| 1 |
On 3/14/20 7:20 AM, Samuel Bernardo wrote: |
| 2 |
> Hi again, |
| 3 |
> |
| 4 |
> As a suggestion, I would propose for each overlay to override the eclass |
| 5 |
> functions for those necessary to EAPI 6 within their ebuilds. |
| 6 |
|
| 7 |
Yes, either do that or upgrade EAPI in the ebuilds. |
| 8 |
|
| 9 |
> For instance, go-overlay could use old eclasses that support EAPI 6 |
| 10 |
> overriding the current ones. For some reason they updated their eclasses |
| 11 |
> to EAPI 7 before updating the ebuilds. |
| 12 |
|
| 13 |
Apparently the go-overlay sabotaged itself with this commit: |
| 14 |
|
| 15 |
https://github.com/Dr-Terrible/go-overlay/commit/2b5bfa01bf8777f28f57eb5b7da2ccde4ace744b#diff-2e747d04b1138fffb99dbb6a5fdd6c1b |
| 16 |
|
| 17 |
> This brings me to the main issue that is the portage cache and possible |
| 18 |
> conficts between those eclass defined in the overlay and those from |
| 19 |
> gentoo main stream. Maybe a name convention would be useful here for |
| 20 |
> future upgrades as a code styling guide for gentoo developers. |
| 21 |
|
| 22 |
There's no conflict. Overlays are free to override eclasses as needed. |
| 23 |
|
| 24 |
> What do you think about this kind of issues? |
| 25 |
|
| 26 |
It's not a good idea for overlays to make eclass changes that break |
| 27 |
their own ebuilds. |
| 28 |
|
| 29 |
> Best, |
| 30 |
> |
| 31 |
> Samuel |
| 32 |
> |
| 33 |
> On 12/03/2020 10.42, Samuel Bernardo wrote: |
| 34 |
>> Hi, |
| 35 |
>> |
| 36 |
>> Is there any way to allow EAPI 6 for some overlays with current portage |
| 37 |
>> stable version (2.3.89-r1)? |
| 38 |
|
| 39 |
The portage version is irrelevant. All versions of portage are backward |
| 40 |
compatible with all old EAPIs. |
| 41 |
|
| 42 |
-- |
| 43 |
Thanks, |
| 44 |
Zac |
Archives