Subdividing quarter note into 33-tuplet obscures the next beat in front unless it is the last beat of the measure

Problem. When subdividing a quarter note into a 33-tuplet (or a 99-tuplet but not a 32- or 34-tuplet), the measure hides the following beat within the 33rd subdivision of the tuplet. I've tried this in 4/4, 3/8, and 5/4, so I assume it affects most if not all time signatures. So where does that beat go? Compare:

Here is where MuseScore says beat two is (approx.) when I subdivide a half rest. This behavior is expected.

Here is where MuseScore says beat two is (exact) when I subdivide a quarter rest. This behavior is not expected.

The denominator of the tuplet ratio also does not seem to matter.

Edit: I should also note that 99-tuplets also have this problem, but when I create one in the last beat(s), there is even stranger behavior.

Workaround. I can attempt to workaround this by creating the tuplet on the last beat(s) in the measure that take up a quarter rest. In 4/4, it would be on beat 4, and in 3/8, it would be on beats 2 and 3. Then, I can copy that together with beat one of the next measure and paste that where I want it. However, the outcome is still not desirable because the actual subdivision is 32 rather than 33, and beat two still appears as if it is within the tuplet.

And indeed, if I divide beat two, we see the problem is only semi-solved.

This method actually does nothing if the denominator of the tuplet ratio is 1.

Versions. I've checked this problem in the following versions:
V1.3, rev:5702 (QT 4.8.4)
OS: macOS 10.14, Arch.: x86_64, MuseScore version (64-bit): 3.0.0, revision: 3abd82b
OS: macOS 10.14, Arch.: x86_64, MuseScore version (64-bit): 3.0.0, revision: c07ed54
OS: OS X 10.14, Arch.: x86_64, MuseScore version (64-bit): 2.3.2, revision: 4592407

This problem does not seem to exist in 1.3, so this is a regression (unless it affects a different subdivision of the quarter or other note(s)).