This test caused other tests to fail due to missing cleanup.

Configure the replaygain analysis by passing arguments to the Backends. This
avoids the difference between ReplayGain and EBU r128 backends; every Backend
can now fulfil both tasks. Additionally it eases Backend development as the
difference between the two tag formats is now completely handled in the main
Plugin, not in the Backends.

Allow to configure the target level for R128_* tags separately from REPLAYGAIN_*
tags and skip peak calculation for R128_* tags if possible.

Assert that analysing the same track with different target levels yields
different gain adjustments.

Assert that the replaygain plugin does not write REPLAYGAIN_* tags but R128_*
tags, when instructed to do so.

This test is skipped for the `command` backend as it does not support OPUS.

- document `r128_targetlevel`
- explain difference between `targetlevel` and `r128_targetlevel`
- deprecate `method` option: use `targetlevel` instead.

Add changelog entries for the introduction of the `r128_targetlevel`
configuration option, superseding the `method` option.

Apply improvements suggested in GitHub PullRequest beetbox#3065:
- be idiomatic
  - 0 is falsy
  - check enum equality, not identity
  - mutate list by constructing a new one
- improve documentation
  - fix a typo
  - do not mention deprecation of a config option