| 1 |
On 04/09/10 08:10, Nirbheek Chauhan wrote: |
| 2 |
> Hello! |
| 3 |
> |
| 4 |
> So, I can't find any documentation about this; nor can I find a |
| 5 |
> best-practices list. Can we add broken ebuilds in-tree as long as they |
| 6 |
> are package.masked? automagic deps, wrong deps, missing deps, file |
| 7 |
> collisions, etc etc? Even if it makes the ebuild completely unusable |
| 8 |
> by itself? |
| 9 |
> |
| 10 |
> If yes: |
| 11 |
> |
| 12 |
> So we can add completely broken and useless stuff to tree as long as |
| 13 |
> it's package.masked? |
| 14 |
> |
| 15 |
> If no: |
| 16 |
> |
| 17 |
> What's the minimum amount of "working-ness" that an ebuild must have |
| 18 |
> to be added to tree? Who decides this? The QA team? |
| 19 |
|
| 20 |
Use common sense: if it's work in progress then committing a broken ebuild which |
| 21 |
is p.masked is IMHO acceptable (especially if you need to bump/add more ebuilds |
| 22 |
to get this one working). At the same time if you don't plan on improving it and |
| 23 |
just want to get it committed somewhere - use overlay. |
| 24 |
|
| 25 |
-- |
| 26 |
Krzysztof Pawlik <nelchael at gentoo.org> key id: 0xF6A80E46 |
| 27 |
desktop-misc, java, apache, ppc, vim, kernel, python... |
Archives