1 |
On Friday 11 of December 2009 09:17:54 Brian Harring wrote: |
2 |
> On Thu, Dec 10, 2009 at 10:27:30PM +0000, Ciaran McCreesh wrote: |
3 |
> > * Don't mess with kdebuild until you're sure that no-one has any |
4 |
> > kdebuild packages installed. |
5 |
> |
6 |
> I'm not looking to start a fight, and frankly after a year or so I've |
7 |
> learned to just subconsciously/automatically ignore kdebuild, but why |
8 |
> exactly must this be in pms? |
9 |
> |
10 |
> It's experimental w/ low user acceptance, and was fundamentally |
11 |
> outside the gentoo mainstream. If long term maintenance of it is |
12 |
> desired for anyone who hasn't yet punted those ebuilds from their |
13 |
> vdb, maintain it in a branch rather then the effective head (don't |
14 |
> make the mainline suffer maintenance for something that was outside |
15 |
> mainline). |
16 |
> |
17 |
> My two cents mind you- at this point, flipping through the source, the |
18 |
> kdebuild bits disrupt the flow from my view (and more importantly the |
19 |
> resultant read of it due to folks trying to structure the text to be |
20 |
> agnostic to non-kdebuild), more importantly doing so w/ minimal gain |
21 |
> for the mainstream. |
22 |
> |
23 |
> Either way, take it as a +1 for punting it from mainstream and making |
24 |
> the kdebuild specific folk maintain their own branch rather then |
25 |
> general eapi (gentoo specific) having to maintain it. Branching of |
26 |
> this sort is presumably one of the reasons pms uses a dvcs after all. |
27 |
> |
28 |
> And to head off one angle of argument, I fully expect if I ever get |
29 |
> ambitious enough to derive an eapi extension that I'll have to |
30 |
> maintain it outside of pms mainline- being nonstandard, I'd expect |
31 |
> nothing less (hence the +1 for punting kdebuild). |
32 |
> |
33 |
> My two cents either way, and well aware it's probably not something |
34 |
> everyone wants to hear. |
35 |
> |
36 |
> Not a hard +1 since I've zero interest in a fight also, but a +1 |
37 |
> either way. |
38 |
> |
39 |
> Sorry, but tis my views. |
40 |
> ~harring |
41 |
|
42 |
PMS document is meant to provide information that can be relied upon. |
43 |
kdebuild-1 was used only during short period of time two years ago and only |
44 |
supported by one package manager. Gentoo KDE team has already expressed no |
45 |
interest in kdebuild-1 long time ago as well[1]. |
46 |
That being said kdebuild-1 can no longer be relied upon as a effective |
47 |
specification - it does not have implementation in official package manager |
48 |
and no new kdebuild-1 is going to be created ever by any Gentoo project - it's |
49 |
effectively dead besides all known . |
50 |
Also, what's the most important - any official packages in kdebuild-1 format |
51 |
has already been replaced (with no loss of functionality, we're talking about |
52 |
KDE 4 ebuilds here) by official packages in EAPI-2 format - format supported |
53 |
by three most popular package managers and users as advised to migrate (if |
54 |
they haven't already) to those packages. |
55 |
Otherwise they're unsupported anyway - and PMS needs to document things that |
56 |
are supported and can be relied upon. |
57 |
|
58 |
There's anything more to add - kdebuild-1 PMS specification should be |
59 |
maintained by those interested in separate git branch, and completely removed |
60 |
from PMS trunk as it only serves as LaTeX code obfuscator. |
61 |
|
62 |
http://www.mail-archive.com/gentoo-dev@l.g.o/msg33146.html |
63 |
|
64 |
On behalf of Gentoo KDE |
65 |
|
66 |
-- |
67 |
regards |
68 |
MM |