Hi folks, Getting this on the OpenFlow mailing list. Also changed the subject to be slightly more related to the actual topic than "Gerrit is stuck" :)
Anyway Josh - do you want to summarize your solution (and the problem) for the folks new to this discussion? Thanks, Abhijit On Tue, Jul 19, 2016 at 4:19 PM, Josh Hershberg <[email protected]> wrote: > Which I think my patch is! The differences between the Match objects in > question are syntactic and not semantic. I added a compare method that > ignores the structural differences. > > On Tue, Jul 19, 2016 at 12:45 PM, Sam Hague <[email protected]> wrote: > >> As Josh mentions the issue is that the match object coming back is not >> the same as what was written to mdsal so it ends up as a different flow. We >> need a better way to compare the flows. >> >> On Jul 19, 2016 2:43 AM, "Josh Hershberg" <[email protected]> wrote: >> >> Adding Sam. >> >> Sam, this is the alien flow issue. >> >> On Tue, Jul 19, 2016 at 5:35 AM, Anil Vishnoi <[email protected]> >> wrote: >> >>> +adding other folks >>> >>> Hi Josh, >>> >>> Thanks for reminding me of the patch. >>> >>> Andrej/Jozef, This is the patch, regarding which i asked a question on >>> how we currently compare the operational flow and config flow in lithium >>> plugin? >>> >>> In the helium plugin we have custom comparators to compare the config >>> and operational flow that do explicit match using the matching construct of >>> the flow ( + cookie). Looks like in lithium plugin we are using >>> hashcode/equals for the same ( i didn't get a chance to look at the code >>> yet), which i am afraid won't work for all the scenario, specially >>> augmentation. Using custom comparator for matching flow for augmentation is >>> also not really a good idea, because there can be many extension (openflow >>> and vendor specific) and handling them in the plugin is not a good idea. In >>> helium plugin, i added cookie as a part of the flow comparison, so if you >>> have two flows that is different only by the extension matches, then user >>> can use cookie to differentiate the flows, so custom comparator don't have >>> to match based on the augmentation. >>> >>> Can you please share some more details on the existing comparison >>> mechanism and lets discuss how we can resolve this issue in better way. >>> >>> Thanks >>> Anil >>> >>> On Sat, Jul 16, 2016 at 9:49 PM, Josh Hershberg <[email protected]> >>> wrote: >>> >>>> Hi Anil, >>>> >>>> Could you please take a look at this gerrit? Just want to make sure >>>> it's not buried or stuck or something. >>>> >>>> thanks, >>>> >>>> Josh >>>> >>>> https://git.opendaylight.org/gerrit/#/c/41542/ >>>> >>> >>> >>> >>> -- >>> Thanks >>> Anil >>> >> >> >> >
_______________________________________________ openflowplugin-dev mailing list [email protected] https://lists.opendaylight.org/mailman/listinfo/openflowplugin-dev
