Merge lp://staging/~mvo/ubuntu-release-upgrader/use-str.format into lp://staging/ubuntu-release-upgrader

Maybe str.format is better for other reasons, but you can certainly still get tracebacks from it:

$ python3 -c "'hello {0'.format('world')"
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ValueError: expected '}' before end of string

Hi Michael, thanks for your comment. Indeed, I should have been more clear.

The issue that is fixed is:

$ python -c 'print("foo" % "bar")'
Traceback (most recent call last):
  File "<string>", line 1, in <module>
TypeError: not all arguments converted during string formatting

$ python -c 'print("foo".format("bar"))'
foo

But yes, if a translation has a incorrect number of {} it will still crash, so maybe its not worth
doing all the work and instead use the "Foo %(bar)s" % { "bar": "some-str" } approach where its
also ok to leave out a argument. I.e.
$ python -c 'print("foo" % {"bar": "bar"})'
foo

It might be a lot of work to port, but I will suggest using my flufl.i18n package, which has a lot of nice convenience APIs and safeguards for doing i18n interpolation.

http://pythonhosted.org/flufl.i18n/docs/using.html

And of course, it's packaged up :).

Also note that with modern Pythons

'{0}'.format('foo')

is equivalent to

'{}'.format('foo')