Release Notes: This release adds memory and performance improvements (up to two times faster than the 3.1 codebase), several other lexers, and the new flags -s (to split large code trees), and -u (to break up num,num,num,num runs in CTF files so that these runs of tokens are not compared).
Release Notes: The codebase was completely rewritten, making the tool 5 to 10 times faster than the 2.x versions. The codebase is now structured as a library with command-line front-end tools.
Release Notes: ctcompare's speed has been doubled, with some hand-coding of the hotspot loops. A new database has been added, with read speeds comparable to Berkeley DB 4, but with faster write speeds. Compile-time support for code comparisons within a source tree has been added.
Release Notes: Significant comparison performance improvements (8 times or more).
Release Notes: Speed has been significantly improved, especially when there are large amounts of similarities between code trees. This release has a command line option to sort the results by the number of tokens in a run. Initial support for Java source files has been added.
Release Notes: The comparison method has been completely rewritten. A database is now used to hold 16-token "tuples" as keys; the result attached to each key is the list of source files which have that tuple. Tuples with multiple files from different source trees indicate potential code similarity. These are then fully tested to find actual code similarity.