1 |
On Sun, 2004-08-01 at 06:02 +0100, Ciaran McCreesh wrote: |
2 |
> | [And note, I actually think the argument we all tend to use to -march |
3 |
> | in CFLAGS is, in Gentoo's world, the relevant "architecture"] |
4 |
> |
5 |
> -march= is an x86 thing which is going to go away at some point. The |
6 |
> correct way to do it is to use either CHOST or CCHOST. |
7 |
|
8 |
Ok, fine, but the mandated CHOST of "i686-pc-linux-gnu" doesn't actually |
9 |
sufficiently describe the limitations of code generated - if I'm |
10 |
compiling pentium3 code, the it won't run on a i686. |
11 |
|
12 |
I poked around looking for definitions of CCHOST; I read |
13 |
crosscompile.eclass and it seems not to differentiate in the pentium |
14 |
family newer than pentium II (i686). |
15 |
|
16 |
Or are you saying that you expect Gentoo to drop support for compiling |
17 |
specific processor optimizations provided by GCC? |
18 |
|
19 |
AfC |