1 |
On Sat, Sep 24, 2011 at 6:34 AM, Fabian Groffen <grobian@g.o> wrote: |
2 |
> On 24-09-2011 12:19:14 +0200, Moritz Schlarb wrote: |
3 |
>> -----BEGIN PGP SIGNED MESSAGE----- |
4 |
>> Hash: SHA1 |
5 |
>> |
6 |
>> Hey, |
7 |
>> |
8 |
>> same problem here on SLES 10.1. (glibc-2.4) |
9 |
>> It seems like there is some kind of infinite loop, because the error |
10 |
>> messages always repeat like this many many times. |
11 |
>> |
12 |
>> But! (I want to bootstrap prefix on a high performance computing cluster |
13 |
>> system, where I work and have made available GCC 4.4 and GCC 4.6 by |
14 |
>> compiling my hand and integrating it to the base system) If I compile |
15 |
>> GCC 4.4 (or 4.6) by hand (without doing any magic), It works fine! |
16 |
>> So I'm sorry to say that, but there must be something caused by the |
17 |
>> ebuild... |
18 |
>> If my time allows it, I'll look into it in five minutes, this is just a |
19 |
>> message to everyone having this issue that there may be a light on the |
20 |
>> horizon! ;) |
21 |
> |
22 |
> try USE=vanilla, see if that works |
23 |
|
24 |
Be careful with that. I was able to install gcc 4.5 using -vanilla bit |
25 |
then later on random packages would fail to compile. In particular I |
26 |
had a problem with libxml2, which seemed to get linked to system |
27 |
libraries instead of the ones in prefix. I'm not sure if this is |
28 |
related, all I know is that the problem goes away when using gcc 4.2. |
29 |
|
30 |
Also, have a look at the gcc ebuild the regular portage tree, gcc |
31 |
4.5.2 requires >=glibc.2.8, I assume there is a reason for that. |
32 |
|
33 |
Best, |
34 |
|
35 |
Martin |
36 |
|
37 |
> |
38 |
> -- |
39 |
> Fabian Groffen |
40 |
> Gentoo on a different level |
41 |
> |