GSoC/GCI Archive
Google Code-in 2013 KDE

Optimize Baloo Indexing - FileIndexingQueue

completed by: Benjamin Kaiser

mentors: Vishesh Handa

Baloo is the new file indexing service in KDE. It currently indexes all the files in 2 phases - Basic Indexing and File Indexing. The Basic Indexer just indexes the file url and mimetype. The File Indexing in contrast looks into the contents of the file.

This is implemented via a FileIndexingQueue and a FileIndexingJob. Both of these jobs currently store the entire url of the file. This is not required as we just need the unique identifier of the file, which is an unsigned integer.

Optimize the FileIndexingQueue and FileIndexingJob to only use integers. It should also pass integers to the baloo_file_extractor and that should lookup the url using the FileMapping class.

This will reduce the memory consumed by the file indexer, and it should increase performance.

Repository: kde:baloo
Code: baloo/src/file/