1 |
Sorry for the delay. |
2 |
|
3 |
On 05-11-2008 03:02:34 -0700, David Hubbard wrote: |
4 |
> John, Fabian, |
5 |
> |
6 |
> I got the cannot compute sizeof (long long) error when attempting to |
7 |
> build sys-devel/gcc-apple-4.2.1_p5564 on OS X 10.4.11 (intel), and |
8 |
> found the discussion on gentoo-alt. |
9 |
> |
10 |
> Though I can't say what the "unknown one-operand" error is (in |
11 |
> config.log), here is the patch that adds a call to dsymutil there: |
12 |
> http://gcc.gnu.org/ml/gcc-patches/2006-09/msg00057.html |
13 |
> |
14 |
> And a follow-up which may indicate dsymutil isn't really the right solution: |
15 |
> http://gcc.gnu.org/ml/gcc/2008-10/msg00083.html |
16 |
> |
17 |
> So I reverted the part of the patch that calls dsymutil, and the |
18 |
> ebuild now works. (Patches attached) |
19 |
|
20 |
Wow, thanks! That's very nice work! |
21 |
|
22 |
> Thus I would call it a gcc bug, but only a gcc-apple-4.2.1 bug on OS X |
23 |
> 10.4.11, and the revert leaves the original problem of finding the |
24 |
> dsym folder for objects packaged in a library. And none of this |
25 |
> applies on 10.5 and above. So all in all, not worth mentioning to the |
26 |
> gcc developers. |
27 |
|
28 |
I prefer to call it a gcc-apple-4.2.1 bug on OSX 10.4/x86, as it works |
29 |
for me on 10.4/ppc. 10.5 indeed is unaffected. |
30 |
|
31 |
Why is the -no-dynamic thing necessary? For bootstrapping? |
32 |
|
33 |
|
34 |
> P.S. Please CC me on the gentoo-alt thread |
35 |
|
36 |
|
37 |
|
38 |
|
39 |
-- |
40 |
Fabian Groffen |
41 |
Gentoo on a different level |