The goal of the visibility increase is to help in cases where timing is
an issue (by showing the approach circle etc.). This doesn't need to
apply to catch.
@smoogipoo interested as to whether you agree with this one. Visually it
looks better to me but it does change the behaviour for only osu!catch,
so I'm not 100% confident on it.
Closes#13367.
This seems to be the simplest way to avoid calls to `Play` after the
underlying sample may have been disposed. As per the issue thread, a
local workaround is acceptable here.
Closes#13223.
* Move steps from ctor to a separate basic test.
* Wait for barrage to complete in basic test, as not doing so polluted
state of other tests.
* Reset score processor after every test.
It's not valid in the base `HitErrorMeter`, as the colour meter only
displays colour for a given judgement, so it is still valid to add new
items to it even if the hit window is 0, as misses are still possible.
`CatchHitWindows` were a vestige from the past, and were not actually
used anywhere except for the hit error meter test, giving off an
appearance that the hit error meter was working properly.
`CatchHitObject` actually specifies empty hit windows.