1 |
On Monday 22 December 2008 00:07:01 Graham Murray wrote: |
2 |
> Alan McKinnon <alan.mckinnon@×××××.com> writes: |
3 |
> > postgresql-server is not something new, it's a new way of packaging an |
4 |
> > existing product. It's like the monolithic/split KDE ebuilds, the new |
5 |
> > split postgresql packages make the dev's life easier, and make it |
6 |
> > possible for you to make more modular choices about what you want support |
7 |
> > for. |
8 |
> > |
9 |
> > Plus, the old packages only go as far a version 8.3.1 |
10 |
> > Only the new packages are supported after that version |
11 |
> |
12 |
> The other advantage is that it makes the upgrade process considerably |
13 |
> easier and safer. The new postresql-[base|server|docs] ebuilds are |
14 |
> slotted. |
15 |
|
16 |
What's the rationale behind that? I can see why someone might want two or more |
17 |
versions of php, python, perl or mysql. |
18 |
|
19 |
But postgresql? I can't imagine why it would be useful to the majority to have |
20 |
SLOTs for postgresql. People tend to run one version doing one major job, |
21 |
which is often not the case for the other examples above. |
22 |
|
23 |
-- |
24 |
alan dot mckinnon at gmail dot com |