Hi, an import thing for fuzzing performance is the raw speed. As it currently stands, there is a lot of waiting on the thread in the resolver to join. I tried to hack around this by invoking getaddrinfo_a instead of getaddrinfo, so I could add a timeout, but I failed. I got it to work, but had some lifetime issues so I gave up.
Disabling the threaded name resolution makes the fuzzing not work at all, for some reason. Is there an easy way to (locally) add a hack to get the resolving done quicker? Perhaps by hard coding everything to resolve to a fixed adress? The fuzzing speed is below 100 tries per second, and that is at least a factor of ten too low. Paul ------------------------------------------------------------------- Unsubscribe: https://cool.haxx.se/list/listinfo/curl-library Etiquette: https://curl.haxx.se/mail/etiquette.html
