I'm not sure I agree with your presentation of this.

First, it's not a matter of "robustness" but of feature coverage.

"Robustness" is, roughly speaking, resistance to weathering (environmental factors), error, damage, or breakage.

The test as I wrote it is designed to verify the correct indentation behavior of the various paragraph types, because each type differs in whether indentation is applied or not.

That the paragraph indentation amount is configurable in mm is another matter entirely.

I can foresee a separate test that checks whether the paragraph indentation as specified by the user (by modifying the `Pi` register is honored.

Also, if we should inadvertently change the O or Pi register defaults, the existing test would catch that, and with your changes it would not, since it extracts their values and accepts them whatever they may be.

Thus, I am rejecting the patch, dropping the summary tag, redrawing the scope of the ticket, and retaining it.  I don't object to the idea; the more test coverage we can get of mm compared to its current amount, the better.