Fulltext task search

[idpaterson]

Currently task searching (not released yet) uses a very basic substring match against task names. It is not particularly efficient and provides no way to rank results. By creating a fulltext search model of the task id and title the workflow can provide much better task search performance and functionality.

An option should be included to improve search results for English tasks by indexing with the Porter stemming algorithm. This allows search terms like "fly" to match tasks containing the word "flies" based on comparison of word stems. Though there are stemming implementations for other languages, SQLite only supports English.