On Wed, May 13, 2009 at 11:46:58AM +0200, Daniel Veillard wrote:
On Mon, May 11, 2009 at 03:29:42PM -0400, Laine Stump wrote:
> Yes, tun interfaces too. Since this is binary data rather than a
> null-terminated string,
> we need to decide among the following three choices:
>
> 1) have a fixed length (how long? is 16 bytes long enough?) and
> zero-fill the shorter ones.
>
> 2) Add a macLen arg to any API function that uses mac address (this will
> need to be a return arg in some cases too)
>
> 3) Only provide the versions of the functions that accept/use ASCII mac
> address args.
IMHO, I would play safe at this point and pick 3)
First it's sufficient, from the ASCII version people can usually derive
the binary one if they really need it, but mostly I think people asked
for those interfaces at the libvirt level because they wanted the
ability to not mess with the low level stuff, so we should focuse on
the high level. And if this proves unsufficient we can stil add new APIs
based on the people feedback.
Ok, so we just need to get rid of the constants in the public header
file, and remove these APis
virInterfacePtr virInterfaceLookupByMAC (virConnectPtr conn,
const unsigned char *mac);
int virInterfaceGetMAC (virInterfacePtr interface,
unsigned char *mac);
and change the remote_protocol.x file to use 'remote_nonnull_string'
instead of the fixed length mac addr on the wire.
And in the virInterfacePtr struct in src/datatypes.c also use the null
terminated string, instead of fixed byte array.
Regards,
Daniel
--
|: Red Hat, Engineering, London -o-
http://people.redhat.com/berrange/ :|
|:
http://libvirt.org -o-
http://virt-manager.org -o-
http://ovirt.org :|
|:
http://autobuild.org -o-
http://search.cpan.org/~danberr/ :|
|: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|