1 |
On Sun, Mar 27, 2011 at 7:40 PM, Nirbheek Chauhan <nirbheek@g.o> wrote: |
2 |
> On Mon, Mar 28, 2011 at 12:47 AM, Alec Warner <antarus@g.o> wrote: |
3 |
>> On Sun, Mar 27, 2011 at 1:43 PM, Nirbheek Chauhan <nirbheek@g.o> wrote: |
4 |
>>> Just start removing old[1] maintainer-needed packages. If people |
5 |
>>> complain, tell them to start maintaining it. If they continue to |
6 |
>>> complain, ignore them. As tree-cleaner, you have the power to do this |
7 |
>>> and not take bullshit from people about it. |
8 |
>> |
9 |
>> The intent of the TreeCleaner project (years ago) was to essentially |
10 |
>> look for packages in bugzilla that had lots of bugs and no maintainer. |
11 |
>> For a while beandog essentially maintained a site that tracked this |
12 |
>> for us (Gentoo Package that need Lovin' was the awesome title.) |
13 |
>> |
14 |
>> From that list you either fixed the problems and commited them (e.g. |
15 |
>> you were a roving package maintainer) or you pmasked it and marked it |
16 |
>> for the deadpool. |
17 |
>> |
18 |
>> There is not much policy on treecleaning a package just because no one |
19 |
>> has touched it. Time since last touch was just one of a dozen |
20 |
>> indicators used to find packages that are broken (because a package |
21 |
>> not touched since 2006 is also not likely to compile.) |
22 |
>> |
23 |
> |
24 |
> Sure, that's the history. But what made sense back then doesn't make |
25 |
> sense now. Back then we didn't have 600+ packages that no one |
26 |
> maintains, and whose bugs go almost entirely unread. We had crazy |
27 |
> amounts of manpower back then. |
28 |
|
29 |
We probably had more than 600 unmaintained packages because no one was |
30 |
removing dead packages from the tree. I also dispute your manpower |
31 |
logic. Gentoo has been short on developers for years. I don't see |
32 |
how 2011 is any different than 2007 in this aspect. |
33 |
|
34 |
> |
35 |
> As we evolve, the responsibilities of the different parts of Gentoo |
36 |
> also evolve. As such, the tree-cleaners project has evolved, and if |
37 |
> the team isn't allowed to clean the tree, then why do we even have it |
38 |
> anymore? |
39 |
|
40 |
The community got pissed when I deleted unmaintained packages from the |
41 |
tree 'just because it was unmaintained.' Thats why there were a set |
42 |
of criteria for removal. Maybe they changed their mind and you can |
43 |
convince them. Ignoring people's opinions because they are whiners |
44 |
and you are Treecleaners is a thin edge to walk though; so I'd be |
45 |
careful. At least during my tenure there were still hundreds of |
46 |
unmaintained and broken packages; so I didn't much care about |
47 |
unmaintained but working stuff (since there was plenty of work to do.) |
48 |
I would argue the tree is still in a similar state. |
49 |
|
50 |
> |
51 |
> I really don't understand *why* people want to keep around |
52 |
> unmaintained packages. If a package is not maintained, we should come |
53 |
> up and say it outright. Trying to maintain the illusion of maintenance |
54 |
> is really bad — for each person reporting a bug about a package, 100 |
55 |
> people who got that same bug don't report it at all. So what happens |
56 |
> when there are just 50 users for some packages? Half the time we won't |
57 |
> even know that one of them is broken[1]. The rest of the time, users |
58 |
> will get a bad impression of Gentoo saying "Man, half the packages |
59 |
> don't even work". |
60 |
|
61 |
Properly tagged I don't think there is any illusion. |
62 |
Maintainer-needed is maintainer-needed after all. If half of the |
63 |
packages *in the tree* don't work then we have a problem. If half the |
64 |
packages *a user tries to install* are broken then they should |
65 |
certainly use another distro. Perhaps Gentoo is not for their area |
66 |
(and the key point is that it doesn't have to be.) |
67 |
|
68 |
> |
69 |
> It's really simple: |
70 |
> |
71 |
> (a) If the package has plenty of users, there should be no problems |
72 |
> finding a maintainer or a proxy-maintainer. |
73 |
> (b) If the package has few users and is high-maintenance, it's either |
74 |
> already broken, or will get broken soon without a maintainer. Find one |
75 |
> or remove it! |
76 |
> (c) If the package has few users and is low-maintenance, package.mask |
77 |
> it so we can figure out who the users are, and we can get them to |
78 |
> proxy-maintain it, it's so little work anyway, right? |
79 |
> (d) If the package has very few or no users, what the hell is it doing |
80 |
> unmaintained in the tree? It's just eating up disk inodes and space. |
81 |
|
82 |
So launch gstats and get usage numbers. If no one is using a package |
83 |
that is a keen indicator it can be punted; however no one will get off |
84 |
their ass and get more data to back anything up (myself included...) |
85 |
All of your points above assume we have a decent metric of 'how many |
86 |
users a package has' and about the only way we know that is when users |
87 |
file bugs for it (version bump, bug, feature req, etc..) |
88 |
|
89 |
> |
90 |
> We all like to boast about how gentoo has 15,000 packages, but we |
91 |
> neglect to mention that more than 1000 of these are either |
92 |
> unmaintained or very poorly maintained. And this is a very |
93 |
> conservative number. |
94 |
|
95 |
But again this is all made up...m-n was 670-odd packages last I |
96 |
checked. Do we still have m-w these days? |
97 |
|
98 |
> |
99 |
> Let's not turn portage into a graveyard for packages. Let's just remove crap. |
100 |
> |
101 |
> 1. Writer is bad at statistics, this is probably inaccurate. |
102 |
> |
103 |
> -- |
104 |
> ~Nirbheek Chauhan |
105 |
> |
106 |
> Gentoo GNOME+Mozilla Team |
107 |
> |
108 |
> |