| 1 |
Dne čtvrtek 23 Duben 2009 18:13:00 Fabio Erculiani napsal(a): |
| 2 |
> I see a lack of coherence in doing that instead. |
| 3 |
> And also, could you tell me the rationale behind your sentence: "I |
| 4 |
> don't think this is interesting for Gentoo". |
| 5 |
> Because I already have explained what advantages it could bring (with |
| 6 |
> a bare minimum effort!). |
| 7 |
> So, i suppose that the opinions around here are just political or |
| 8 |
> driven by resilience to change. Why not offering more? |
| 9 |
> I just wonder... :) |
| 10 |
Well for example sets support |
| 11 |
now we add l10n ebuild to the set. |
| 12 |
User change linguas recompile l10n since it is in the set. |
| 13 |
With your approach we cant include all the l10n ebuilds to the set, even |
| 14 |
adding them optionaly per useflag is quite annoying against current approach. |
| 15 |
And compile time is this: |
| 16 |
Wed Apr 22 21:27:24 2009 >>> kde-base/kde-l10n-4.2.2 |
| 17 |
merge time: 31 seconds. |
| 18 |
on 1.3 ghz machine for 1 language. So i guess user can wait the 2 minutes and |
| 19 |
it cause not too big slow-down. Disk space needed for compilation all lingua |
| 20 |
packages is smaller than needed for firefox or kdelibs. |
| 21 |
I really agree with one point it is pain in the ass for binary distros to |
| 22 |
maintain. So the solution for it is better to allow packagemanager to handle |
| 23 |
it. |
| 24 |
|
| 25 |
Cheers |
Archives