1 |
Well it's not an issue for me anymore as I just exported the absolute path to avoid |
2 |
having problems on my system. But I wanted to report it anyhow in case you wanted to |
3 |
fix this in a more permanent way. |
4 |
|
5 |
I tend to agree I like the idea of leaving in a location independent system in case |
6 |
there is automounting, etc. So, maybe it's possible to just fix the QA check or |
7 |
provide a means of dynamically resolving the absolute path when you do the QA check |
8 |
rather than fixing the ebuild itself ? |
9 |
|
10 |
----- Original Message ----- |
11 |
From: "Fabian Groffen" <grobian@g.o> |
12 |
To: gentoo-alt@l.g.o |
13 |
Sent: Wednesday, November 7, 2007 11:43:02 AM (GMT-0600) America/Chicago |
14 |
Subject: Re: [gentoo-alt] symlinks in EPREFIX path |
15 |
|
16 |
On 07-11-2007 18:41:44 +0100, Fabian Groffen wrote: |
17 |
> > Is this something prefixed-portage should handle, or the ebuild or |
18 |
> > should I solve it manually via the export trick I did ? |
19 |
> |
20 |
> prefixed-portage. |
21 |
> |
22 |
> We run path.normpath() on EPREFIX provided by you, but I guess we should |
23 |
> use the absolute path thinghy that portage internally also uses for |
24 |
> almost everything to avoid this kind of things from happening. |
25 |
> |
26 |
> I don't really like abspath things, because it can possibly screw up |
27 |
> location independence (think of an automounter path being expanded on |
28 |
> e.g. OSX and BSD). |
29 |
> |
30 |
> That said maybe python-fchcksum is easily fixable. |
31 |
|
32 |
I was sort of an optimist. This is python cruft, so of course a hell. |
33 |
|
34 |
|
35 |
-- |
36 |
Fabian Groffen |
37 |
Gentoo on a different level |
38 |
-- |
39 |
gentoo-alt@g.o mailing list |
40 |
|
41 |
|
42 |
-- |
43 |
gentoo-alt@g.o mailing list |