1 |
On Mon, Jul 31, 2006 at 09:11:35PM -0400, Mike Frysinger wrote: |
2 |
> On Monday 31 July 2006 18:48, Brian Harring wrote: |
3 |
> > Should be discussed on -dev, not here imo; |
4 |
> |
5 |
> why ? short term we're gaining functionality: |
6 |
> simply add .* |
7 |
|
8 |
1) portage isn't the only package manager |
9 |
2) it's not a backwards compatible change |
10 |
3) EAPI was added explicitly to address issues stemming from #2, and |
11 |
it's also an excuse to role out at least |
12 |
3a) default src_install |
13 |
3b) final cleanup of DEPEND=${DEPEND-${RDEPEND}} and friends |
14 |
3c) any cleanup reqiurements needed (fex, disallowing direct vdb |
15 |
access and requiring usage of new metadata functionality). |
16 |
3d) potentially splitting out a src_configure (potentially a pipe |
17 |
dream till eclass2 occurs). |
18 |
|
19 |
Granted, that stuff isn't strictly required (although most of it's |
20 |
simple changes), but the fact is that portage will _not_ handle .* |
21 |
properly now means it's an EAPI bump for incompatibility versioning. |
22 |
|
23 |
Since it has to be versioned, might as well throw in a few other |
24 |
things that have been held up for similar reasons. |
25 |
|
26 |
|
27 |
> if we want to go long term and cut #*, then we can chat on gentoo-dev |
28 |
|
29 |
A) doesn't hurt to ask |
30 |
B) doesn't take any real time to ask |
31 |
C) someone might have a suggestion on better way to do the |
32 |
representation. |
33 |
D) folks are at least aware of what's occuring, rather then having 300 |
34 |
devs gradually screw up and find out that .* started working at some |
35 |
point (as was how .* died off iirc). |
36 |
|
37 |
Plus, it's a real bitch when something slides in and gets deployed, |
38 |
along comes someone pointing out some corner case they think needs |
39 |
fixing. Doubt that will occur, but the "yo... this is going to be |
40 |
changing" is required imo. |
41 |
|
42 |
~harring |