1 |
On Monday 12 of March 2012 01:49:35 Brian Harring wrote: |
2 |
> On Sun, Mar 11, 2012 at 04:14:33PM +0100, Ch??-Thanh Christopher Nguy???n |
3 |
wrote: |
4 |
> > Ciaran McCreesh schrieb: |
5 |
> > >> Is there really much of a benefit to this? I guess for anybody who |
6 |
> > >> runs scripts to mass-manipulate ebuilds it might be helpful, but I |
7 |
> > >> think all the package managers planned on supporting all the EAPIs for |
8 |
> > >> quite a while longer. |
9 |
> > > |
10 |
> > > We have to support them indefinitely. It's not possible to uninstall a |
11 |
> > > package whose EAPI is unknown. |
12 |
> > |
13 |
> > Would it be feasible to do a pkg_pretend() check and refuse |
14 |
> > install/upgrade if packages with unsupported EAPI are detected? |
15 |
> |
16 |
> The question should be "is it worth doing it", rather than "can we |
17 |
> hack out something". |
18 |
> |
19 |
> As Ciaran said, PM's are going to be supporting EAPI1 indefinitely |
20 |
|
21 |
In principle, it's actually entirely possible for any PM to start supporting |
22 |
exclusively completely API and ABI-breaking EAPI. |
23 |
|
24 |
There's always the problem with self-upgrading software as it needs to somehow |
25 |
predict (and limit) its development to keep upgrade paths. |
26 |
We have this exact problem where I work (with updating basestation SW) and |
27 |
some solution for this software is to stop being self-upgrading. |
28 |
|
29 |
With external upgrade tool (which would rsync package tree do any |
30 |
vdb/metadata-magic needed) one could replace current PM with latest, API/ABI- |
31 |
incompatible PM version. |
32 |
|
33 |
Now, it may not really make sense for PM as upgrade process of PM itself isn't |
34 |
any different from upgrade process of any other package, still it would allow |
35 |
any API/ABI-breaking ebuild format changes to be introduced if we really need |
36 |
them so badly. |
37 |
|
38 |
-- |
39 |
regards |
40 |
MM |