On Wed, Apr 18, 2018 at 9:49 AM, Robert Varga <[email protected]> wrote:

> Hello Anil,
>
> No, support is not being removed, but Ipv4Address.equals() is defined in a
> way which returns false when compared with Ipv4AddressNoZone of the same
> value IIRC, hence things would break.
>
​This is not correct behavior in my opinion, because ietf clearly says that
user "may" add zone, so the equals should not fail if two ipv4-address
(without zone) is compared.​


>
> Yes, applications can support zones and conversion from NoZone is cheap,
> so it is a matter of preference.
>
> Sent from my BlackBerry - the most secure mobile device - via the Orange
> Network
> *From:* [email protected]
> *Sent:* April 18, 2018 6:13 PM
> *To:* [email protected]
> *Cc:* [email protected]; [email protected];
> [email protected]; [email protected];
> [email protected]
> *Subject:* Re: Build breakage in openflowplugin due IP address NoZone
> changes
>
> Hi Robert,
>
> Please see inline..
>
> On Wed, Apr 18, 2018 at 1:33 AM, Robert Varga <[email protected]> wrote:
>
>> On 18/04/18 08:09, D Arunprakash wrote:
>> > Hello,
>> >
>> > The following review in mdsal might have impacted openflowplugin
>> > functionality.
>> >
>> > https://git.opendaylight.org/gerrit/#/c/70769/
>>
>> Sorry about that.
>>
>> > https://jenkins.opendaylight.org/releng/job/openflowplugin-m
>> aven-verify-fluorine-mvn33-openjdk8/259/console
>> >
>> >
>> >
>> > org.opendaylight.yang.gen.v1.urn.ietf.params.xml.ns.yang.iet
>> f.inet.types.rev130715.Ipv4Address<Ipv4Address{_value=0.1.2.3}>
>> > but was:
>> > org.opendaylight.yang.gen.v1.urn.ietf.params.xml.ns.yang.iet
>> f.inet.types.rev130715.Ipv4AddressNoZone<Ipv4Address{_value=0.1.2.3}>
>> >
>> > is it new expectation on openflowplugin to change from Ipv4Address to
>> > Ipv4AddressNoZone ?
>>
>> There are two aspects to this.
>>
>> As an interim, the codecs from wire need to be updated to convert
>> Ipv4AddressNoZone to Ipv4Address, so that equality works as expected and
>> the breakage is recovered -- https://git.opendaylight.org/gerrit/71072
>> does that.
>>
> ​This seems like regression. ietf-type (2013-07-15) supports both the
> version and this revision seems to be backward compatible. So are you
> removing the support for ipv4-address type? ​
>
>
>>
>> Going forward, though, I believe the openflow models need to be updated
>> to require ipv4-address-no-zone rather than ipv4-address (and same goes
>> for ip-address and ipv6-address). This really is the correct thing to do
>> -- ipv4-address is not really the IPv4 address used in OpenFlow protocol.
>>
> ​Yeah for flows /groups true, but that does not mean application can't use
> ipv4_address with zone.​
>
>
>>
>> Regards,
>> Robert
>>
>>
>
>
> --
> Thanks
> Anil
>



-- 
Thanks
Anil
_______________________________________________
openflowplugin-dev mailing list
[email protected]
https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev

Reply via email to