1 |
On Sunday 02 February 2014 21:12:30 Jeremy Olexa wrote: |
2 |
> On Sat, Feb 1, 2014 at 8:31 PM, Ruud Koolen <redlizard@g.o> wrote: |
3 |
> > Hi all, |
4 |
> > |
5 |
> > I have recently been doing some cleanup work to the bootstrap script, |
6 |
> > getting |
7 |
> > rid of cruft here and there. My goals in this have been twofold: part of |
8 |
> > the |
9 |
> |
10 |
> I do have to say that I respect your goals here. The essential problem is |
11 |
> that bootstrapping is fragile. We all agree there. However, to introduce a |
12 |
> fundamental change to the process, you will encounter resistance. Now, you |
13 |
> are testing on the "easiest" platform that Gentoo Prefix supports - linux. |
14 |
|
15 |
Very true. I've tried to do more broader testing, but I didn't get far. |
16 |
|
17 |
> Such that there is no automated testing, I'd strongly advise that you start |
18 |
> compiling a list of platforms where the "new" method works. Since you are |
19 |
> the champion of this new idea, it should be your responsibility to ensure |
20 |
> that there is no loss of support. |
21 |
|
22 |
Naturally. As I said, I would certainly not try merging this stuff without the |
23 |
wire-range testing I spoke of. |
24 |
|
25 |
> Now, the hard part is that one person does not have access to many |
26 |
> platforms. I would get ahold of SLES 10&11, which has been historically, the |
27 |
> hardest linux flavor to support. Ping Alan (one of our most active |
28 |
> contributors over the years) to see how FreeMiNT goes. You get some access |
29 |
> to solaris from Fabian (but I think there is some version open sourced?). Of |
30 |
> course there is others too.. |
31 |
> |
32 |
> Does that make sense? |
33 |
|
34 |
Quite. I'll spend some time building a VM fleet for testing purposes, then, |
35 |
and fish around for access to the more restricted platforms. |
36 |
|
37 |
-- Ruud |