1 |
On Fri, 2009-04-03 at 09:04 +0200, Markus Duft wrote: |
2 |
> On Thu, 2009-04-02 at 22:22 +0200, Fabian Groffen wrote: |
3 |
> [snip] |
4 |
[snip] |
5 |
> > |
6 |
> > Now your chaining patches. |
7 |
> > |
8 |
> > Apparently you take the total opposite direction now, where you first |
9 |
> > have a Portage in EPREFIX (how you got it is questionable), and within |
10 |
> > this Portage you expect it to take stuff it works with (bash, sed) from |
11 |
> > a different location, let's call it CPREFIX. |
12 |
|
13 |
another thing i'd like to propose. prefix-chaining does not change the |
14 |
behaviour of anything as long as no READONLY_EPREFIX is set in |
15 |
make.conf. i'm taking much care not to break things by accidently taking |
16 |
executables for example from / or from somewhere they might be broken. |
17 |
this is true for all of portage/baselayout/etc. (i think i manage to do |
18 |
it right ;)), but aparently not for the ebuilds. |
19 |
|
20 |
in some places i simply put a type -P instead of EPREFIX/... as you |
21 |
said, this is wrong. so i'd like to do something like the portage |
22 |
function i added yesterday: a function that does this (pseudocode): |
23 |
|
24 |
find_in_eprefix_or_ro_root(path) { |
25 |
if -x $EPREFIX/$path |
26 |
return $EPREFIX/$path |
27 |
|
28 |
for root in readonly_roots { |
29 |
if allows_DEPEND($root) && -x $root/$path |
30 |
return $root/$path |
31 |
} |
32 |
|
33 |
return $EPREFIX/$path |
34 |
} |
35 |
|
36 |
i'd like to add this to prefix.eclass. this does not change the |
37 |
behaviour of an evuild if there is no readonly root set, but makes it |
38 |
work if there is... |
39 |
|
40 |
ideas? |
41 |
|
42 |
Cheers, Markus |
43 |
|
44 |
[snip] |
45 |
> Cheers, Markus |
46 |
> |
47 |
> > |
48 |
> > |
49 |
> |
50 |
> |