performance improvement:
In the past: to prevent duplicate protocol names (and alike), each time a new protocol was registered, the list of protocols were iterated and the name compared with each existing name using strcasecmp, which is slow as we have >500 protocols right now.
Now: the protocol name to check against duplicates is first converted into a hashvalue and then only this value is checked and stored in a hashtable. This way the string to check for, has to be converted/compared only a single time!
git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@13007
f5534014-38df-0310-8fa8-
9805f1628bb7