I pushed a (hopefully) improved version of the the locked narrowing
feature in the feature/improved-locked-narrowing branch. It is now
possible:
1. to unlock a locked narrowing,
2. to use narrow-to-region and widen within the bounds of a locked
narrowing,
3. to enter and leave a locked narrowing within a locked narrowing, and
4. to use that feature in Lisp code (either directly or with a helper
macro).
As far as I understand, given that it is possible to unlock a locked
narrowing, adding the possibility to choose how the bounds of the locked
narrowings are computed has become unnecessary.