Richard Maciel wrote:
On 11/18/2009 11:10 PM, Kaitlin Rupert wrote:
>> diff -r f834593694d2 -r 815eeee7df1d
>> suites/libvirt-cim/lib/XenKvmLib/vxml.py
>> --- a/suites/libvirt-cim/lib/XenKvmLib/vxml.py Wed Nov 18 17:24:36
>> 2009 -0200
>> +++ b/suites/libvirt-cim/lib/XenKvmLib/vxml.py Wed Nov 18 17:42:15
>> 2009 -0200
>> @@ -987,7 +987,7 @@
>> mem, const.default_mallocunits, None, grstype, address, port_num,
>> keymap, irstype, btype, vnc_passwd)
>> self._os(const.LXC_init_path)
>> - self._devices(mac, ntype, net_name, const.LXC_default_tty)
>> + self._devices(const.LXC_default_emulator, mac, ntype, net_name,
>> const.LXC_default_tty)
>
> I would modify the __init__ function of LXCXML to accept an emulator
> param. We do something similar with KVM. This provides the option for
> specifying other emulators (which might be useful in the future, or
> useful for testing negative test cases).
None of the other virtualization classes (KVMXML, XenXML, XenFVXML)
include the emulator as a param in their __init__. Instead, they define
a set_emulator method that lets the user change it. So, I changed my
code following the established convention.
Ah, yes - you're right. My mistake. I thought Deepti had added this at
some point.
--
Kaitlin Rupert
IBM Linux Technology Center
kaitlin(a)linux.vnet.ibm.com