diff --git a/beets/config_default.yaml b/beets/config_default.yaml
index ba58debe7..545fa9638 100644
--- a/beets/config_default.yaml
+++ b/beets/config_default.yaml
@@ -65,6 +65,7 @@ ui:
 format_item: $artist - $album - $title
 format_album: $albumartist - $album
 time_format: '%Y-%m-%d %H:%M:%S'
+format_raw_length: no
 
 sort_album: albumartist+ album+
 sort_item: artist+ album+ disc+ track+
diff --git a/beets/dbcore/db.py b/beets/dbcore/db.py
index d54e3a0d3..202cf006f 100644
--- a/beets/dbcore/db.py
+++ b/beets/dbcore/db.py
@@ -467,6 +467,11 @@ class Model(object):
 
         return cls._type(key).parse(string)
 
+    def set_parse(self, key, string):
+        """Set the object's key to a value represented by a string.
+        """
+        self[key] = self._parse(key, string)
+
 
 # Database controller and supporting interfaces.
 
diff --git a/beets/dbcore/query.py b/beets/dbcore/query.py
index f0adac665..d8a3a0ea8 100644
--- a/beets/dbcore/query.py
+++ b/beets/dbcore/query.py
@@ -653,6 +653,33 @@ class DateQuery(FieldQuery):
         return clause, subvals
 
 
+class DurationQuery(NumericQuery):
+    """NumericQuery that allow human-friendly (M:SS) time interval formats.
+
+    Converts the range(s) to a float value, and delegates on NumericQuery.
+
+    Raises InvalidQueryError when the pattern does not represent an int, float
+    or M:SS time interval.
+    """
+    def _convert(self, s):
+        """Convert a M:SS or numeric string to a float.
+
+        Return None if `s` is empty.
+        Raise an InvalidQueryError if the string cannot be converted.
+        """
+        if not s:
+            return None
+        try:
+            return util.raw_seconds_short(s)
+        except ValueError:
+            try:
+                return float(s)
+            except ValueError:
+                raise InvalidQueryArgumentTypeError(
+                    s,
+                    "a M:SS string or a float")
+
+
 # Sorting.
 
 class Sort(object):
diff --git a/beets/library.py b/beets/library.py
index 870c46856..f8d226dbe 100644
--- a/beets/library.py
+++ b/beets/library.py
@@ -195,6 +195,28 @@ class MusicalKey(types.String):
             return self.parse(key)
 
 
+class DurationType(types.Float):
+    """Human-friendly (M:SS) representation of a time interval."""
+    query = dbcore.query.DurationQuery
+
+    def format(self, value):
+        if not beets.config['format_raw_length'].get(bool):
+            return beets.ui.human_seconds_short(value or 0.0)
+        else:
+            return value
+
+    def parse(self, string):
+        try:
+            # Try to format back hh:ss to seconds.
+            return util.raw_seconds_short(string)
+        except ValueError:
+            # Fall back to a plain float.
+            try:
+                return float(string)
+            except ValueError:
+                return self.null
+
+
 # Library-specific sort types.
 
 class SmartArtistSort(dbcore.query.Sort):
@@ -426,7 +448,7 @@ class Item(LibModel):
         'original_day':         types.PaddedInt(2),
         'initial_key':          MusicalKey(),
 
-        'length':      types.FLOAT,
+        'length':      DurationType(),
         'bitrate':     types.ScaledInt(1000, u'kbps'),
         'format':      types.STRING,
         'samplerate':  types.ScaledInt(1000, u'kHz'),
diff --git a/beets/util/__init__.py b/beets/util/__init__.py
index e2d09c3ab..55c599a05 100644
--- a/beets/util/__init__.py
+++ b/beets/util/__init__.py
@@ -843,3 +843,16 @@ def case_sensitive(path):
     lower = _windows_long_path_name(path.lower())
     upper = _windows_long_path_name(path.upper())
     return lower != upper
+
+
+def raw_seconds_short(string):
+    """Formats a human-readable M:SS string as a float (number of seconds).
+
+    Raises ValueError if the conversion cannot take place due to `string` not
+    being in the right format.
+    """
+    match = re.match('^(\d+):([0-5]\d)$', string)
+    if not match:
+        raise ValueError('String not in M:SS format')
+    minutes, seconds = map(int, match.groups())
+    return float(minutes * 60 + seconds)
diff --git a/beetsplug/edit.py b/beetsplug/edit.py
new file mode 100644
index 000000000..8d6b168f5
--- /dev/null
+++ b/beetsplug/edit.py
@@ -0,0 +1,326 @@
+# This file is part of beets.
+# Copyright 2015
+#
+# Permission is hereby granted, free of charge, to any person obtaining
+# a copy of this software and associated documentation files (the
+# "Software"), to deal in the Software without restriction, including
+# without limitation the rights to use, copy, modify, merge, publish,
+# distribute, sublicense, and/or sell copies of the Software, and to
+# permit persons to whom the Software is furnished to do so, subject to
+# the following conditions:
+#
+# The above copyright notice and this permission notice shall be
+# included in all copies or substantial portions of the Software.
+
+"""Open metadata information in a text editor to let the user edit it.
+"""
+from __future__ import (division, absolute_import, print_function,
+                        unicode_literals)
+
+from beets import plugins
+from beets import util
+from beets import ui
+from beets.dbcore import types
+from beets.ui.commands import _do_query
+import subprocess
+import yaml
+from tempfile import NamedTemporaryFile
+import os
+
+
+# These "safe" types can avoid the format/parse cycle that most fields go
+# through: they are safe to edit with native YAML types.
+SAFE_TYPES = (types.Float, types.Integer, types.Boolean)
+
+
+class ParseError(Exception):
+    """The modified file is unreadable. The user should be offered a chance to
+    fix the error.
+    """
+
+
+def edit(filename):
+    """Open `filename` in a text editor.
+    """
+    cmd = util.shlex_split(util.editor_command())
+    cmd.append(filename)
+    subprocess.call(cmd)
+
+
+def dump(arg):
+    """Dump a sequence of dictionaries as YAML for editing.
+    """
+    return yaml.safe_dump_all(
+        arg,
+        allow_unicode=True,
+        default_flow_style=False,
+    )
+
+
+def load(s):
+    """Read a sequence of YAML documents back to a list of dictionaries
+    with string keys.
+
+    Can raise a `ParseError`.
+    """
+    try:
+        out = []
+        for d in yaml.load_all(s):
+            if not isinstance(d, dict):
+                raise ParseError(
+                    'each entry must be a dictionary; found {}'.format(
+                        type(d).__name__
+                    )
+                )
+
+            # Convert all keys to strings. They started out as strings,
+            # but the user may have inadvertently messed this up.
+            out.append({unicode(k): v for k, v in d.items()})
+
+    except yaml.YAMLError as e:
+        raise ParseError('invalid YAML: {}'.format(e))
+    return out
+
+
+def _safe_value(obj, key, value):
+    """Check whether the `value` is safe to represent in YAML and trust as
+    returned from parsed YAML.
+
+    This ensures that values do not change their type when the user edits their
+    YAML representation.
+    """
+    typ = obj._type(key)
+    return isinstance(typ, SAFE_TYPES) and isinstance(value, typ.model_type)
+
+
+def flatten(obj, fields):
+    """Represent `obj`, a `dbcore.Model` object, as a dictionary for
+    serialization. Only include the given `fields` if provided;
+    otherwise, include everything.
+
+    The resulting dictionary's keys are strings and the values are
+    safely YAML-serializable types.
+    """
+    # Format each value.
+    d = {}
+    for key in obj.keys():
+        value = obj[key]
+        if _safe_value(obj, key, value):
+            # A safe value that is faithfully representable in YAML.
+            d[key] = value
+        else:
+            # A value that should be edited as a string.
+            d[key] = obj.formatted()[key]
+
+    # Possibly filter field names.
+    if fields:
+        return {k: v for k, v in d.items() if k in fields}
+    else:
+        return d
+
+
+def apply(obj, data):
+    """Set the fields of a `dbcore.Model` object according to a
+    dictionary.
+
+    This is the opposite of `flatten`. The `data` dictionary should have
+    strings as values.
+    """
+    for key, value in data.items():
+        if _safe_value(obj, key, value):
+            # A safe value *stayed* represented as a safe type. Assign it
+            # directly.
+            obj[key] = value
+        else:
+            # Either the field was stringified originally or the user changed
+            # it from a safe type to an unsafe one. Parse it as a string.
+            obj.set_parse(key, unicode(value))
+
+
+class EditPlugin(plugins.BeetsPlugin):
+
+    def __init__(self):
+        super(EditPlugin, self).__init__()
+
+        self.config.add({
+            # The default fields to edit.
+            'albumfields': 'album albumartist',
+            'itemfields': 'track title artist album',
+
+            # Silently ignore any changes to these fields.
+            'ignore_fields': 'id path',
+        })
+
+    def commands(self):
+        edit_command = ui.Subcommand(
+            'edit',
+            help='interactively edit metadata'
+        )
+        edit_command.parser.add_option(
+            '-f', '--field',
+            metavar='FIELD',
+            action='append',
+            help='edit this field also',
+        )
+        edit_command.parser.add_option(
+            '--all',
+            action='store_true', dest='all',
+            help='edit all fields',
+        )
+        edit_command.parser.add_album_option()
+        edit_command.func = self._edit_command
+        return [edit_command]
+
+    def _edit_command(self, lib, opts, args):
+        """The CLI command function for the `beet edit` command.
+        """
+        # Get the objects to edit.
+        query = ui.decargs(args)
+        items, albums = _do_query(lib, query, opts.album, False)
+        objs = albums if opts.album else items
+        if not objs:
+            ui.print_('Nothing to edit.')
+            return
+
+        # Get the fields to edit.
+        if opts.all:
+            fields = None
+        else:
+            fields = self._get_fields(opts.album, opts.field)
+        self.edit(opts.album, objs, fields)
+
+    def _get_fields(self, album, extra):
+        """Get the set of fields to edit.
+        """
+        # Start with the configured base fields.
+        if album:
+            fields = self.config['albumfields'].as_str_seq()
+        else:
+            fields = self.config['itemfields'].as_str_seq()
+
+        # Add the requested extra fields.
+        if extra:
+            fields += extra
+
+        # Ensure we always have the `id` field for identification.
+        fields.append('id')
+
+        return set(fields)
+
+    def edit(self, album, objs, fields):
+        """The core editor function.
+
+        - `album`: A flag indicating whether we're editing Items or Albums.
+        - `objs`: The `Item`s or `Album`s to edit.
+        - `fields`: The set of field names to edit (or None to edit
+          everything).
+        """
+        # Present the YAML to the user and let her change it.
+        success = self.edit_objects(objs, fields)
+
+        # Save the new data.
+        if success:
+            self.save_write(objs)
+
+    def edit_objects(self, objs, fields):
+        """Dump a set of Model objects to a file as text, ask the user
+        to edit it, and apply any changes to the objects.
+
+        Return a boolean indicating whether the edit succeeded.
+        """
+        # Get the content to edit as raw data structures.
+        old_data = [flatten(o, fields) for o in objs]
+
+        # Set up a temporary file with the initial data for editing.
+        new = NamedTemporaryFile(suffix='.yaml', delete=False)
+        old_str = dump(old_data)
+        new.write(old_str)
+        new.close()
+
+        # Loop until we have parseable data and the user confirms.
+        try:
+            while True:
+                # Ask the user to edit the data.
+                edit(new.name)
+
+                # Read the data back after editing and check whether anything
+                # changed.
+                with open(new.name) as f:
+                    new_str = f.read()
+                if new_str == old_str:
+                    ui.print_("No changes; aborting.")
+                    return False
+
+                # Parse the updated data.
+                try:
+                    new_data = load(new_str)
+                except ParseError as e:
+                    ui.print_("Could not read data: {}".format(e))
+                    if ui.input_yn("Edit again to fix? (Y/n)", True):
+                        continue
+                    else:
+                        return False
+
+                # Show the changes.
+                self.apply_data(objs, old_data, new_data)
+                changed = False
+                for obj in objs:
+                    changed |= ui.show_model_changes(obj)
+                if not changed:
+                    ui.print_('No changes to apply.')
+                    return False
+
+                # Confirm the changes.
+                choice = ui.input_options(
+                    ('continue Editing', 'apply', 'cancel')
+                )
+                if choice == 'a':  # Apply.
+                    return True
+                elif choice == 'c':  # Cancel.
+                    return False
+                elif choice == 'e':  # Keep editing.
+                    # Reset the temporary changes to the objects.
+                    for obj in objs:
+                        obj.read()
+                    continue
+
+        # Remove the temporary file before returning.
+        finally:
+            os.remove(new.name)
+
+    def apply_data(self, objs, old_data, new_data):
+        """Take potentially-updated data and apply it to a set of Model
+        objects.
+
+        The objects are not written back to the database, so the changes
+        are temporary.
+        """
+        if len(old_data) != len(new_data):
+            self._log.warn('number of objects changed from {} to {}',
+                           len(old_data), len(new_data))
+
+        obj_by_id = {o.id: o for o in objs}
+        ignore_fields = self.config['ignore_fields'].as_str_seq()
+        for old_dict, new_dict in zip(old_data, new_data):
+            # Prohibit any changes to forbidden fields to avoid
+            # clobbering `id` and such by mistake.
+            forbidden = False
+            for key in ignore_fields:
+                if old_dict.get(key) != new_dict.get(key):
+                    self._log.warn('ignoring object whose {} changed', key)
+                    forbidden = True
+                    break
+            if forbidden:
+                continue
+
+            id = int(old_dict['id'])
+            apply(obj_by_id[id], new_dict)
+
+    def save_write(self, objs):
+        """Save a list of updated Model objects to the database.
+        """
+        # Save to the database and possibly write tags.
+        for ob in objs:
+            if ob._dirty:
+                self._log.debug('saving changes to {}', ob)
+                ob.try_sync(ui.should_write())
diff --git a/beetsplug/fetchart.py b/beetsplug/fetchart.py
index 1b2089b58..a56f9f95a 100644
--- a/beetsplug/fetchart.py
+++ b/beetsplug/fetchart.py
@@ -157,34 +157,6 @@ class AlbumArtOrg(ArtSource):
             self._log.debug(u'no image found on page')
 
 
-class GoogleImages(ArtSource):
-    URL = 'https://ajax.googleapis.com/ajax/services/search/images'
-
-    def get(self, album):
-        """Return art URL from google.org given an album title and
-        interpreter.
-        """
-        if not (album.albumartist and album.album):
-            return
-        search_string = (album.albumartist + ',' + album.album).encode('utf-8')
-        response = self.request(self.URL, params={
-            'v': '1.0',
-            'q': search_string,
-            'start': '0',
-        })
-
-        # Get results using JSON.
-        try:
-            results = response.json()
-            data = results['responseData']
-            dataInfo = data['results']
-            for myUrl in dataInfo:
-                yield myUrl['unescapedUrl']
-        except:
-            self._log.debug(u'error scraping art page')
-            return
-
-
 class ITunesStore(ArtSource):
     # Art from the iTunes Store.
     def get(self, album):
@@ -196,11 +168,19 @@ class ITunesStore(ArtSource):
         try:
             # Isolate bugs in the iTunes library while searching.
             try:
-                itunes_album = itunes.search_album(search_string)[0]
+                results = itunes.search_album(search_string)
             except Exception as exc:
                 self._log.debug('iTunes search failed: {0}', exc)
                 return
 
+            # Get the first match.
+            if results:
+                itunes_album = results[0]
+            else:
+                self._log.debug('iTunes search for {:r} got no results',
+                                search_string)
+                return
+
             if itunes_album.get_artwork()['100']:
                 small_url = itunes_album.get_artwork()['100']
                 big_url = small_url.replace('100x100', '1200x1200')
@@ -380,7 +360,7 @@ class FileSystem(ArtSource):
 
 # Try each source in turn.
 
-SOURCES_ALL = [u'coverart', u'itunes', u'amazon', u'albumart', u'google',
+SOURCES_ALL = [u'coverart', u'itunes', u'amazon', u'albumart',
                u'wikipedia']
 
 ART_SOURCES = {
@@ -388,7 +368,6 @@ ART_SOURCES = {
     u'itunes': ITunesStore,
     u'albumart': AlbumArtOrg,
     u'amazon': Amazon,
-    u'google': GoogleImages,
     u'wikipedia': Wikipedia,
 }
 
diff --git a/beetsplug/lyrics.py b/beetsplug/lyrics.py
index 0bd49589c..1af34df99 100644
--- a/beetsplug/lyrics.py
+++ b/beetsplug/lyrics.py
@@ -454,28 +454,32 @@ class Google(Backend):
     BY_TRANS = ['by', 'par', 'de', 'von']
     LYRICS_TRANS = ['lyrics', 'paroles', 'letras', 'liedtexte']
 
-    def is_page_candidate(self, urlLink, urlTitle, title, artist):
+    def is_page_candidate(self, url_link, url_title, title, artist):
         """Return True if the URL title makes it a good candidate to be a
         page that contains lyrics of title by artist.
         """
         title = self.slugify(title.lower())
         artist = self.slugify(artist.lower())
         sitename = re.search(u"//([^/]+)/.*",
-                             self.slugify(urlLink.lower())).group(1)
-        urlTitle = self.slugify(urlTitle.lower())
+                             self.slugify(url_link.lower())).group(1)
+        url_title = self.slugify(url_title.lower())
+
         # Check if URL title contains song title (exact match)
-        if urlTitle.find(title) != -1:
+        if url_title.find(title) != -1:
             return True
+
         # or try extracting song title from URL title and check if
         # they are close enough
         tokens = [by + '_' + artist for by in self.BY_TRANS] + \
                  [artist, sitename, sitename.replace('www.', '')] + \
             self.LYRICS_TRANS
-        songTitle = re.sub(u'(%s)' % u'|'.join(tokens), u'', urlTitle)
-        songTitle = songTitle.strip('_|')
-        typoRatio = .9
-        ratio = difflib.SequenceMatcher(None, songTitle, title).ratio()
-        return ratio >= typoRatio
+        tokens = [re.escape(t) for t in tokens]
+        song_title = re.sub(u'(%s)' % u'|'.join(tokens), u'', url_title)
+
+        song_title = song_title.strip('_|')
+        typo_ratio = .9
+        ratio = difflib.SequenceMatcher(None, song_title, title).ratio()
+        return ratio >= typo_ratio
 
     def fetch(self, artist, title):
         query = u"%s %s" % (artist, title)
@@ -492,12 +496,12 @@ class Google(Backend):
 
         if 'items' in data.keys():
             for item in data['items']:
-                urlLink = item['link']
-                urlTitle = item.get('title', u'')
-                if not self.is_page_candidate(urlLink, urlTitle,
+                url_link = item['link']
+                url_title = item.get('title', u'')
+                if not self.is_page_candidate(url_link, url_title,
                                               title, artist):
                     continue
-                html = self.fetch_url(urlLink)
+                html = self.fetch_url(url_link)
                 lyrics = scrape_lyrics_from_html(html)
                 if not lyrics:
                     continue
diff --git a/beetsplug/scrub.py b/beetsplug/scrub.py
index 4636b477d..4e37ad7ff 100644
--- a/beetsplug/scrub.py
+++ b/beetsplug/scrub.py
@@ -45,9 +45,6 @@ _MUTAGEN_FORMATS = {
 }
 
 
-scrubbing = False
-
-
 class ScrubPlugin(BeetsPlugin):
     """Removes extraneous metadata from files' tags."""
     def __init__(self):
@@ -55,44 +52,17 @@ class ScrubPlugin(BeetsPlugin):
         self.config.add({
             'auto': True,
         })
-        self.register_listener("write", self.write_item)
+
+        if self.config['auto']:
+            self.register_listener("import_task_files", self.import_task_files)
 
     def commands(self):
         def scrub_func(lib, opts, args):
-            # This is a little bit hacky, but we set a global flag to
-            # avoid autoscrubbing when we're also explicitly scrubbing.
-            global scrubbing
-            scrubbing = True
-
             # Walk through matching files and remove tags.
             for item in lib.items(ui.decargs(args)):
                 self._log.info(u'scrubbing: {0}',
                                util.displayable_path(item.path))
-
-                # Get album art if we need to restore it.
-                if opts.write:
-                    try:
-                        mf = mediafile.MediaFile(util.syspath(item.path),
-                                                 config['id3v23'].get(bool))
-                    except IOError as exc:
-                        self._log.error(u'could not open file to scrub: {0}',
-                                        exc)
-                    art = mf.art
-
-                # Remove all tags.
-                self._scrub(item.path)
-
-                # Restore tags, if enabled.
-                if opts.write:
-                    self._log.debug(u'writing new tags after scrub')
-                    item.try_write()
-                    if art:
-                        self._log.info(u'restoring art')
-                        mf = mediafile.MediaFile(util.syspath(item.path))
-                        mf.art = art
-                        mf.save()
-
-            scrubbing = False
+                self._scrub_item(item, opts.write)
 
         scrub_cmd = ui.Subcommand('scrub', help='clean audio tags')
         scrub_cmd.parser.add_option('-W', '--nowrite', dest='write',
@@ -140,8 +110,36 @@ class ScrubPlugin(BeetsPlugin):
                 self._log.error(u'could not scrub {0}: {1}',
                                 util.displayable_path(path), exc)
 
-    def write_item(self, item, path, tags):
-        """Automatically embed art into imported albums."""
-        if not scrubbing and self.config['auto']:
-            self._log.debug(u'auto-scrubbing {0}', util.displayable_path(path))
-            self._scrub(path)
+    def _scrub_item(self, item, restore=True):
+        """Remove tags from an Item's associated file and, if `restore`
+        is enabled, write the database's tags back to the file.
+        """
+        # Get album art if we need to restore it.
+        if restore:
+            try:
+                mf = mediafile.MediaFile(util.syspath(item.path),
+                                         config['id3v23'].get(bool))
+            except IOError as exc:
+                self._log.error(u'could not open file to scrub: {0}',
+                                exc)
+            art = mf.art
+
+        # Remove all tags.
+        self._scrub(item.path)
+
+        # Restore tags, if enabled.
+        if restore:
+            self._log.debug(u'writing new tags after scrub')
+            item.try_write()
+            if art:
+                self._log.debug(u'restoring art')
+                mf = mediafile.MediaFile(util.syspath(item.path))
+                mf.art = art
+                mf.save()
+
+    def import_task_files(self, session, task):
+        """Automatically scrub imported files."""
+        for item in task.imported_items():
+            self._log.debug(u'auto-scrubbing {0}',
+                            util.displayable_path(item.path))
+            self._scrub_item(item)
diff --git a/docs/changelog.rst b/docs/changelog.rst
index eceda093c..7f07bbb1c 100644
--- a/docs/changelog.rst
+++ b/docs/changelog.rst
@@ -3,6 +3,10 @@ Changelog
 
 1.3.16 (in development)
 -----------------------
+* A new plugin edit helps you manually edit fields from items.
+  You search for items in the normal beets way.Then edit opens a texteditor
+  with the items and the fields of the items you want to edit. Afterwards you can
+  review your changes save them back into the items.
 
 New:
 
@@ -26,6 +30,11 @@ New:
   singles compilation, "1." See :ref:`not_query`. :bug:`819` :bug:`1728`
 * :doc:`/plugins/info`: The plugin now accepts the ``-f/--format`` option for
   customizing how items are displayed. :bug:`1737`
+* Track length is now displayed as ``M:SS`` by default, instead of displaying
+  the raw number of seconds. Queries on track length also accept this format:
+  for example, ``beet list length:5:30..`` will find all your tracks that have
+  a duration over 5 minutes and 30 seconds. You can turn off this new behavior
+  using the ``format_raw_length`` configuration option. :bug:`1749`
 
 For developers:
 
@@ -73,6 +82,16 @@ Fixes:
   ImageMagick on Windows. :bug:`1721`
 * Fix a crash when writing some Unicode comment strings to MP3s that used
   older encodings. The encoding is now always updated to UTF-8. :bug:`879`
+* :doc:`/plugins/fetchart`: The Google Images backend has been removed. It
+  used an API that has been shut down. :bug:`1760`
+* :doc:`/plugins/lyrics`: Fix a crash in the Google backend when searching for
+  bands with regular-expression characters in their names, like Sunn O))).
+  :bug:`1673`
+* :doc:`/plugins/scrub`: In ``auto`` mode, the plugin now *actually* only
+  scrubs files on import---not every time files were written, as it previously
+  did. :bug:`1657`
+* :doc:`/plugins/scrub`: Also in ``auto`` mode, album art is now correctly
+  restored. :bug:`1657`
 
 .. _Emby Server: http://emby.media
 
diff --git a/docs/plugins/edit.rst b/docs/plugins/edit.rst
new file mode 100644
index 000000000..507d56950
--- /dev/null
+++ b/docs/plugins/edit.rst
@@ -0,0 +1,36 @@
+Edit Plugin
+===========
+
+The ``edit`` plugin lets you modify music metadata using your favorite text
+editor.
+
+Enable the ``edit`` plugin in your configuration (see :ref:`using-plugins`) and
+then type::
+
+     beet edit QUERY
+
+Your text editor (i.e., the command in your ``$EDITOR`` environment variable)
+will open with a list of tracks to edit. Make your changes and exit your text
+editor to apply them to your music.
+
+Command-Line Options
+--------------------
+
+The ``edit`` command has these command-line options:
+
+- ``-a`` or ``--album``: Edit albums instead of individual items.
+- ``-f FIELD`` or ``--field FIELD``: Specify an additional field to edit
+  (in addition to the defaults set in the configuration).
+- ``--all``: Edit *all* available fields.
+
+Configuration
+-------------
+
+To configure the plugin, make an ``edit:`` section in your configuration
+file. The available options are:
+
+- **itemfields**: A space-separated list of item fields to include in the
+  editor by default.
+  Default: ``track title artist album``
+- **albumfields**: The same when editing albums (with the ``-a`` option).
+  Default: ``album albumartist``
diff --git a/docs/plugins/fetchart.rst b/docs/plugins/fetchart.rst
index c04517b11..80149d478 100644
--- a/docs/plugins/fetchart.rst
+++ b/docs/plugins/fetchart.rst
@@ -50,7 +50,7 @@ file. The available options are:
 - **sources**: List of sources to search for images. An asterisk `*` expands
   to all available sources.
   Default: ``coverart itunes amazon albumart``, i.e., everything but
-  ``wikipedia`` and ``google``. Enable those two sources for more matches at
+  ``wikipedia``. Enable those two sources for more matches at
   the cost of some speed.
 
 Note: ``minwidth`` and ``enforce_ratio`` options require either `ImageMagick`_
@@ -94,7 +94,7 @@ no resizing is performed for album art found on the filesystem---only downloaded
 art is resized. Server-side resizing can also be slower than local resizing, so
 consider installing one of the two backends for better performance.
 
-When using ImageMagic, beets looks for the ``convert`` executable in your path.
+When using ImageMagick, beets looks for the ``convert`` executable in your path.
 On some versions of Windows, the program can be shadowed by a system-provided
 ``convert.exe``. On these systems, you may need to modify your ``%PATH%``
 environment variable so that ImageMagick comes first or use Pillow instead.
@@ -106,8 +106,9 @@ Album Art Sources
 -----------------
 
 By default, this plugin searches for art in the local filesystem as well as on
-the Cover Art Archive, the iTunes Store, Amazon, AlbumArt.org,
-and Google Image Search, and Wikipedia, in that order. You can reorder the sources or remove
+the Cover Art Archive, the iTunes Store, Amazon, and AlbumArt.org, in that
+order.
+You can reorder the sources or remove
 some to speed up the process using the ``sources`` configuration option.
 
 When looking for local album art, beets checks for image files located in the
@@ -126,23 +127,16 @@ iTunes Store
 To use the iTunes Store as an art source, install the `python-itunes`_
 library. You can do this using `pip`_, like so::
 
-    $ pip install python-itunes
+    $ pip install https://github.com/ocelma/python-itunes/archive/master.zip
 
+(There's currently `a problem`_ that prevents a plain ``pip install
+python-itunes`` from working.)
 Once the library is installed, the plugin will use it to search automatically.
 
+.. _a problem: https://github.com/ocelma/python-itunes/issues/9
 .. _python-itunes: https://github.com/ocelma/python-itunes
 .. _pip: http://pip.openplans.org/
 
-Google Image Search
-'''''''''''''''''''
-
-You can optionally search for cover art on `Google Images`_. This option uses
-the first hit for a search query consisting of the artist and album name. It
-is therefore approximate: "incorrect" image matches are possible (although
-unlikely).
-
-.. _Google Images: http://images.google.com/
-
 
 Embedding Album Art
 -------------------
diff --git a/docs/plugins/index.rst b/docs/plugins/index.rst
index d09139837..5da5ee0be 100644
--- a/docs/plugins/index.rst
+++ b/docs/plugins/index.rst
@@ -40,6 +40,7 @@ Each plugin has its own set of options that can be defined in a section bearing
    discogs
    duplicates
    echonest
+   edit
    embedart
    embyupdate
    fetchart
@@ -96,6 +97,7 @@ Metadata
 * :doc:`bpm`: Measure tempo using keystrokes.
 * :doc:`echonest`: Automatically fetch `acoustic attributes`_ from
   `the Echo Nest`_ (tempo, energy, danceability, ...).
+* :doc:`edit`: Edit metadata from a texteditor.
 * :doc:`embedart`: Embed album art images into files' metadata.
 * :doc:`fetchart`: Fetch album cover art from various sources.
 * :doc:`ftintitle`: Move "featured" artists from the artist field to the title
diff --git a/docs/reference/query.rst b/docs/reference/query.rst
index 1349d755e..82a16869d 100644
--- a/docs/reference/query.rst
+++ b/docs/reference/query.rst
@@ -144,6 +144,11 @@ and this command finds MP3 files with bitrates of 128k or lower::
 
     $ beet list format:MP3 bitrate:..128000
 
+The ``length`` field also lets you use a "M:SS" format. For example, this
+query finds tracks that are less than four and a half minutes in length::
+
+    $ beet list length:..4:30
+
 
 .. _datequery:
 
diff --git a/test/rsrc/unicode’d.mp3 b/test/rsrc/unicode’d.mp3
index ef732eba6..a14181941 100644
Binary files a/test/rsrc/unicode’d.mp3 and b/test/rsrc/unicode’d.mp3 differ
diff --git a/test/test_art.py b/test/test_art.py
index 04bfe3eed..cb29f3769 100644
--- a/test/test_art.py
+++ b/test/test_art.py
@@ -226,38 +226,6 @@ class AAOTest(_common.TestCase):
         self.assertEqual(list(res), [])
 
 
-class GoogleImageTest(_common.TestCase):
-
-    _google_url = 'https://ajax.googleapis.com/ajax/services/search/images'
-
-    def setUp(self):
-        super(GoogleImageTest, self).setUp()
-        self.source = fetchart.GoogleImages(logger)
-
-    @responses.activate
-    def run(self, *args, **kwargs):
-        super(GoogleImageTest, self).run(*args, **kwargs)
-
-    def mock_response(self, url, json):
-        responses.add(responses.GET, url, body=json,
-                      content_type='application/json')
-
-    def test_google_art_finds_image(self):
-        album = _common.Bag(albumartist="some artist", album="some album")
-        json = b"""{"responseData": {"results":
-            [{"unescapedUrl": "url_to_the_image"}]}}"""
-        self.mock_response(self._google_url, json)
-        result_url = self.source.get(album)
-        self.assertEqual(list(result_url)[0], 'url_to_the_image')
-
-    def test_google_art_dont_finds_image(self):
-        album = _common.Bag(albumartist="some artist", album="some album")
-        json = b"""bla blup"""
-        self.mock_response(self._google_url, json)
-        result_url = self.source.get(album)
-        self.assertEqual(list(result_url), [])
-
-
 class ArtImporterTest(UseThePlugin):
     def setUp(self):
         super(ArtImporterTest, self).setUp()
diff --git a/test/test_edit.py b/test/test_edit.py
new file mode 100644
index 000000000..ae0500029
--- /dev/null
+++ b/test/test_edit.py
@@ -0,0 +1,260 @@
+# This file is part of beets.
+# Copyright 2015, Adrian Sampson and Diego Moreda.
+#
+# Permission is hereby granted, free of charge, to any person obtaining
+# a copy of this software and associated documentation files (the
+# "Software"), to deal in the Software without restriction, including
+# without limitation the rights to use, copy, modify, merge, publish,
+# distribute, sublicense, and/or sell copies of the Software, and to
+# permit persons to whom the Software is furnished to do so, subject to
+# the following conditions:
+#
+# The above copyright notice and this permission notice shall be
+# included in all copies or substantial portions of the Software.
+
+from __future__ import (division, absolute_import, print_function,
+                        unicode_literals)
+import codecs
+
+from mock import patch
+from test._common import unittest
+from test.helper import TestHelper, control_stdin
+
+
+class ModifyFileMocker(object):
+    """Helper for modifying a file, replacing or editing its contents. Used for
+    mocking the calls to the external editor during testing."""
+
+    def __init__(self, contents=None, replacements=None):
+        """ `self.contents` and `self.replacements` are initalized here, in
+        order to keep the rest of the functions of this class with the same
+        signature as `EditPlugin.get_editor()`, making mocking easier.
+            - `contents`: string with the contents of the file to be used for
+            `overwrite_contents()`
+            - `replacement`: dict with the in-place replacements to be used for
+            `replace_contents()`, in the form {'previous string': 'new string'}
+
+        TODO: check if it can be solved more elegantly with a decorator
+        """
+        self.contents = contents
+        self.replacements = replacements
+        self.action = self.overwrite_contents
+        if replacements:
+            self.action = self.replace_contents
+
+    def overwrite_contents(self, filename):
+        """Modify `filename`, replacing its contents with `self.contents`. If
+        `self.contents` is empty, the file remains unchanged.
+        """
+        if self.contents:
+            with codecs.open(filename, 'w', encoding='utf8') as f:
+                f.write(self.contents)
+
+    def replace_contents(self, filename):
+        """Modify `filename`, reading its contents and replacing the strings
+        specified in `self.replacements`.
+        """
+        with codecs.open(filename, 'r', encoding='utf8') as f:
+            contents = f.read()
+        for old, new_ in self.replacements.iteritems():
+            contents = contents.replace(old, new_)
+        with codecs.open(filename, 'w', encoding='utf8') as f:
+            f.write(contents)
+
+
+class EditCommandTest(unittest.TestCase, TestHelper):
+    """ Black box tests for `beetsplug.edit`. Command line interaction is
+    simulated using `test.helper.control_stdin()`, and yaml editing via an
+    external editor is simulated using `ModifyFileMocker`.
+    """
+    ALBUM_COUNT = 1
+    TRACK_COUNT = 10
+
+    def setUp(self):
+        self.setup_beets()
+        self.load_plugins('edit')
+        # make sure that we avoid invoking the editor except for making changes
+        self.config['edit']['diff_method'] = ''
+        # add an album, storing the original fields for comparison
+        self.album = self.add_album_fixture(track_count=self.TRACK_COUNT)
+        self.album_orig = {f: self.album[f] for f in self.album._fields}
+        self.items_orig = [{f: item[f] for f in item._fields} for
+                           item in self.album.items()]
+
+        # keep track of write()s
+        self.write_patcher = patch('beets.library.Item.write')
+        self.mock_write = self.write_patcher.start()
+
+    def tearDown(self):
+        self.write_patcher.stop()
+        self.teardown_beets()
+        self.unload_plugins()
+
+    def run_mocked_command(self, modify_file_args={}, stdin=[], args=[]):
+        """Run the edit command, with mocked stdin and yaml writing, and
+        passing `args` to `run_command`."""
+        m = ModifyFileMocker(**modify_file_args)
+        with patch('beetsplug.edit.edit', side_effect=m.action):
+            with control_stdin('\n'.join(stdin)):
+                self.run_command('edit', *args)
+
+    def assertCounts(self, album_count=ALBUM_COUNT, track_count=TRACK_COUNT,
+                     write_call_count=TRACK_COUNT, title_starts_with=''):
+        """Several common assertions on Album, Track and call counts."""
+        self.assertEqual(len(self.lib.albums()), album_count)
+        self.assertEqual(len(self.lib.items()), track_count)
+        self.assertEqual(self.mock_write.call_count, write_call_count)
+        self.assertTrue(all(i.title.startswith(title_starts_with)
+                            for i in self.lib.items()))
+
+    def assertItemFieldsModified(self, library_items, items, fields=[]):
+        """Assert that items in the library (`lib_items`) have different values
+        on the specified `fields` (and *only* on those fields), compared to
+        `items`.
+        An empty `fields` list results in asserting that no modifications have
+        been performed.
+        """
+        changed_fields = []
+        for lib_item, item in zip(library_items, items):
+            changed_fields.append([field for field in lib_item._fields
+                                   if lib_item[field] != item[field]])
+        self.assertTrue(all(diff_fields == fields for diff_fields in
+                            changed_fields))
+
+    def test_title_edit_discard(self):
+        """Edit title for all items in the library, then discard changes-"""
+        # edit titles
+        self.run_mocked_command({'replacements': {u't\u00eftle':
+                                                  u'modified t\u00eftle'}},
+                                # Cancel.
+                                ['c'])
+
+        self.assertCounts(write_call_count=0,
+                          title_starts_with=u't\u00eftle')
+        self.assertItemFieldsModified(self.album.items(), self.items_orig, [])
+
+    def test_title_edit_apply(self):
+        """Edit title for all items in the library, then apply changes."""
+        # edit titles
+        self.run_mocked_command({'replacements': {u't\u00eftle':
+                                                  u'modified t\u00eftle'}},
+                                # Apply changes.
+                                ['a'])
+
+        self.assertCounts(write_call_count=self.TRACK_COUNT,
+                          title_starts_with=u'modified t\u00eftle')
+        self.assertItemFieldsModified(self.album.items(), self.items_orig,
+                                      ['title'])
+
+    def test_single_title_edit_apply(self):
+        """Edit title for one item in the library, then apply changes."""
+        # edit title
+        self.run_mocked_command({'replacements': {u't\u00eftle 9':
+                                                  u'modified t\u00eftle 9'}},
+                                # Apply changes.
+                                ['a'])
+
+        self.assertCounts(write_call_count=1,)
+        # no changes except on last item
+        self.assertItemFieldsModified(list(self.album.items())[:-1],
+                                      self.items_orig[:-1], [])
+        self.assertEqual(list(self.album.items())[-1].title,
+                         u'modified t\u00eftle 9')
+
+    def test_noedit(self):
+        """Do not edit anything."""
+        # do not edit anything
+        self.run_mocked_command({'contents': None},
+                                # no stdin
+                                [])
+
+        self.assertCounts(write_call_count=0,
+                          title_starts_with=u't\u00eftle')
+        self.assertItemFieldsModified(self.album.items(), self.items_orig, [])
+
+    def test_album_edit_apply(self):
+        """Edit the album field for all items in the library, apply changes.
+        By design, the album should not be updated.""
+        """
+        # edit album
+        self.run_mocked_command({'replacements': {u'\u00e4lbum':
+                                                  u'modified \u00e4lbum'}},
+                                # Apply changes.
+                                ['a'])
+
+        self.assertCounts(write_call_count=self.TRACK_COUNT)
+        self.assertItemFieldsModified(self.album.items(), self.items_orig,
+                                      ['album'])
+        # ensure album is *not* modified
+        self.album.load()
+        self.assertEqual(self.album.album, u'\u00e4lbum')
+
+    def test_single_edit_add_field(self):
+        """Edit the yaml file appending an extra field to the first item, then
+        apply changes."""
+        # append "foo: bar" to item with id == 1
+        self.run_mocked_command({'replacements': {u"id: 1":
+                                                  u"id: 1\nfoo: bar"}},
+                                # Apply changes.
+                                ['a'])
+
+        self.assertEqual(self.lib.items('id:1')[0].foo, 'bar')
+        self.assertCounts(write_call_count=1,
+                          title_starts_with=u't\u00eftle')
+
+    def test_a_album_edit_apply(self):
+        """Album query (-a), edit album field, apply changes."""
+        self.run_mocked_command({'replacements': {u'\u00e4lbum':
+                                                  u'modified \u00e4lbum'}},
+                                # Apply changes.
+                                ['a'],
+                                args=['-a'])
+
+        self.album.load()
+        self.assertCounts(write_call_count=self.TRACK_COUNT)
+        self.assertEqual(self.album.album, u'modified \u00e4lbum')
+        self.assertItemFieldsModified(self.album.items(), self.items_orig,
+                                      ['album'])
+
+    def test_a_albumartist_edit_apply(self):
+        """Album query (-a), edit albumartist field, apply changes."""
+        self.run_mocked_command({'replacements': {u'album artist':
+                                                  u'modified album artist'}},
+                                # Apply changes.
+                                ['a'],
+                                args=['-a'])
+
+        self.album.load()
+        self.assertCounts(write_call_count=self.TRACK_COUNT)
+        self.assertEqual(self.album.albumartist, u'the modified album artist')
+        self.assertItemFieldsModified(self.album.items(), self.items_orig,
+                                      ['albumartist'])
+
+    def test_malformed_yaml(self):
+        """Edit the yaml file incorrectly (resulting in a malformed yaml
+        document)."""
+        # edit the yaml file to an invalid file
+        self.run_mocked_command({'contents': '!MALFORMED'},
+                                # Edit again to fix? No.
+                                ['n'])
+
+        self.assertCounts(write_call_count=0,
+                          title_starts_with=u't\u00eftle')
+
+    def test_invalid_yaml(self):
+        """Edit the yaml file incorrectly (resulting in a well-formed but
+        invalid yaml document)."""
+        # edit the yaml file to an invalid file
+        self.run_mocked_command({'contents': 'wellformed: yes, but invalid'},
+                                # no stdin
+                                [])
+
+        self.assertCounts(write_call_count=0,
+                          title_starts_with=u't\u00eftle')
+
+
+def suite():
+    return unittest.TestLoader().loadTestsFromName(__name__)
+
+if __name__ == b'__main__':
+    unittest.main(defaultTest='suite')
diff --git a/test/test_info.py b/test/test_info.py
index aaabed980..4a85b6dc9 100644
--- a/test/test_info.py
+++ b/test/test_info.py
@@ -111,7 +111,7 @@ class InfoTest(unittest.TestCase, TestHelper):
         self.add_item_fixtures()
         out = self.run_with_output('--library', '--format',
                                    '$track. $title - $artist ($length)')
-        self.assertEqual(u'02. tïtle 0 - the artist (1.1)\n', out)
+        self.assertEqual(u'02. tïtle 0 - the artist (0:01)\n', out)
 
 
 def suite():
diff --git a/test/test_library.py b/test/test_library.py
index c544e007e..9aeaad719 100644
--- a/test/test_library.py
+++ b/test/test_library.py
@@ -25,6 +25,7 @@ import shutil
 import re
 import unicodedata
 import sys
+import time
 
 from test import _common
 from test._common import unittest
@@ -1127,6 +1128,58 @@ class ParseQueryTest(unittest.TestCase):
             beets.library.parse_query_string(b"query", None)
 
 
+class LibraryFieldTypesTest(unittest.TestCase):
+    """Test format() and parse() for library-specific field types"""
+    def test_datetype(self):
+        t = beets.library.DateType()
+
+        # format
+        time_local = time.strftime(beets.config['time_format'].get(unicode),
+                                   time.localtime(123456789))
+        self.assertEqual(time_local, t.format(123456789))
+        # parse
+        self.assertEqual(123456789.0, t.parse(time_local))
+        self.assertEqual(123456789.0, t.parse('123456789.0'))
+        self.assertEqual(t.null, t.parse('not123456789.0'))
+        self.assertEqual(t.null, t.parse('1973-11-29'))
+
+    def test_pathtype(self):
+        t = beets.library.PathType()
+
+        # format
+        self.assertEqual('/tmp', t.format('/tmp'))
+        self.assertEqual(u'/tmp/\xe4lbum', t.format(u'/tmp/\u00e4lbum'))
+        # parse
+        self.assertEqual(b'/tmp', t.parse('/tmp'))
+        self.assertEqual(b'/tmp/\xc3\xa4lbum', t.parse(u'/tmp/\u00e4lbum/'))
+
+    def test_musicalkey(self):
+        t = beets.library.MusicalKey()
+
+        # parse
+        self.assertEqual('C#m', t.parse('c#m'))
+        self.assertEqual('Gm', t.parse('g   minor'))
+        self.assertEqual('Not c#m', t.parse('not C#m'))
+
+    def test_durationtype(self):
+        t = beets.library.DurationType()
+
+        # format
+        self.assertEqual('1:01', t.format(61.23))
+        self.assertEqual('60:01', t.format(3601.23))
+        self.assertEqual('0:00', t.format(None))
+        # parse
+        self.assertEqual(61.0, t.parse('1:01'))
+        self.assertEqual(61.23, t.parse('61.23'))
+        self.assertEqual(3601.0, t.parse('60:01'))
+        self.assertEqual(t.null, t.parse('1:00:01'))
+        self.assertEqual(t.null, t.parse('not61.23'))
+        # config format_raw_length
+        beets.config['format_raw_length'] = True
+        self.assertEqual(61.23, t.format(61.23))
+        self.assertEqual(3601.23, t.format(3601.23))
+
+
 def suite():
     return unittest.TestLoader().loadTestsFromName(__name__)
 
diff --git a/test/test_lyrics.py b/test/test_lyrics.py
index 33b8c6bb5..515e96587 100644
--- a/test/test_lyrics.py
+++ b/test/test_lyrics.py
@@ -376,6 +376,17 @@ class LyricsGooglePluginTest(unittest.TestCase):
         self.assertEqual(google.is_page_candidate(url, urlTitle, s['title'],
                          s['artist']), False, url)
 
+    def test_is_page_candidate_special_chars(self):
+        """Ensure that `is_page_candidate` doesn't crash when the artist
+        and such contain special regular expression characters.
+        """
+        # https://github.com/sampsyo/beets/issues/1673
+        s = self.source
+        url = s['url'] + s['path']
+        url_title = u'foo'
+
+        google.is_page_candidate(url, url_title, s['title'], 'Sunn O)))')
+
 
 def suite():
     return unittest.TestLoader().loadTestsFromName(__name__)