 |
|
| 4 |
> On 8/18/2014 8:56 AM, hasufell wrote: |
| 5 |
> > Almost forgot, of course this does not work if you expect |
| 6 |
> > unpacker_src_unpacker() to run: |
| 7 |
> > inherit unpacker games base |
| 8 |
> > |
| 9 |
> > as well as |
| 10 |
> > inherit unpacker base games |
| 11 |
> > |
| 12 |
> > however |
| 13 |
> > inherit games unpacker base |
| 14 |
> > |
| 15 |
> > will work. |
| 16 |
> > |
| 17 |
> > And now... guess why the games herd made it a policy to always inherit |
| 18 |
> > games.eclass last. Because of the unpredictability of eclasses and that |
| 19 |
> > they may randomly add exported phase functions. It's a bit paranoid, but |
| 20 |
> > understandable, since we don't have any real rules here. |
| 21 |
> > |
| 22 |
> > So in the end 3 eclasses all tell you "inherit me last! really!". Good |
| 23 |
> > luck with figuring out how to make a gnome game with python and multilib |
| 24 |
> > support work together. I can predict the days such a review would take |
| 25 |
> > in #gentoo-sunrise. Not less than 3. |
| 26 |
> > |
| 27 |
> Would it be feasible to add a repoman check for situations like this, |
| 28 |
> where the behavior of a phase is dependent on inherit order? If so, it |
| 29 |
> seems reasonable to me to require explicit calls to eclass functions in |
| 30 |
> these cases to make it clear what's being called when. |
| 31 |
|
| 32 |
Right now, we have no kind of repoman for eclasses. If you have time to |
| 33 |
work on such a thing, please do. Otherwise, all we can do is put more |
| 34 |
checks in ebuilds but that triggers the warning for the wrong people... |
| 35 |
|
| 36 |
-- |
| 37 |
Best regards, |
| 38 |
Michał Górny |
Archives