1 |
Andrew Savchenko <bircoph <at> gentoo.org> writes: |
2 |
|
3 |
> |
4 |
> On Sat, 2 Jul 2016 14:00:29 -0400 Rich Freeman wrote: |
5 |
> > On Sat, Jul 2, 2016 at 1:34 PM, James <wireless <at> tampabay.rr.com> wrote: |
6 |
> > > |
7 |
> > > No wonder the gentoo dev graveyard is so much bigger than those who are |
8 |
> > > still active.... |
9 |
> > |
10 |
> > You're probably conflating effect with cause. It isn't like the |
11 |
> > treecleaners arose and drove off all the devs. (How could they? |
12 |
> > There are only a few of them, and Gentoo policy does operate by |
13 |
> > majority rules.) |
14 |
> |
15 |
> I agree with James, the tree cleaners are on the verge of abusing |
16 |
> their power as well as the security team in PMASK-related issues. |
17 |
> |
18 |
> A lot of packages are being removed just because upstream is AWOL |
19 |
> and package has no maintainer[1]. If package is not |
20 |
> seriously broken, there are _no valid reasons_ to remove it. If |
21 |
> homepage is not availed or was not updated for a few years, but |
22 |
> package still works fine, it should not be removed. "Packages are |
23 |
> still sitting in ~arch" is even less grounded reason: some people |
24 |
> do not use arch at all (including myself). |
25 |
> |
26 |
> Of course such packages have higher probability of being broken in |
27 |
> the future, but as long as they work, they must remain in the tree. |
28 |
> |
29 |
> Same applies for security team, some packages are being masked for |
30 |
> removal due to either minor security issues[2] or issues affecting |
31 |
> very limited number of application use cases[3]. |
32 |
> |
33 |
> I understand that people are probably irritated like "we don't want |
34 |
> more of this crap in the tree", but they may do more harm than good |
35 |
> with such approach. |
36 |
|
37 |
Since Sunrise is closing down, the attic is not being refreshed and there |
38 |
is not current github based system like retreiving old codes from the attic, |
39 |
I think something need to 'arranged' to effect the ability of users to |
40 |
retrieve old codes, and work on them, perhaps overlays. |
41 |
|
42 |
|
43 |
> > It is more like Gentoo's popularity has waned somewhat and we don't |
44 |
> > have as many devs as we used to |
45 |
> |
46 |
> Cant agree with this: we have approximately the same number of devs |
47 |
> for several last years (actually it is increasing a bit from ~230 |
48 |
> to ~240 with VCS write access) and is surely large than <200 about |
49 |
> 5 years ago. And we have more contributors via git-powered |
50 |
> proxy-maintaining now. Of course we need more people engaged and by |
51 |
> all means new developers are welcomed. |
52 |
|
53 |
Yes. Well, I'm still struggling with github. |
54 |
|
55 |
> But the real problem is in package complexity rise. Not only number |
56 |
> of packages is being increased, but they became more and more |
57 |
> complex, sometimes insane. Large packages often reinvent wheels by |
58 |
> creating their own build systems, tools, complex bootstrapping and |
59 |
> even bundling of their own compilers which can't be unbundled in a |
60 |
> sane way. Sometimes I have a feeling that software developers have |
61 |
> a global contest on the most crazy and b0rked build system. |
62 |
|
63 |
Very true. I've ask for a basic github (gentoo centric) wiki doc to |
64 |
help folks get accustomed to github. As an older hack it just a new |
65 |
way of doing things and I'm struggling with those news ways and when to |
66 |
use which one for github. I know there are often many ways to do things, |
67 |
but for folks at my level (strong-user learning the proxy-maint tricks, |
68 |
there should be a few more docs and well defined ways to do things, |
69 |
so the basics are routinely possible by just reading and looking at a few |
70 |
examples. I have no problem digging through the latest version of the |
71 |
dev-manual, or asking for help. |
72 |
|
73 |
|
74 |
> Even Debian has similar issues these days (lack of the manpower to |
75 |
> support over complex stuff)[4]. |
76 |
|
77 |
Agreed. Since my interest is running codes on clusters (that I control), |
78 |
it's just a natural fit to try and help the sys-cluster team by hacking |
79 |
on older, not critical codes as I learn. sys-frabric and rDMA is also |
80 |
a close project to clustering that I am interested in. |
81 |
|
82 |
Also, working on minimal and embedded systems often has a coder |
83 |
use old tools as todays embedded linux systems often resemble systems |
84 |
from years ago and those old CLI tools and C codes are very often usable. |
85 |
|
86 |
I routinely go these purge lists and install the codes I want to keep |
87 |
in /usr/local/portage. I really which there was an attic equivalent |
88 |
in github. That would ease a lot of pain, imho. |
89 |
|
90 |
|
91 |
|
92 |
|
93 |
> Best regards, |
94 |
> Andrew Savchenko |
95 |
|
96 |
hth (and thanks Andrew) |
97 |
James |