[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [MiNT] [Mint-cvs] [FreeMiNT CVS] freemint/sys/sockets/inet4
On Sat, 2010-01-02 at 00:10 +0100, Vincent Rivière wrote:
> Odd Skancke wrote:
> > My main developmentenvironment is GCC 4.4.2/KDE
> > on Fedora. On this beast, it takes 16 seconds to build the kernel from
> > distclean. I dont want to loose that luxury :)
> This is the voice of reason.
> It is a fact that modern versions of GCC require strong hardware. It
> would be a nonsense to run it on older Atari machines, since the *exact*
> same result can be achieved with a cross-GCC running on a modern computer.
Not always. There are some situations where building natively gets the
correct answer. Think of runtime tests, whereas in the cross-compiler
situation the software has to guess, which may not always be right.
> There are a lot of people that hate cross-compilers. Fortunately, it is
> not the case of everyone. I hope the developers of games on GameBoy or
> cell phones don't run GCC on that hardware :-)
> Be sure that bytes are clean entities. The bytes produced by GCC running
> on Atari hardware and the ones produced on a cross-GCC running on
> Windows or Linux are exactly the same. One to one. They don't smell bad.
> For me, the choice has been made for a long time. I use my main modern
> computer to write and compile programs, then I can run them on my
> favourite hardware (or emulators).
Remember I'm building Gentoo on my CT60 Falcon and it works building
from source, albeit some packages can take a while to compile such as
GCC, but the majority are relatively quick.
> > vice versa, it will be very difficult keeping this under control via
> > #ifdefs within a single source tree. And I hate these kinds of #ifdefs :-)
> What are those huge differences ?
> GCC 4.x is more picky than earlier versions, thus the code has sometimes
> to be cleaned up. But the resulting code should work perfectly on older
> compilers. Or except extremely specific cases.
> Am I wrong ?