Back in the bad old days (about 30 minutes ago) DNS Crawler would store the entire zone transfer into memory before printing the results to the webpage. This would cause long delays before you see anything happen for large zone transfers. So, I've rearranged the code a bit so that things are more modularized. The result is less code and the transfers gets printed to the webpage as it is transferred. This makes for immediate results for zone transfers that consist of many packets. Zone transfers with only one packet with many answers will still get loaded into memory before getting printed.
There are other ramifications. You will no longer see a count of how many records were in the zone near the top of the page since this is unknown at the time the header was printed. But, if you scroll all the way to the bottom of the page after the transfer is complete, you can see the count there. Also, for lagged dns servers, the transfer may actually terminate prematurely if the connection was lost during the transfer. Again, scroll to the bottom of the page to see how many records were transferred and the error that resulted.
I've also fixed a minor bug that effected all TCP queries, but was mostly a problem in exceptionally large zone transfers that would prematurely terminate the transfer. I've known about this bug for a long time, but didn't realize until now that it actually had a significant impact.
Dave Clark [email] [irc chat]
DollarDNS Services
|