| 1 |
On Thu, 12 Jun 2008 09:52:13 +0200 |
| 2 |
Luca Barbato <lu_zero@g.o> wrote: |
| 3 |
> > You should instead be asking the pkgcore guys why they should be |
| 4 |
> > allowed to continue keeping a package in the tree when they're |
| 5 |
> > blatantly ignoring the EAPI process. |
| 6 |
> |
| 7 |
> The eapi process is something not defined so they cannot do much |
| 8 |
> about it, same for the portage people. |
| 9 |
|
| 10 |
The EAPI process requires that any package manager that claims to |
| 11 |
support a particular EAPI really does. When someone releases a package |
| 12 |
manager that has significant bugs in new EAPI handling, we have to |
| 13 |
decide: |
| 14 |
|
| 15 |
* whether we can use the EAPI in the tree |
| 16 |
* whether we have to avoid the bits of that EAPI that are broken |
| 17 |
* whether we have to release a new EAPI n+1 that's identical to EAPI n, |
| 18 |
and completely ban EAPI n. |
| 19 |
|
| 20 |
Package manager maintainers refusing to do basic testing before |
| 21 |
claiming support for a new EAPI has very messy consequences. If package |
| 22 |
manager maintainers aren't going to do the responsible thing, the whole |
| 23 |
point of EAPIs is lost. |
| 24 |
|
| 25 |
-- |
| 26 |
Ciaran McCreesh |
Archives