1 |
Ühel kenal päeval, E, 02.07.2018 kell 04:48, kirjutas M. J. Everitt: |
2 |
> A couple of passing thoughts:- |
3 |
> |
4 |
> 1) What is/are your own current use cases for using Gentoo? Server? |
5 |
> Desktop? Cluster? Embedded? on what platforms/arches? And why Gentoo? |
6 |
|
7 |
I basically just have Gentoo everywhere but my phones, usually ~amd64. |
8 |
Including a mips netbook, arm64 board and some arm boards; these are |
9 |
usually for tinkering with gentoo on them, not in daily use. |
10 |
Now also a server running stable amd64. |
11 |
|
12 |
In the past I've done some minimal images work for fitting things in |
13 |
64MB flash (I think it ended up quite comfy below 40MB without trying |
14 |
much, but it was a long time ago). |
15 |
|
16 |
For arches that's then amd64, arm64, arm, mips and x86 in the past. |
17 |
For arm64 I formed arch project some time ago, could use more help... |
18 |
|
19 |
Gentoo offers me familiarity and being in power of things and |
20 |
understanding what is what. |
21 |
|
22 |
> 2) What improvements to the distro (ignoring all structural |
23 |
> discussions) |
24 |
> would you like to see, to enhance your experience of using Gentoo vs |
25 |
> something else? |
26 |
|
27 |
I would like to be able to care less about choices from time to time, |
28 |
when my choice is to let that be done by the maintainers, who know the |
29 |
per-package choice best, so I don't have to research. An example of |
30 |
that is REQUIRED_USE usage in ebuilds that force to make a choice, |
31 |
instead of having a maintainer selected priority when a choice is not |
32 |
made by user. Though it's not actually that much of an issue in |
33 |
practice (this specific example), but I mean in general such things. |
34 |
|
35 |
The package manager could use some speedups and improvements. Getting a |
36 |
@world upgrade list in 25 minutes without some special non-default |
37 |
flags on a dual-core APU with spinning disk is not OK. |
38 |
|
39 |
And I'd like to see more innovation happening within Gentoo. Clear |
40 |
Linux, Fedora and such things shouldn't be taking the cake here; we |
41 |
have potential and often easier to implement (don't have to force it on |
42 |
everyone from binary packages for testing, etc) opportunities too.. |
43 |
|
44 |
|
45 |
Mart |