= DEP-11 Data Extractor TODO List =

=== Known issues ===

 * If an icon is symlinked, we just ignore it at time, since the symlink itself
   doesn't contain data.
   We should either support that case or drop a meaningful error message
   (Reference package: git-cola == 2.2.1-1)
 * The extractor can't get a list of files for certain .deb packages, see
   'qtwebkit-opensource-src/libqt5webkit5-dbg_5.4.2+dfsg-3_amd64.deb' as an example.
   Might be an issue with python-apt, some multiprocessing insanity leading to this
   happening sometimes, or could even be a problem with the package itself.
   This issue happens only rarely.

=== Planned Features ===

 * Extract localizstion status for AppStream components and add them as
   `Languages` field.
 * Extract more metadata from things which do not have AppStream upstream
   metadata yet.
 * Expand the HTML pages to include more and more useful information.

=== Whishlist / Random Ideas ===

 * Maybe pre-filter for interesting packages based on Contents.gz
   to speed up the generator.
 * Could we maybe scan just one .deb package per architecture, and simply
   reuse the previously extracted metadata for all other architectures, given
   that checksums on the files relevent for the metadata match between archs?
   Would need to be stored on the database somewhere, and might grow into a pretty
   big storage of checksums, so do we want that?
   Also, would something like this save a sufficient amount of disk-space for
   screenshots/icons and improve the generator speed to make implementing this
   functionality worth the effort?
