test: update likelihood scan reference visualization for latest matplotlib

[alexander-held]

The matplotlib 3.9 release caused a change in the legend positioning of the likelihood scan visualization used as reference. This is presumably due to https://matplotlib.org/stable/api/prev_api_changes/api_changes_3.9.0.html#loc-best-for-legend-now-considers-text-and-polycollections. Updating the reference here to be up to date with the latest legend positioning.

The Python 3.8 test uses an older version of matplotlib with different legend positioning. Since support for that version will presumably be dropped sooner or later anyway, it seems reasonable to accept a failing reference figure check. The old legend positioning is still perfectly fine, so this is flagged as xfail for that specific version.

* update legend positioning in likelihood scan visualization test reference
* update test to handle expected failure in Python 3.8

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 100.00%. Comparing base (9aa0447) to head (24fae53).

Additional details and impacted files

```diff @@ Coverage Diff @@ ## master #476 +/- ## ========================================= Coverage 100.00% 100.00% ========================================= Files 22 22 Lines 2093 2093 Branches 346 346 ========================================= Hits 2093 2093 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.