Last modified: 2012-12-13 11:19:22 UTC
To avoid later problems the hashable clash should use SHA-1 and not MD5. The later has weaker properties, but is probably good enough. Still SHA-1 is now used for hashing revisions and it could be better to use the same algorithm.
This shift should be done before anything goes into production. It has implications on SnakObject (line 27), ClaimObject (line 100), MapValueHasher (line 57) and StatementObject (line 185).
Verified in Wikidata demo time for sprint 17