What is this?
This is a set of test cases to look at if you're building a DNS recursive resolver. It includes weird queries, weird answers, and other things your server might want to handle.
How is it organized?
Each test is in its own folder that contains a couple of files:
README.mdwill tell you how to run the test. It doesn't give too many details on what the test is doing (on purpose)!whats_happening.mdhas everything. What's going on in the test, what the expected behavior should be. My intent is that you check it after you think you've got a solution for the behavior in the test, but do whatever helps you learn the best!run_test.sh, where it exists, is a script to run the test by sending a packet against a DNS server.packet, where it exists, is the packet contents itself. DNS packets aren't printable binary, so you'll probably want to open it in a hex editor or examine it with a tool like Wireshark!
The test folders are just numbered in the order I wrote them, which is somewhat arbitrary but tries to be in increasing "trickyness." Tests 1-5 are straightforward and don't do anything too tricky with DNS. Tests after that are a bit more sinister.
Will it work in my DNS server written in Rust/Haskell/Clojure/Shakespeare?
These tests all treat the DNS server as a black box, so they should be fully language agnostic.
How do I know if the test passed?
On a first pass, you might not! Some of the fun of debugging these is breaking down what's happening and determining what behavior should occur. The README will explain if the test has a somewhat obscure pass condition or if there's a particular part of your server to pay attention to when running the test. For others, failures will be pretty clear (e.g. not return a reply or crash the whole server).
You can always look in the whats_happening file if you're unsure!
Can I use these against a production server?
Yes! If you don't control that server, you should probably check with the administrator of it first, though. Some of these tests have the possibility of crashing servers.
What work is still todo on this?
I'm going to organize the tests better than just test1, test2, and so on. I
also want to add support to the test run script for showing the response
received (and maybe even validating it) at some point.
And there's a lot more things I want to test.