1 |
On Sun, 24 Mar 2013 08:40:17 -0400 |
2 |
Rich Freeman <rich0@g.o> wrote: |
3 |
|
4 |
> On Sun, Mar 24, 2013 at 8:31 AM, Markos Chandras |
5 |
> <hwoarang@g.o> wrote: |
6 |
> > I don't mind adding that link to every package mask. Do note thought |
7 |
> > that this is not the only way for a package to be rescued (assuming |
8 |
> > it can be rescued). Providing fixes without becoming the maintainer |
9 |
> > is also a viable solution, which is probably something we need to |
10 |
> > add to that page as well. |
11 |
> |
12 |
> I started something at: |
13 |
> http://dev.gentoo.org/~rich0/treeclean.txt |
14 |
|
15 |
It is relief to see that someone is trying to listen and do something |
16 |
constructive about this long standing problem. However, with all due |
17 |
respect, I do not think that the document will help very much. I |
18 |
think most users are aware of the possibilities they have to save a |
19 |
package. They just do not have will, time or priority to do so for a |
20 |
particular package they are using (this fact is essential to accept in |
21 |
order to understand the problem here). |
22 |
|
23 |
What I have expressed in rather theatrical way in my previous mail, is |
24 |
the fact that unresolved bugs contributes to the package removal. |
25 |
This may lead to _not_ reporting a bugs on purpose in order to lower |
26 |
the possibility that the package will be removed. I may say that I am |
27 |
afraid to submit a bug for some packages and for some cases I |
28 |
willingly do not report any for the very same reason. Sad but true. |
29 |
|
30 |
How it can be mitigated? In my opinion by applying 30 days removal |
31 |
policy only to packages that are completely broken. So packages that |
32 |
-- according to current policy -- would have been 30d masked for |
33 |
removal, would be _just masked_ (no 30d removal). This has following |
34 |
advantages: |
35 |
|
36 |
1. Users can submit as many bugs as they want without being afraid |
37 |
that it will contribute to removal of the package. Documented bugs |
38 |
are better than hidden bugs. |
39 |
|
40 |
2. Users can still use the package while being aware that it is |
41 |
partially broken. They can find known bugs in bugzilla. |
42 |
|
43 |
3. Users can still submit new bugs or workarounds to existing bugs. |
44 |
|
45 |
4. Users can submit patches, effectively maintain the package even no |
46 |
official proxy maintainer was established. (If from time to time |
47 |
some dev would bring provided patches to the tree, even better.) |
48 |
|
49 |
5. Since the mask period will be likely longer than 30d there is a |
50 |
bigger chance that someone will take proxy/maintainership, or that |
51 |
someone will submit provided patches to the upstream. Even users |
52 |
or devs that usually do not have will, time or priority to take |
53 |
care of the particular package could find some -- e. g. during |
54 |
summer or so -- and provide a patch. |
55 |
|
56 |
During the mask period Gentoo will basically be providing just the |
57 |
infrastructure. |
58 |
|
59 |
Sorry that I am addressing the policy here even you explicitly said in |
60 |
the document not to do so. I will not make this post longer than it |
61 |
is in trying to explain why I am doing it. I just hope you (and other |
62 |
devs) will try to listen. |
63 |
|
64 |
Robert |
65 |
|
66 |
|
67 |
-- |
68 |
Róbert Èeròanský |
69 |
E-mail: openhs@×××××××××.com |
70 |
Jabber: hs@××××××.sk |