This avoids naming conflicts in the source directory. In particular, when
encoding MP3 -> MP3, the previous scheme would overwrite the original file
(and hang ffmpeg waiting for input). This should also work in
situations where the source directory is read-only.
Here's another little experiment: to make flexattrs a little easier to use for
end users, you can now get and set them by using 'namespace-key' as the
argument to __getattr__ or __setattr__.
For example, try:
$ beet mod foo-bar=baz
$ beet ls -f '${foo-bar}'
baz
baz
baz
...
I introduced a regression a few commits ago when I started using
lib.destination with the basedir keyword argument as opposed to doing
os.path.join manually.
The major functional change here is how files move around when in keep_new
mode. Now, files are first moved to the destination directory and then
copied/transcoded back into the library.
This avoids problems where naming conflicts could occur when transcoding from
MP3 to MP3 (and thus not changing the filename).
As _print_and_apply_changes itself does for items, we now shortcut
modifications (metadata and filesystem) for albums when no changes are
required for a given album. This avoids effectively doing a "beet move" on an
album even when nothing has changed.
This change uses _album_for_id and _track_for_id instead of the full
autotag.match.* functions. This should be faster (requiring fewer calls to the
MusicBrainz API) while also being more predictable. It also won't, for
example, use acoustic fingerprinting even if the chroma plugin is installed.
Finally, this change catches the error case in which MBIDs are erroneous. This
can happen, for example, if the user has some track MBIDs left over from
before the NGS transition.
The main change here is to use shorter transactions -- one per matching entity
-- rather than one large one. This avoids very long transactions when the
network happens to move slowly.
