Hello Seth, On Sep 3, 2012, at 19:54 , Seth Mattinen wrote:
>> 1b. Is VARBINARY the best way to do it for MySQL? > > Yes. A similar issue was seen in sql-based bayes databases for > SpamAssassin tokens and the solution is to use BINARY for the token col > instead of CHAR. Generic fixes are best. Agreed. Always good to hear from experience :) >> 2b. Do you think type=NULL (SQL NULL) is an ugly hack? If so, what else >> should we do? > > If it's internal-only and handled automatically by rectify-zone it makes > sense to not have a type and that it would be null. External management > tools can be easily modified to ignore rows were type is null. Yes, agreed. >> 2c. If we accept type=NULL as an acceptable notation, should we still have >> this extra field just to make cleanup easier? > > A 'virt' flag wold be more future-proof if in the future virtual records > needed a type for some currently unforeseen reason. Management tools can > likewise ignore 'virt=true' records. If they have a type, they're not virtual - as far as I can tell for now. I also don't foresee many more changes to the fundamentals of DNS, after DNSSEC. We will ponder it some more... Kind regards, -- Peter van Dijk Netherlabs Computer Consulting BV - http://www.netherlabs.nl/ _______________________________________________ Pdns-users mailing list [email protected] http://mailman.powerdns.com/mailman/listinfo/pdns-users
