1 |
On 09/21/10 12:41, Kevin O'Gorman wrote: |
2 |
> On Mon, Sep 20, 2010 at 3:23 PM, Beau Henderson <beau@××××××××××××.com |
3 |
> <mailto:beau@××××××××××××.com>> wrote: |
4 |
> |
5 |
> I had this same problem and decided I had bad RAM. Before I could order any, I rebuild my system |
6 |
> and it happens that I did so with an image that had GCC 4.3* rather than 4.4. Funny enough, |
7 |
> firefox worked just fine. I did some searching and apparently nspr has issues with a certain |
8 |
> function enabled in -O2 @ gcc 4.4. |
9 |
> |
10 |
> From the following: https://bugzilla.redhat.com/show_bug.cgi?id=487844 |
11 |
> |
12 |
> Apparently if you rebuild nspr @ gcc 4.4 with -O2 -fno-strict-aliasing |
13 |
> |
14 |
> I haven't confirmed this, as I haven't had time to jump back to 4.4 but if someone can confirm |
15 |
> this fixes the issue, I'd certainly be greatful! |
16 |
> |
17 |
> |
18 |
> I'm still at 4.3.4, and having these problems. I wouldn't be holding my breath for a silver bullet. |
19 |
> I'm writing this on chormium, having just given up on Opera for being slow as FF. Sigh. |
20 |
> |
21 |
> |
22 |
> -- |
23 |
> Kevin O'Gorman, PhD |
24 |
> |
25 |
|
26 |
Are you compiling nspr with -O3 by any chance ? The flag that is responsible was apparently moved |
27 |
from -O3 in gcc 4.3* to -O2 in 4.4*. |
28 |
|
29 |
Are you getting the seg fault when you strace firefox ? |
30 |
|
31 |
-- |
32 |
Beau Henderson |