1 |
On Thursday 06 July 2006 19:51, Ciaran McCreesh wrote: |
2 |
> And for a single compile? |
3 |
I always leave the two of them in sync, even C++ apps might have parts |
4 |
building CFLAGS. In case you know you're going to use only C++ is not |
5 |
difficult to use |
6 |
|
7 |
CFLAGS=${CXXFLAGS} has_cpuset 3dnow |
8 |
|
9 |
don't you think? |
10 |
|
11 |
> And your assumption would be wrong. I can assure you that relying upon |
12 |
> -march doing anything sensible with __MAGIC__ is entirely unsafe. Go |
13 |
> and look at the VIS handling if you want a perfect example. |
14 |
Okay, maybe VIS handling is broken. But we can rely pretty safely on x86, |
15 |
amd64 and PPC gcc to know the table of arches and extensions supported. |
16 |
Remember that I asked to talk with SPARC team for VIS just because I only |
17 |
know about the other three arches. |
18 |
|
19 |
> No no. Where "regain control" means the user has to screw around with |
20 |
> nasty hacks and pray, rather than setting a well defined, specific |
21 |
> variable. |
22 |
Find me a reason to do that, a part for broken MMX code that should be |
23 |
disabled on the ebuild itself already. |
24 |
|
25 |
> Uh. USE flags are available at DEPEND time. |
26 |
If you talk about the nasm dependency, then it is rare, most of the MMX |
27 |
support is inline in C sources anyway. |
28 |
|
29 |
> And at the metadata phase? |
30 |
Should be already transparent or something is strange. nasm is simpler to add |
31 |
the dependency for x86, there is really few people not enabling mmx already. |
32 |
Yes it is a bit of regression, but for a small percentage of users, while |
33 |
there's more safety for many other people. |
34 |
|
35 |
> Er. No. Not at all. The __MAGIC__ isn't guaranteed. Quite the contrary. |
36 |
This ain't no magic. The magic is in the _CODE_ that GCC creates, but not in |
37 |
the _DEFINES_ that GCC emits. |
38 |
|
39 |
> You're trying to guess what the user wants based upon hairy magic, |
40 |
No, about their chosen architecture. |
41 |
|
42 |
> rather than doing what the user says (aren't you always yelling at |
43 |
> upstreams for doing that?) |
44 |
The user asks for athlon64 support? They get athlon64 (mmx, 3dnow, 3dnowex, |
45 |
sse, sse2) |
46 |
The user asks for G3 support? They get G3 (nothing) |
47 |
The user asks for Pentium4 support? They get what they want (mmx, sse, sse2, |
48 |
sse3 in case) |
49 |
|
50 |
-- |
51 |
Diego "Flameeyes" Pettenò - http://farragut.flameeyes.is-a-geek.org/ |
52 |
Gentoo/Alt lead, Gentoo/FreeBSD, Video, AMD64, Sound, PAM, KDE |