Changes for version 2.00_02
- ENHANCEMENTS
- Rewrote Algorithm::SkipList::Node. Now based on the Tree::Node package, which is written in C for significant improvement in memory usage, and a slight improvement in speed
- Cleaner code in various methods
- Option to set minimum level
- INCOMPATIBLE CHANGES
- changed interface for Node package, but since it is meant for internal use, this should not be a problem for most users
- accessing the key and value accessors of the header will cause an error
- _search() and _search_with_finger() return their values in a different order; again these are for internal use...
- duplicates option renamed to allow_duplicates
- other name and interface changes to internals
- the levels method does not return a value equal to the number of levels in the list header node (list method)
- the order of arguments to the copy method has changed to copy(low, high, finger) so as to be consistent with other methods
- The following methods are not yet implemented in this release: merge find_duplicates _node_by_index key_by_index index_by_key value_by_index
- OTHER CHANGES
- Updated and reorganized documentation.
- Rewrote tests from scratch.
- Level threshold not readjusted during deletes.
- Benchmarks.txt file not included with this release.
Modules
Perl implementation of skip lists
header node class for Algorithm::SkipList
a dummy node class for Algorithm::SkipList