Minor Description
The selenium-based protocol plugins (protocol-selenium, protocol-interactiveselenium, protocol-htmlunit) are partial copies of protocol-http to sniff HTTP headers in order to determine the content type and if it's not HTML to implement a fall-back protocol to download the content.
While there have been significant improvements and fixes in protocol-http (NUTCH-2549 and related), these haven't been transferred to the selenium-based protocol plugins. They should, or even better the code for the fall-back protocol implementation should be shared.