On 11/06/2012 01:56 PM, Laine Stump wrote:
On 11/06/2012 12:40 PM, Gene Czarcinski wrote:
> On 11/06/2012 11:50 AM, Gene Czarcinski wrote:
>> I am looking into an alternate method of obtaining the dnsmasq
>> version but the dnsmasq version will be the deciding rule for radvd
>> versus dnsmasq usage.
> Try "dig chaos txt version.bind"
Yeah, I saw that thread (I subscribed to dnsmasq-discuss last week). I
think that's overkill for what we're doing. running "dig" from libvirt
just increases the number of external binaries we're dependent on, but
you still need to parse the text output. Is there any reason for
preferring this over simply running "dnsmasq --help"? (other than maybe
hoping that the output would be less likely to change format over time).
I actually think that it's more important to reduce the count of
external binary dependencies...
It is not clear to me how to make "dig" work anyway ... dnsmasq is not
running yet. So it is back to doing "dnsmasq --version" and capturing
the output in a buffer. Then parsing the buffer for the release and
setting a flag appropriately. Once the flag is set, the actually doing
the code to use dnsmasq or radvd is "simple". Development is a "work in
progress." Naturally, I am doing this on top of the three patches I
submitted earlier today.
For the version, I am going to be looking for "2.63" or later. Besides
being for radvd versus dnsmasq, for DHCPv6 I could add a warning if it
was 2.63 and deny/fail/ignore DHCPv6 definitions if it was earlier.
Overkill?
Gene