1 |
On Wed, 09 Jan 2008 17:49:40 +0100 |
2 |
"Wulf C. Krueger" <philantrop@g.o> wrote: |
3 |
> >> What's the proper fix for when keyword requests stagnate in |
4 |
> >> bugzilla? |
5 |
> > That depends upon whether the keyword request is important. |
6 |
> |
7 |
> Let's take a real world example: KDE 3.5.5 is old, buggy and has |
8 |
> some important issues which won't be fixed anymore. |
9 |
|
10 |
Yet it's the most proven version on mips. |
11 |
|
12 |
> > If it is (and legitimately so |
13 |
> |
14 |
> I hope you'll accept it when I say that 3.5.5 is such a legitimate |
15 |
> case now. |
16 |
|
17 |
Why? It was good enough to be keyworded stable at one point. |
18 |
|
19 |
> What would you suggest to do now? I think we've done all we could |
20 |
> short of the following: |
21 |
> |
22 |
> a) Drop all keywords but those of mips. Leaves mips and, more |
23 |
> importantly, its users with a vulnerable and unmaintained set of |
24 |
> packages. |
25 |
|
26 |
...and break the tree spectacularly, causing huge amounts of pain for |
27 |
your fellow developers when they encounter horrible repoman output when |
28 |
they try to do anything. |
29 |
|
30 |
> b) package.mask 3.5.5 with a big, fat warning and let the users |
31 |
> decide. Same drawbacks as a). |
32 |
|
33 |
...and break the tree spectacularly, causing huge amounts of pain for |
34 |
your fellow developers when they encounter horrible repoman output when |
35 |
they try to do anything. |
36 |
|
37 |
> c) Drop 3.5.5 from the tree. The cleanest but most radical solution. |
38 |
> If mips' users want KDE, they would have to bug (sic!) the mips team. |
39 |
|
40 |
...and break the tree spectacularly, causing huge amounts of pain for |
41 |
your fellow developers when they encounter horrible repoman output when |
42 |
they try to do anything. |
43 |
|
44 |
> The solution I favour by far is c). What's your suggestion or did I |
45 |
> miss any other viable solution? Just doing nothing is not an option |
46 |
> here, I'd say, but state your case. |
47 |
|
48 |
3.5.5 was good enough to be keyworded stable at one point. Thus, it |
49 |
can't be *that* bad. |
50 |
|
51 |
-- |
52 |
Ciaran McCreesh |