1 |
On Sunday 20 September 2009 13:28:40 Richard Freeman wrote: |
2 |
> Ryan Hill wrote: |
3 |
> > So, should we always keep a working EAPI 0 version around? If not, when |
4 |
> > can we drop support for old EAPIs? Your opinions please. |
5 |
> |
6 |
> You might want to define what you mean by dropping support for old |
7 |
> EAPIs? Do you mean: |
8 |
> |
9 |
> 1. No longer ensuring that users who have pre-EAPI versions of portage |
10 |
> have a clean upgrade path. |
11 |
> |
12 |
> or |
13 |
> |
14 |
> 2. No longer supporting EAPI=0/1 in package managers. |
15 |
|
16 |
I think he means neither. We should no longer tolerate pre-EAPI2 ebuilds being |
17 |
added to the tree and should work on migrating all "old" ebuilds as the need |
18 |
arises. |
19 |
|
20 |
Having 3+x EAPIs around (depending on how you count you can end up with |
21 |
twelve) makes things confusing for devs and introduces an unneeded source for |
22 |
accidental errors. |
23 |
|
24 |
> If we want to drop support for EAPI=0/1 then EVERY package in portage |
25 |
> needs to be updated to a more recent EAPI. I suspect we're not there |
26 |
> yet (at the very least all those system packages that are deliberately |
27 |
> held back need to change). |
28 |
It doesn't have to be a forced process. Just prevent "old" EAPIs from being |
29 |
actively used, that'll pull up most packages to a newish EAPI soon. The rest |
30 |
is undermaintained anyway, so if anyone feels bored he can migrate them. |
31 |
|
32 |
> I can see why package managers would benefit from fewer cases to |
33 |
> support, however... |
34 |
I think we should care more about developers and the tools they have instead |
35 |
of package manager developers and the things they may or may not do. |
36 |
|
37 |
At the same time I consider the current development of having EAPI 3 mostly |
38 |
finished, features for EAPI 4 discussed and ideas for EAPI 5 being thrown |
39 |
around to be quite stupid. The whole process needs to slow down. Most devs |
40 |
can't list you the differences between 0, 1 and 2 ... how do you expect them |
41 |
to have any benefit from adding more stuff? |
42 |
|
43 |
Just my 3 cents (inflation-corrected), |
44 |
|
45 |
Patrick |