1 |
Patrick Lauer posted on Sun, 01 Nov 2015 13:16:31 +0100 as excerpted: |
2 |
|
3 |
> I'm getting mildly very irritated with the lack of easily accessible |
4 |
> ChangeLogs for our packages. |
5 |
> |
6 |
> Apparently updating them stopped some time in August, so now there are |
7 |
> some outdated ChangeLogs that don't really serve any purpose, and the |
8 |
> easiest way for users to figure out why something changed is to yell at |
9 |
> the clumsy gitweb.g.o interface. So instead of grep we now need lots of |
10 |
> patience. |
11 |
> |
12 |
> This does not look reasonable to me. |
13 |
> |
14 |
> Can we please either properly remove ChangeLogs and tell people to not |
15 |
> be curious about changes, or make them useful again? |
16 |
|
17 |
++ |
18 |
|
19 |
I got irritated with that within a week of the git switch, and pretty |
20 |
quickly switched to the github mirror with pregenerated metadata. |
21 |
|
22 |
That gives me the git log, which is perfectly fine by me. =:^) |
23 |
|
24 |
But, that's not at all suitable for the general user. |
25 |
|
26 |
For one thing, while one would hope that github should have no problem |
27 |
with the traffic, there's little question that recommending that as the |
28 |
general gentoo-wide solution runs headlong into the social-contract issue |
29 |
currently being debated on the NFP list. While simply taking commits via |
30 |
github from users who want to use it for that is one thing (and fine by |
31 |
me, I'm not /that/ extreme, as long as the proprietary method doesn't |
32 |
come to be favored), having pretty much the entire userbase syncing from |
33 |
it would be pretty hard *not* to call a "dependency", and thus a |
34 |
violation of the contract. |
35 |
|
36 |
Which means the generally recommended sync method really needs to stay on |
37 |
gentoo infrastructure, or at least with direct gentoo volunteer hosts, |
38 |
tho some may be hosting on proprietary platforms. |
39 |
|
40 |
|
41 |
Personally, I'd love the primary sync method to be git, and the primary |
42 |
change logs conveyed via native git log methods, but in ordered to do |
43 |
that, all those rsync mirrors need to switch to git mirrors. |
44 |
|
45 |
Meanwhile, as pointed out elsewhere in the thread, git syncing simply |
46 |
won't work under some circumstances, as it's all or nothing, and for some |
47 |
people the connection simply isn't stable enough to get all, so it ends |
48 |
up being nothing. |
49 |
|
50 |
But as long as webrsync continues to work, and assuming it can be |
51 |
incrementally fetched/resumed, that should be a sufficient solution for |
52 |
them, so again, rsync isn't really necessary. And for those users, if |
53 |
they do need to check the changelog, the packages.gentoo.org method, |
54 |
following the link to the git log if necessary, should be fine. |
55 |
|
56 |
|
57 |
But clearly, something needs to change. Either the changelogs need to be |
58 |
fixed to be generated again, or the old ones need removed, with people |
59 |
not using git pointed at packages.gentoo.org as mentioned. If the |
60 |
general rsync mirrors are eventually switched to git, with webrsync |
61 |
remaining for those who can't do git, great, but meanwhile, I know if I |
62 |
were still on rsync (or webrsync), I'd be raising hell about the lack of |
63 |
changelogs well before now, and the github mirror I'm actually using |
64 |
clearly isn't suited as general gentoo user solution because it /would/ |
65 |
then be a gentoo dependency and thus a violation of the social contract, |
66 |
so I'm definitely with Patrick on this one. |
67 |
|
68 |
-- |
69 |
Duncan - List replies preferred. No HTML msgs. |
70 |
"Every nonfree program has a lord, a master -- |
71 |
and if you use the program, he is your master." Richard Stallman |