|
From: Christian P. <tr...@ge...> - 2005-08-10 02:44:21
|
Heya, well, well known, I'm on amd64, however, everytime I'm [re]building my loca= l=20 trunk I must see, that it's also compiling for ppc and x86; I *could* understand why it's compiling x86 (are they sharing code?), but f= or=20 sure not why it is compiling files containing the substring "ppc"). Is there any reason for this? Thanks in advance, Christian Parpart. =2D-=20 02:50:26 up 138 days, 15:58, 1 user, load average: 1.68, 2.43, 2.20 |
|
From: Tom H. <to...@co...> - 2005-08-10 06:15:14
|
In message <200...@ge...>
Christian Parpart <tr...@ge...> wrote:
> well, well known, I'm on amd64, however, everytime I'm [re]building my local
> trunk I must see, that it's also compiling for ppc and x86;
> I *could* understand why it's compiling x86 (are they sharing code?), but for
> sure not why it is compiling files containing the substring "ppc").
This seems very unlikely given at least some of those files won't compile
on the wrong architecture. It would also tend to cause duplicate symbol
errors as those files tend to provide the same functions for each platform.
Can you provide an example of a file which you thing is being wrongly
compiled?
Oh, hang on, I bet you're talking about the VEX code? That does compile
support for all guest and host platforms because the whole point is that
it can (at least in theory) translate from one instruction set to another.
Tom
--
Tom Hughes (to...@co...)
http://www.compton.nu/
|
|
From: Julian S. <js...@ac...> - 2005-08-10 10:08:49
|
> Oh, hang on, I bet you're talking about the VEX code? That does compile > support for all guest and host platforms because the whole point is that > it can (at least in theory) translate from one instruction set to another. Yes. Another and more immediate reason for compiling in all host and guest architectures on vex is to reduce the number of different build-time configurations that we have to ensure keep working. J |