On 8/30/19 9:58 AM, Richard W.M. Jones wrote:
As we discussed on IRC, patch 1 is obvious and uncontroversial,
let's
do it.
I think patch 2 is overkill really, and weird pointer tricks have a
habit of breaking on big endian / 32 bit / obscure platforms. Why
don't we just create a brand new filter API number (6 I think), reject
everything with API != 6, and as a further test verify that
strcmp (filter->version, PACKAGE_VERSION) == 0 ?
Works for me.
--
Eric Blake, Principal Software Engineer
Red Hat, Inc. +1-919-301-3226
Virtualization:
qemu.org |
libvirt.org