Re: [Libvir] [patch 3/3] Use libuuid to generate UUIDs
On Mon, Feb 26, 2007 at 02:28:30PM +0000, Mark McLoughlin wrote:
If libuuid from e2fsprogs is available, we use that to generate UUIDs.
On the one hand I think its good to re-use existing libraries where they fullfill our needs, but then we need to keep a built-in impl of UUID generator for non-Linux systems which typically won't have a e2fsprogs available. So rather than maintaining two separate impls of the UUID code, which splits the testing coverage, I'd be inclined to just always use the builtin version. Guarenteeing consistent algorithm / semantics for UUID generation across all platforms. Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|
On Mon, 2007-02-26 at 14:50 +0000, Daniel P. Berrange wrote:
On Mon, Feb 26, 2007 at 02:28:30PM +0000, Mark McLoughlin wrote:
If libuuid from e2fsprogs is available, we use that to generate UUIDs.
On the one hand I think its good to re-use existing libraries where they fullfill our needs, but then we need to keep a built-in impl of UUID generator for non-Linux systems which typically won't have a e2fsprogs available. So rather than maintaining two separate impls of the UUID code, which splits the testing coverage, I'd be inclined to just always use the builtin version. Guarenteeing consistent algorithm / semantics for UUID generation across all platforms.
I don't have a strong opinion either way ... I'd prefer us to not have any UUID generation code of our own at all since we'll inevitably have the rfc4122 zealots coming down on us, but we'll always have to have that code, so ... I didn't optionally use uuid_parse() for the exact reasoning you describe, but I think that reasoning is less important for UUID generation since we're effectively treating them as opaque random numbers. You're happy with the other two patches, then? Cheers, Mark.
On Mon, Feb 26, 2007 at 03:10:24PM +0000, Mark McLoughlin wrote:
On Mon, 2007-02-26 at 14:50 +0000, Daniel P. Berrange wrote:
On Mon, Feb 26, 2007 at 02:28:30PM +0000, Mark McLoughlin wrote:
If libuuid from e2fsprogs is available, we use that to generate UUIDs.
On the one hand I think its good to re-use existing libraries where they fullfill our needs, but then we need to keep a built-in impl of UUID generator for non-Linux systems which typically won't have a e2fsprogs available. So rather than maintaining two separate impls of the UUID code, which splits the testing coverage, I'd be inclined to just always use the builtin version. Guarenteeing consistent algorithm / semantics for UUID generation across all platforms.
I don't have a strong opinion either way ... I'd prefer us to not have any UUID generation code of our own at all since we'll inevitably have the rfc4122 zealots coming down on us, but we'll always have to have that code, so ...
I didn't optionally use uuid_parse() for the exact reasoning you describe, but I think that reasoning is less important for UUID generation since we're effectively treating them as opaque random numbers.
You're happy with the other two patches, then?
Yeah, no problem. Dan. -- |=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=| |=- Perl modules: http://search.cpan.org/~danberr/ -=| |=- Projects: http://freshmeat.net/~danielpb/ -=| |=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|
participants (2)
-
Daniel P. Berrange -
Mark McLoughlin