[Qemu-devel] case A ... B:

[Qemu-devel] case A ... B:

[Nathaniel Filardo]

Greetings all.

I am attempting to port QEMU to Plan 9 for the Google Summer of Code.
A big chunk (in terms of diff line count) of the work that has been
done on this before was to convert all the lines of the form
    case A ... B:
to be individual case arms as required by Plan 9's C compiler.  There
are currently 200 lines in the code base that use this GNU C
extension, which limits the portability of QEMU to platforms with GCC
or a C compiler supporting this particular GNUism.  Would a
bulk-conversion patch be accepted?

Thanks much.
--nwf;

Re: [Qemu-devel] case A ... B:

[Paul Brook]

On Tuesday 24 July 2007, Nathaniel Filardo wrote:
> Greetings all.
>
> I am attempting to port QEMU to Plan 9 for the Google Summer of Code.
> A big chunk (in terms of diff line count) of the work that has been
> done on this before was to convert all the lines of the form
>     case A ... B:
> to be individual case arms as required by Plan 9's C compiler.  There
> are currently 200 lines in the code base that use this GNU C
> extension, which limits the portability of QEMU to platforms with GCC
> or a C compiler supporting this particular GNUism.  Would a
> bulk-conversion patch be accepted?

qemu already required other gcc extensions for dyngen to work, so IMHO there's 
no point removing the other extensions until you remove that.

Paul

Re: [Qemu-devel] case A ... B:

[Thiemo Seufer]

Nathaniel Filardo wrote:
> Greetings all.
>
> I am attempting to port QEMU to Plan 9 for the Google Summer of Code.
> A big chunk (in terms of diff line count) of the work that has been
> done on this before was to convert all the lines of the form
>    case A ... B:
> to be individual case arms as required by Plan 9's C compiler.  There
> are currently 200 lines in the code base that use this GNU C
> extension, which limits the portability of QEMU to platforms with GCC
> or a C compiler supporting this particular GNUism.  Would a
> bulk-conversion patch be accepted?

The Qemu code generation depends (at least for now) heavily on GCC3 and
its particular behaviour. As long as this is the case I don't see the
point of replacing convenient GNU C extensions.

(Once you manage to get Qemu to work with a different compiler the case
for removing GNU C features will be stronger.)


Thiemo