This bug was fixed in the package ubiquity - 2.11.0 --------------- ubiquity (2.11.0) quantal; urgency=low [ Colin Watson ] * Install oem-config-slideshow-ubuntu in a separate pass from oem-config-$frontend and ubiquity-frontend-$frontend, since it may be missing from images; and only do this for the GTK frontend in any case, since other frontends don't currently use the slideshow (LP: #987050). * Re-enable alpha warning for quantal. * Upgrade to debhelper 9 for improved handling of compiler flags. Make sure that /usr/lib/girepository-1.0/ stays where it's supposed to be despite changes in the default libdir for multiarch. * Port to Python 3: - Use Python 3-style print functions. - Use "except Exception as e" syntax rather than the old-style "except Exception, e". - Use reduce from functools rather than relying on the builtin. - Use list comprehensions rather than filter or map. - Use open() rather than file(). - Import configparser rather than ConfigParser if available. - Use input() rather than raw_input() when running under Python 3. - Use set comprehensions. - Import quote from urllib.parse rather than urllib if available. - Use new-style octal literals. - Use test.support rather than test.test_support if available. - Add --python2 and --python3 options to tests/run to force running the tests under Python 2 or 3 respectively. - Handle renaming of __builtin__ to builtins in Python 3. - Only pass unicode=1 to gettext.install in Python 2. - Port ubiquity.auto_update to python-apt 0.8 progress classes. - Use Python 3 names for itertools.izip and itertools.izip_longest if available. - Use helpers from the six module to deal with some bytes/unicode differences. - As a general rule, open subprocesses with universal_newlines=True when expecting to read text from them. This has no effect on Python 2 (aside from \r\n conversion and the like, which is mostly a no-op for us), but causes Python 3 to read str rather than bytes. The exceptions at the moment are: debconf-copydb subprocesses, which return mixed-encoding data that needs to be handled specially; and when feeding the output of a subprocess to hashlib. - Use six.reraise rather than the three-argument form of raise. - Adjust test_filteredcommand for Python 3 text handling. - Fix test_ubi_partman.question_has_variables to handle templates files as binary data, since they're mixed-encoding. - Cope with assertItemsEqual/assertCountEqual naming difference between Python 2.7 and 3.2. - Rearrange ubiquity.i18n.get_translations to treat debconf-copydb output as binary data and do field-dependent decoding. - Adjust test_misc and test_upower to cope with file type changes in Python 3. - Use xml.etree.cElementTree instead of libxml2; it's faster, has a smaller footprint by virtue of being built into the standard library, arguably easier to read, and works with Python 3. - Fix a slew of file handle leaks, including making much more liberal use of context managers. - When creating a default username based on the user's full name, decode it back to a Unicode type after running it through the ascii_transliterate encoder. - Remove __pycache__ directories on clean. - Use 'from __future__ import unicode_literals' in tests requiring Unicode literals. - Cope with various builtins and dict methods returning iterators rather than lists in Python 3. - Use chr rather than unichr in Python 3. - Use six.string_types rather than types.StringTypes. - Pass a byte string to Gtk.CssProvider.load_from_data. - Replace all calls to unicode() with calls to six.text_type() or misc.utf8() as appropriate. - Use unicode_literals in ubiquity.keyboard_names. - Fix up ubiquity.filteredcommand.debug and its tests for Python 3. In Python 3, we can write Unicode strings (i.e. str) directly to sys.stderr, and its defaults are such that the worst case is that they get backslash-encoded. Arrange the tests to simulate this. - Simplify ubiquity.nm.decode_ssid using bytearray(), allowing it to work with Python 3 too. - Open .debs in binary mode when verifying their contents. - Decode text read from debconffilter subprocesses in Python 3. - Only encode preseeded values in Python 2. Python 3's debconf module accepts Unicode strings directly. - Encode the second argument to struct.pack. - Fix byte array handling when generating /etc/iftab. - Fix a missing bit of python-apt 0.8 porting in ubiquity.install_misc. - Open /var/lib/dpkg/status in binary mode when compressing it. - Pass bytes rather than text to kdecore.ki18n. - Call sip.setapi("QVariant", 1) on KDE frontend startup, pending a more complete port to the new API. - Use boolean tests rather than isNull, to cope with getting ordinary Python strings rather than QStrings. - In Python 3, pass strings directly to Qt rather than creating QStrings. - (Build-)depend on python3-gi, since python3-gi-cairo doesn't depend on that and in any case we use it directly. * Add Burmese translations. * Remove obsolete uxlaunch handling, since it's no longer used and has been removed from the archive. * Avoid locking failures when clicking on the "update this installer" link more than once. * Move a bit more code into ubiquity.install_misc, including a new target_file helper method. * Remove unnecessary #! lines from non-executable files. * Improve download verification to handle systems not configured for multiarch, i.e. anything other than amd64 by default (LP: #998492). * Automatic update of included source packages: choose-mirror 2.39ubuntu5, clock-setup 0.110ubuntu1, debian-installer-utils 1.89ubuntu1, localechooser 2.40ubuntu1. [ Stéphane Graber ] * Spawn dconf-service from ubiquity-dm and drop all the dbus-launch calls. This should ensure we have a single dconf-service running and that gets killed on ubiquity's exit. * ubiquity-dm: Turn on the compositor in metacity. (LP: #987168) * wallpaper: Update code to properly set the wallpaper when compositing is enabled. -- Colin Watson