1 |
On 01 Apr 2022 02:27, Sam James wrote: |
2 |
> > On 29 Mar 2022, at 18:56, Mike Frysinger wrote: |
3 |
> > * release management (e.g. distfiles hosting) |
4 |
> |
5 |
> I'm not sure I love this one under the test of "if our GH got wiped tomorrow, would there |
6 |
> be much impact?" |
7 |
> |
8 |
> If downstream and others are using e.g. pax-utils with an unreliable SRC_URI, |
9 |
> that *is* a pain, and it's not much comfort to then tell them that it "wasn't |
10 |
> covered by infra anyway" or something. |
11 |
|
12 |
we're already in that situation now. dev.g.o is very unreliable. devs clean |
13 |
their space, or they retire, and all the archives they accumulated vanish when |
14 |
infra decides to remove their www space. |
15 |
|
16 |
i'll bet hard cash everytime on GH being more reliable in the short, medium, |
17 |
and long term than dev.g.o. |
18 |
|
19 |
> We do need a proper solution in infra for hosting resources though. I thought |
20 |
> we had a bug for it but I can't find it right this second, bu the idea would be to expand |
21 |
> projects.gentoo.org to more easily host distfiles and stuff independently of |
22 |
> individual developers (whose links go dead when they retire). |
23 |
|
24 |
everyone has been saying this for over a decade. probably even 2 decades at |
25 |
this point. wishing for it doesn't make it happen. i grok that complaining |
26 |
about it also doesn't make it happen :p. but folks who have been infra for a |
27 |
long time have already thought about & discussed it much longer than i have, |
28 |
and we're still where we were 20 years ago -- dev.g.o. |
29 |
|
30 |
it doesn't make sense to me to ban a solution that exists now, and would be |
31 |
trivial to migrate off of if Gentoo infra ever does come up with a solution. |
32 |
especially considering many Gentoo devs are using GH right now for Gentoo |
33 |
projects and the only place you can find their releases are on GH, not even |
34 |
on dev.g.o. |
35 |
|
36 |
> > * CI runs (e.g. GH actions) |
37 |
> |
38 |
> I don't object to this and free CPU is free CPU. I just wouldn't want to |
39 |
> create binary artefacts from it, but I don't think you're proposing that. |
40 |
|
41 |
i would never trust such artifacts in the first place. it's begging for |
42 |
supply chain abuse. |
43 |
|
44 |
> > * Projects for task management |
45 |
> |
46 |
> I struggle with this a bit more because it'd hurt archeology efforts |
47 |
> if GitHub got wiped. |
48 |
|
49 |
i agree, although i note that exact situation occurs now _a lot_ because we |
50 |
accept PRs and a ton of conversation happens there. so if we ever migrate |
51 |
off of GH, any migration process presumably would already have to compensate |
52 |
for moving metadata. |
53 |
|
54 |
> > * possibly even Discussions since it'll provide a clear/scoped space for |
55 |
> > non-Gentoo users & devs. Gentoo forums are huge and require custom accts, |
56 |
> > and mailing lists are huge and a bit restrictive old timey. |
57 |
> |
58 |
> I'm not opposed to this if it's just for user support / queries rather than |
59 |
> Bugs. Making it easier for people to seek help isn't a bad thing. |
60 |
|
61 |
right, bugs belong on bugs.g.o. |
62 |
-mike |