1 |
On Fri, Aug 10, 2012 at 4:21 PM, Jeroen Roovers <jer@g.o> wrote: |
2 |
> On Fri, 10 Aug 2012 14:03:23 +0200 |
3 |
> Gilles Dartiguelongue <eva@g.o> wrote: |
4 |
> |
5 |
>> Since you are proposing this, a side question is: |
6 |
>> Why should we write SRC_URI in ebuilds if that info is now available |
7 |
>> in metadata.xml ? (granted that we might still want to keep |
8 |
>> over-riding this information in ebuilds) |
9 |
> |
10 |
> 1) The information in metadata.xml is inaccurate, it's a hint. When it |
11 |
> fails, nothing of value is lost since the ebuild (supposedly) has |
12 |
> what you want. |
13 |
> 2) SRC_URI is precise. |
14 |
> 3) SRC_URI can change over time, and across versions (even with all the |
15 |
> variables in place). |
16 |
> 4) Backward compatibility. |
17 |
> 5) The inversion of your question: Why should we start handling SRC_URI |
18 |
> outside ebuilds and eclasses? Or, how would that be practical, |
19 |
> advantageous, an improvement on the current situation. |
20 |
|
21 |
Right, our proposal is not here to replace SRC_URI, it's here to fix |
22 |
the cases where SRC_URI can't be sanely used to guess new upstream |
23 |
versions (strange mangling rules, unbrowsable directories, etc...). |
24 |
|
25 |
-- |
26 |
Corentin Chary |
27 |
http://xf.iksaif.net |