On Fri, 2005-06-24 at 22:30 +0100, Tim Walker wrote:
> Thanks for the replies.
> I thought the Live CD was a bad idea until I realised it can still be
> booted under QEMU for non x86 users (Live CDs can be created for other
> platforms - have to pick one otherwise it'd be a nightmare). The
> tradeoff would be speed for x86 users vs ease of image maintenance and
> build environment precision (host processor). There is an appealing
> match between the desired uniformity of build environment and QEMU's aim
> to uniformly emulate different platforms regardless of host. It could
> also be a good test of QEMU itself in the same way GCC self compiles.
Yes, the idea would be to have a built QEMU.EXE so that you could boot
the livecd under windows.
I would think the more likely thing to do is to boot it into linux and
run it native. DSL will run completely from RAMdisk (in that mode it's
actually faster than a native install) but it depends on how much memory
As to non x86 platforms, yeah the the live CD could be run under QEMU. I
don't know how many folks are itching to build for Windows under PowerPC
or ARM, but hack, who knows...
> Cross compile to Windows from Linux is an unknown quantity to me - has
> anyone tried it yet with QEMU?
I believe I had it working at some point. mingw cross compiler is in
> Depending on how it's used this could either be a two edged sword or a
> virtuous circle - headlines of "Convenient Build Environment demands
> massive investment of time to keep functional" or "Convenient Build
> Environment helps find QEMU inconsistencies and vice versa". Something
> like that.
Yes, there would be no point in doing this if it is very hard to do and
keep maintained. Which is why I suggest DSL which has some package
management and remastering stuff built-in. Also you have full access to
Debian apt repositories, including mingw.
Qemu-devel mailing list