1 |
> Step 2 - Metadata.xml contains only a herd |
2 |
> ------------------------------------------ |
3 |
> 1. Take the herd element, and look up the herd in herds.xml to convert |
4 |
> to an email address. This email address must be a valid bugzilla |
5 |
> account. |
6 |
> 2. This email is treated as an implicit maintainer element after this |
7 |
> point. "<maintainer><email>${HERD_EMAIL}</email></maintainer>" |
8 |
> |
9 |
|
10 |
What about <herd>no-herd</herd>? Does that expand to |
11 |
<maintainer><email>maintainer-needed@g.o</email></maintainer>? |
12 |
Should it be added to herds.xml? |
13 |
|
14 |
I'd be all for the expansion. Consequently all ebuilds with metadata |
15 |
<herd>no-herd</herd> and |
16 |
<maintainer><email>maintainer-needed@g.o</email></maintainer> |
17 |
could be reduced to just <herd>no-herd</herd>... |
18 |
|
19 |
|
20 |
> Step 3 - <maintainer> element |
21 |
> ----------------------------- |
22 |
> 1. Add the maintainer element to an ordered list, in the order they are |
23 |
> present in the file. |
24 |
> 2. If an element appears more than once, the later element overrides |
25 |
> the earlier element. (This provides a route when the herd is assigned, |
26 |
> but does not wish to receive email for a specific package). 3. If a |
27 |
> maintainer element contains the 'ignoreauto=1' attribute AND a |
28 |
> non-empty role element (describing why this maintainer should not be |
29 |
> contacted), delete it from the list. |
30 |
|
31 |
What about <maintainer><email>maintainer-needed@g.o</email></maintainer>? |
32 |
And the case with no maintainer and <herd>no-herd</herd>? |
33 |
Those would be resolved by the <herd>no-herd</herd> expansion proposed |
34 |
above. |
35 |
|
36 |
> |
37 |
> Notes/Changes from before: |
38 |
> ==================================================== |
39 |
> - The herd element is always used. |
40 |
> - The 'contact' attribute is now called 'ignoreauto'. |
41 |
|
42 |
nice! |
43 |
|
44 |
> - The 'ignoreauto' is the logical opposite of the original 'contact' |
45 |
> attribute. |
46 |
> - The 'ignoreauto' attribute defaults to false. |
47 |
> - Any usage of the 'ignoreauto' attribute requires the role element to |
48 |
> be present. |
49 |
> - Herds that do not wish to be contacted for specific bugs should add a |
50 |
> maintainer element stating that (and use 'ignoreauto' on the |
51 |
> element). |
52 |
|
53 |
IMHO at least one herd should always be (at least) CC'ed (unless |
54 |
<herd>no-herd</herd> is the only herd) - in order to guarantee that a |
55 |
group is being notified about the problem. |
56 |
Or put differently: if an ebuild is part of a herd, but the herd doesn't |
57 |
want to know about it, why is ebuild part of the herd in the first place? |
58 |
|
59 |
> - Category level metadata.xml is now used for fallback if |
60 |
> this is a bug about a new package in an existing category. |
61 |
> - Category level metadata.xml may contain herd and maintainer elements. |
62 |
|
63 |
By allowing <maintainer> elements in category metadata we possibly create |
64 |
another (very weak) group maintainership. Is there a specific reason to |
65 |
allow more than <herd>? |
66 |
|
67 |
Thanks. |
68 |
Kind regards |
69 |
Thilo |