All hashing libraries i used before had some kind of more or less annoying limitations. For example:

• key sizes are often limited to a few or a few hundred bytes. While this is often enough is sometimes bites.
• key values are sometimes limited to a restricted character set.
• data sizes are also often limited.
• data values are sometimes limited to a restricted character set.
• the number of keys is often limited. Sometimes the libraries haven't even been tested against a large number of keys, which is a recipe for disaster.
• keys may need to be unique.
• the implementations sometimes only have a high-level interface which make it hard to put wrappers around it. Strhash on the other hand only has a low-level interface and is explicitly meant to be used to but higher layers on top of it.

strhash doesn't have any limitation besides a key and data length limitation of 31 bits. strhash has been tested against a million keys, there a no silent limitations anywhere. strhash can deal with duplicate keys (albeit this will somewhat reduce the performance).

strhash hasn't been written to be the fastest possible implementation, but is reasonable fast. On the bright side it can deal with almost any data you throw at it.

Links

the strhash homepage