Gentoo Archives: gentoo-dev

From: Pacho Ramos <pacho@××××××××××××××××××××××××.es>
To: gentoo-dev@l.g.o
Subject: Re: [gentoo-dev] RFC: Project proposal -- maintainer-wanted
Date: Thu, 14 May 2009 06:50:36
Message-Id: 1242284002.6874.9.camel@localhost
In Reply to: [gentoo-dev] RFC: Project proposal -- maintainer-wanted by Mart Raudsepp
1 El jue, 14-05-2009 a las 03:32 +0300, Mart Raudsepp escribió:
2 > Hello,
3 >
4 > I have had this project in my mind for a while, so it's about time to
5 > get it out there, as to see if feedback finds it a good one - and if
6 > that is so, if there are people who want to make it happen.
7 > It is worded as a hypothetical project description for the purpose of
8 > the text perhaps being a draft for the projects official description. So
9 > in the following text instead of terms like "this project would be" I'm
10 > purposely using terms like "this project is", as while writing it, it
11 > got quickly very awkward writing "would be" and such all the time.
12 > Please take it still as a proposal to be judged, commented, improved,
13 > etc etc. And well, do that commenting and improving and volunteering ;)
14 >
15 >
16 > Project maintainer-wanted
17 > =========================
18 >
19 > Abstract:
20 > There are currently quite some package requests (over 3000) languishing
21 > on bugzilla waiting for a developer or team to get interested and
22 > package it in the official gentoo-x86 portage tree. However in quite
23 > some cases that might not happen for quite a while even with very
24 > popular packages desired by users. The purpose of the maintainer-wanted
25 > project is to get as many of such packages to the official tree as
26 > possible as a stopgap solution.
27 >
28 >
29 >
30 > Details/implementation:
31 >
32 > The maintainer-wanted team is actively looking for popular packages in
33 > bugzilla that have been waiting for packaging in the official
34 > "gentoo-x86" tree for a while, and package and maintain as many of them
35 > as the project teams manpower allows without sacrificing quality.
36 >
37 > To decide which libraries/application get packaged in the official tree
38 > by the project, various factors can be considers by the team. These can
39 > include metrics such as:
40 > * bugzilla vote count amongst open maintainer-wanted packages
41 > * recent general useful activity on the packages bugzilla entry
42 > * past and present user community activity in providing ebuild
43 > improvements, version bumps and fixes in the packages bugzilla
44 > maintainer-wanted bug entry
45 > * ease of the packaging thanks to active user contributions or
46 > consistently good upstream packaging making the workload of the
47 > maintainer-team not grow much
48 > * team members interest and personal qualification in the type of the
49 > package and its packaging methods
50 > * ...
51 >
52 >
53 > maintainer-wanted maintained packages are seen as a stopgap solution. As
54 > such it is desired that eventually a team or developer takes over
55 > maintenance to provide the packages dedicated care and free up the
56 > maintainer-wanted team to make available more desired packages that do
57 > not yet have a dedicated maintainer. To achieve that, various methods
58 > are pursued:
59 >
60 > * Other teams and developers are encouraged to take over maintenance.
61 > This includes proxy maintenance when for example a user is found to
62 > consistently and with good quality help out the maintainer-wanted team
63 > in providing fixes, improvements and bumps to an in-tree
64 > maintainer-wanted maintained package. Taking over maintenance is as easy
65 > as for maintainer-needed packages, however a notification to and
66 > acknowledgment from a maintainer-wanted team member is appreciated.
67 >
68 > * Lists of maintainer-wanted packages are generated, sorted by category
69 > of interest. Developers and dedicated package teams are encouraged to
70 > find packages of interest from these lists and take over maintenance.
71 >
72 > * Simply the easier availability (and the resulting exposure) of a
73 > package in the official tree (as opposed to an unreviewed, yet possibly
74 > high quality, ebuild attached to bugzilla, which has thousands of such
75 > entries) could catch the interest of another team or developer and they
76 > are encouraged to take over maintenance when they have the capacity
77 > (manpower/time etc)
78 >
79 >
80 > In other ways the maintainer-wanted team is not significantly different
81 > to other package maintaining teams:
82 > * The project is responsible for their maintained packages. Quality is
83 > not sacrificed; bugs on in-tree packages get acted upon, etc. As such it
84 > is likely desired to have a different alias than the default one for new
85 > packages (or a different good means of differentiating), as to not have
86 > bugs against already in-tree packages get lost amongst the hundreds or
87 > thousands of packages still waiting to get into the official package
88 > tree.
89 > * In-tree maintainer-wanted packages are also tried to be kept up to
90 > date in regards to new stable upstream versions. Users are encouraged to
91 > file bump requests on bugzilla even as 1-day requests due to the
92 > diversity of packages maintained by the project and therefore too many
93 > different places and notification mechanisms to manually check and
94 > monitor for the team (bugzilla version bump requests solve the
95 > diversity); at least until no mechanisms exist for automatic checking of
96 > bumps, which could get implemented in the future.
97 >
98 >
99 > The maintainer-team project hopes to make previously directly
100 > unavailable popular packages of quality easily available to the user
101 > base until other projects and developers are able to take over.
102 >
103 >
104 > ----
105 >
106 > Discuss! :)
107
108 Maybe other option would be to use sunrise project for that:
109 1. Bugs related with new package additions could be automatically
110 assigned to something like sunrise@g.o
111 2. sunrise@g.o would be a group that would try to get these apps
112 in sunrise overlay as soon as possible. Also, as more people would be in
113 that "sunrise" herd, hopefully, updates, version bumps and some bug
114 fixes would be done faster, as the same people could handle more than
115 one app. For example, if I were a sunrise member then, if I had a bit of
116 free time, I could check bug list and try to fix as much bugs I were
117 able to. Also, me, as a member of that herd, would get mails from
118 bugzilla with new bugs related to that apps.
119 3. People, could join to sunrise herd for helping with this task (of
120 course after aproval of current sunrise leaders)
121
122 Regards