1 |
W dniu 26.02.2017, nie o godzinie 21∶16 +0100, użytkownik Lars Wendler |
2 |
napisał: |
3 |
> On Sun, 26 Feb 2017 19:59:19 +0000 Robin H. Johnson wrote: |
4 |
> |
5 |
> > On Sat, Feb 25, 2017 at 03:05:09PM +0100, Ulrich Mueller wrote: |
6 |
> > > As the council has decided in its 2014-10-14 meeting (and confirmed |
7 |
> > > again in the 2016-11-13 meeting), CVS headers should be removed after |
8 |
> > > the migration to Git. |
9 |
> > |
10 |
> > The 2014-10-14 meeting did NOT specify what CVS headers were in |
11 |
> > question, and it was later decided that this was $Header$, not $Id$. |
12 |
> > |
13 |
> > > Until recently, this was blocked by repoman still checking for the |
14 |
> > > $Id$ line. The latter is now fixed in the stable repoman version. |
15 |
> > > |
16 |
> > > Therefore, I am going to remove the remaining CVS headers throughout |
17 |
> > > the tree (except for patches, of course) in two days from now. |
18 |
> > |
19 |
> > This was also discussed in August 2015: |
20 |
> > Subject: 'Infra plans regarding $Id$ - official answer...' |
21 |
> > https://archives.gentoo.org/gentoo-dev/message/d01ce943a9f9404c454c26bdb7efdf0e |
22 |
> > |
23 |
> > $Id$ is used by Git as well, and I was a strong advocate that expansion |
24 |
> > of $Id$ should be ENABLED in the rsync exports, because it allowed |
25 |
> > tracing what version of a file was actually in use. |
26 |
> > |
27 |
> > In the case of Git, $Id$ expands to the blob hash, which can be traced |
28 |
> > to a commit trivially, and several of the council members in the 2015 |
29 |
> > thread did agree it was useful in that format (but I see no formal vote |
30 |
> > was ever taken). |
31 |
> > |
32 |
> |
33 |
> And that's exactly for what I use the $Id$ header. |
34 |
> I am completely against removal of this header line. It does _not_ do |
35 |
> any harm and I don't understand why people want it to be removed so |
36 |
> badly. |
37 |
> Now QA again wants to do a questionable action _without_ any approval |
38 |
> from neither infra nor council. Sorry guys but this is not how things |
39 |
> work. The official answer from infra regarding $Id$ gives enough good |
40 |
> examples why this header line should be kept. |
41 |
> This $Id$ header line is the only way how I can safely keep official |
42 |
> ebuilds and ebuilds from my overlay in sync. I don't like getting my |
43 |
> workflows sabotaged and I consider this a pure act of sabotage... |
44 |
> |
45 |
> How about QA finally starts acting on useful issues or at least do |
46 |
> actions that make sense? |
47 |
|
48 |
How about you give some respect to your fellow developers who simply try |
49 |
to do stuff to improve Gentoo, instead of attributing malice and taking |
50 |
it as personal attack on you? |
51 |
|
52 |
As far as I'm concerned, we could tell as well that the Council decided |
53 |
on header removal, then Infra went rogue and replaced the header with |
54 |
another one, and now it claims that the decision was about $Header$ |
55 |
and not $Id$. Does that sound nice to you? Does it motivate you to work |
56 |
more on Gentoo? |
57 |
|
58 |
Of course, we can dispute that Infra might one day actually start |
59 |
expanding $Id$. And not break random files in the process. And not break |
60 |
Manifest thickening and signing in the process. |
61 |
|
62 |
-- |
63 |
Best regards, |
64 |
Michał Górny |