1 |
On 18/11/2007, Fabian Groffen <grobian@g.o> wrote: |
2 |
> On 18-11-2007 09:05:46 +1000, Peter Ansell wrote: |
3 |
> > Hi, |
4 |
> > |
5 |
> > I did some experimenting with crossdev in my OS X Leopard prefix and |
6 |
> > found that it was failing looking for /etc/portage... so I made up a |
7 |
> > patch for the places where it does this. |
8 |
> > |
9 |
> > I know crossdev isn't keyworded at all, and this is possibly not the |
10 |
> > place to post patches, but I am new to this. The patch possibly isn't |
11 |
> > perfect as I just inserted ${EPREFIX} into a few lines... |
12 |
> > |
13 |
> > Any suggestions for where to submit future patches would be great. |
14 |
> |
15 |
> The patch looks fine. ${EPREFIX} should have been quoted, however. |
16 |
> |
17 |
> I don't have any strict policy on where or what. Here is fine, |
18 |
> especially if you ask for some discussion, which is what you do IMHO. |
19 |
> |
20 |
> Anyhow, how far do/did you get with this? I have experimented a fair |
21 |
> bit with cross-compiling, with not much success. |
22 |
> |
23 |
|
24 |
I was able to use the crossdev script to do some elementary stuff |
25 |
after those changes. |
26 |
|
27 |
I am trying to get an i686-pc-linux-gnu compiler/toolchain inside my |
28 |
i686-apple-darwin9 prefix. I got to the end of the binutils |
29 |
compilation and it complained about symbols/names that were not going |
30 |
to work at installation. I then hacked around that as it kind of |
31 |
suggested by putting FEATURES="allow_broken_install_names" in |
32 |
make.conf, only to find that gcc complained the next time because the |
33 |
.dylib files were not correct, which was the expected behaviour after |
34 |
hacking around the first error. |
35 |
|
36 |
I will privately email you the log files to see whether you can make |
37 |
any sense of them. |
38 |
|
39 |
Peter |
40 |
-- |
41 |
gentoo-alt@g.o mailing list |