1 |
On Friday 21 September 2007, Duncan wrote: |
2 |
> Mike Frysinger <vapier@g.o> posted |
3 |
> 200709201234.42152.vapier@g.o, excerpted below, on Thu, 20 Sep |
4 |
> |
5 |
> 2007 12:34:41 -0400: |
6 |
> > we know that someone taking a stage3 has never configured anything |
7 |
> > before and so we can safely put defaults into /root/. |
8 |
> |
9 |
> Just to point out... I've seen people mention overlaying a stage-3 on an |
10 |
> existing installation for recovery reasons, generally broken gcc or (on |
11 |
> amd64) switching back to multilib from 64-bit only profiles, so it |
12 |
> /cannot/ be rightly assumed that there's not an existing configuration |
13 |
> in /root/. (Whether that's the right way to accomplish such recovery |
14 |
> isn't the point; the point is, it's done, by people desperate to get a |
15 |
> working system once again who know no other way to do it.) |
16 |
|
17 |
there's a ton of other files that'd get blown away (like everything |
18 |
in /etc) ... anyone who blindly unpacks a stage3 onto their system gets what |
19 |
they deserve in my eyes |
20 |
|
21 |
> Chris's idea of testing both USE=build *AND* that there's no existing |
22 |
> file there that's going to get blown away, sounds reasonable, regardless |
23 |
> of the debate over where the code is eventually placed. |
24 |
|
25 |
except that doesnt address the "issue" you raised above at all ... the files |
26 |
are going into /root/ ... how they get there is the subject of the debate |
27 |
-mike |