1 |
On Sunday 05 April 2009 10:26:39 Ciaran McCreesh wrote: |
2 |
> On Sat, 4 Apr 2009 23:58:04 -0400 Mike Frysinger wrote: |
3 |
> > get a *working* implementation first and *then* worry about specing |
4 |
> > it. once you have something running with portage, the spec should |
5 |
> > fall naturally out. previous multilib methods attempted to spec |
6 |
> > things out without any real code and they've all just died. |
7 |
> |
8 |
> You don't know whether it's working until you know what it does. |
9 |
> |
10 |
> You appear to be confusing "writing down roughly what you're doing and |
11 |
> figuring out approximately how it might work before you start writing |
12 |
> code" with "writing the spec up-front". |
13 |
|
14 |
wrong. Thomas is talking about adding things directly to EAPI (EAPI-3 in this |
15 |
case). i dont have a problem with scratching crap out in the sand and getting |
16 |
feedback, but forcing it into the EAPI without actually trying it ends with |
17 |
exactly everything ive said. |
18 |
-mike |