lp:~jjed/archive-crawler/near-rewrite
An as-of-yet experimental, complete overhaul of archive-crawler. Its scope includes:
* A new test suite based on a stable, artificial package archive
* Removal of the ArchiveCrawler and DesktopDataExtr
* Use of ArchiveCache to allow quick metadata extraction without a full archive copy
* A general cleanup and modularization of the codebase
It leverages existing code, but is rather close to a rewrite. The rationale for these changes are:
a) Maintainability. The scope of ArchiveCrawler has expanded well beyond its initial design. Also, its test suite is based on an every-changing archive, making it unstable.
b) Performance. ArchiveCrawler currently has to read every package in the archive to complete; ArchiveCache instead creates and searches a cache in memory, thus making more use of the CPU.
c) Remote extraction. ArchiveCache requires only a tiny fraction of archive to be present locally. The ability to run ArchiveCache without downloading and storing terabytes of data will ease development.
- Get this branch:
- bzr branch lp:~jjed/archive-crawler/near-rewrite
Branch merges
Branch information
Recent revisions
Branch metadata
- Branch format:
- Branch format 7
- Repository format:
- Bazaar repository format 2a (needs bzr 1.16 or later)