1 |
Ryan Hill wrote: |
2 |
> Alin Năstac <mrness@g.o> wrote: |
3 |
> |
4 |
>> I suppose what everyone does in their part of the tree is their |
5 |
>> business, but a small subset of packages I maintain have other |
6 |
>> maintainers as well. It is annoying to see rules you assume being |
7 |
>> respected on your ebuilds being broken at every bump made by others. |
8 |
> |
9 |
> I'm sure they're just as annoyed by your bizarre take on patch version |
10 |
> numbering as you are with theirs. ;) |
11 |
|
12 |
Let me try summarizing and dissecting this issue. |
13 |
Please correct and extend where necessary. |
14 |
|
15 |
|
16 |
Bike shedding versus "real issue" |
17 |
|
18 |
- Issue itself might not be earth-shaking |
19 |
|
20 |
- Repeated frustration can become a problem |
21 |
|
22 |
- Frustration with this is present (me included) |
23 |
|
24 |
- Happy developers stay much longer |
25 |
|
26 |
|
27 |
People split into three groups: |
28 |
|
29 |
- Friends of ${P}-fix-issue.patch naming |
30 |
|
31 |
- Friends of ${PN}-fix-issue.patch naming |
32 |
|
33 |
- Friends of ${PN}-1.2.3-fix-issue.patch naming |
34 |
|
35 |
|
36 |
Qualities |
37 |
|
38 |
- ${P} i.e. ${PN}-${PV} |
39 |
- On version bump either .. |
40 |
- Constant on version bump |
41 |
- Several same-content patch files across ebuilds |
42 |
- .. or .. |
43 |
- Change to ${PN}-<pre-bump-version> |
44 |
- Single patch file across ebuilds |
45 |
- Maybe helps reminding most patches should |
46 |
move from downstream to upstream? |
47 |
- 1:1 patch-ebuild relation (see below) |
48 |
|
49 |
- ${PN} |
50 |
- Constant on version bump |
51 |
- Single patch file across ebuilds |
52 |
- Several ebuilds need to be inspected |
53 |
to find out if a patch file is still used |
54 |
- 1:1 patch-ebuild relation (see below) |
55 |
|
56 |
- ${PN}-1.2.3 |
57 |
- Constant on version bump |
58 |
- Single patch file across ebuilds |
59 |
- Several ebuilds need to be inspected |
60 |
to find out if a patch file is still used |
61 |
- <=${PV} values indicate things: either |
62 |
- the patch is downstream only |
63 |
- upstream has not applied it |
64 |
- 1:n patch-ebuild relation |
65 |
(inverse qualities of 1:1 patch-ebuild |
66 |
relation, see below) |
67 |
|
68 |
|
69 |
- 1:1 patch-ebuild relation |
70 |
- Finding out if a patch is still needed |
71 |
is trivial |
72 |
- In-place patch updates possible without |
73 |
affecting other ebuilds |
74 |
- 1:1 benefits only apply if ${P} is used |
75 |
consistently in the whole tree |
76 |
|
77 |
|
78 |
Possible solutions |
79 |
|
80 |
- *Communicating* your likes to all co-maintainers |
81 |
in hope the will respect and remember your agreement |
82 |
|
83 |
- Add a related local comment (*documenting*) to ebuilds |
84 |
and expect other developers to act accordingly on a bump |
85 |
|
86 |
- Making a GLEP *enforcing* on of these and make people |
87 |
vote on which |
88 |
|
89 |
|
90 |
|
91 |
Sebastian |