1 |
Frank Peters posted on Thu, 30 Jun 2011 21:04:29 -0400 as excerpted: |
2 |
|
3 |
> On Fri, 01 Jul 2011 02:44:36 +0300 Nikos Chantziaras <realnc@×××××.de> |
4 |
> wrote: |
5 |
> |
6 |
>> And here you can read more thorough information about strict aliasing: |
7 |
>> |
8 |
>> http://labs.qt.nokia.com/2011/06/10/type-punning-and-strict-aliasing |
9 |
>> |
10 |
> You've saved the day in more ways than one. |
11 |
> |
12 |
> A few days ago I posted about a possible problem with a floating point |
13 |
> test called the UCBTEST. After examining the source code of this test, |
14 |
> I see violations of aliasing rules throughout. It's hard to efficiently |
15 |
> manipulate variables without them. |
16 |
> |
17 |
> Of course, this code was written before the C99 standard and so is |
18 |
> exempt. But the use of "-fno-strict-aliasing" will apply with the |
19 |
> UCBTEST as well. |
20 |
|
21 |
Truthfully, I thought sure this was a direct follow-on to that! So I'm |
22 |
happy to see that the same cause and solutions apply. =:^) |
23 |
|
24 |
As I said in an earlier post, tho, expect there to be more problems like |
25 |
that with code of that age as time goes on, because they're well past the |
26 |
easy optimizations now, and into stuff like this. So it may be wise to |
27 |
keep an eye out and test the minor gcc version bumps with an eye toward |
28 |
such issues. They generally put out a porting guide, etc, with the new |
29 |
version, that you can read for clues in case the tests start doing |
30 |
unexpected things again. |
31 |
|
32 |
(If you pay attention to flameeyes' blog, he tends to cover such things |
33 |
reasonably early on after release of a new gcc version as well, as he |
34 |
does tinderbox runs to see how bad the interaction of the new version is |
35 |
with the current gentoo ~arch tree as a whole and eventually to test that |
36 |
the latest ~arch packages are updated or patched to fix the problem, |
37 |
before that gcc version gets unmasked to ~arch. By no means do I agree |
38 |
with everything he says, but he really is an asset to gentoo and would be |
39 |
sorely missed should one of his sick episodes make it impossible for him |
40 |
to continue, or if he simply decided he had better things to do with his |
41 |
time. The switch to --as-needed in Gentoo's default ldflags was largely |
42 |
due to his work, explaining the issues for other devs, testing and filing |
43 |
bugs, helping with the hard cases, and in general pushing it until it |
44 |
happened, for instance, and every single gentoo user benefits from that |
45 |
in the form of less mandatory rebuilds, every time they update. It may |
46 |
well have happened without him, but it might have taken a decade longer, |
47 |
too. And because gentoo devs normally push those patches upstream where |
48 |
they can, that has been a benefit to the entire FLOSS community, not just |
49 |
gentoo, as well, reducing "dependency hell" for everyone.) |
50 |
|
51 |
-- |
52 |
Duncan - List replies preferred. No HTML msgs. |
53 |
"Every nonfree program has a lord, a master -- |
54 |
and if you use the program, he is your master." Richard Stallman |