fix: scope Exercise 1.2.2's boundedness hypothesis to its own conjunct

[Chessing234]

Exercise 1.2.2 currently reads

example : ∃ f: ℕ → ℝ → ℝ, ∃ F: ℝ → ℝ, ∃ M, ∀ n, ∀ x ∈ Set.Icc 0 1, |f n x| ≤ M ∧
    (∀ x ∈ Set.Icc 0 1, Filter.atTop.Tendsto (fun n ↦ f n x) (nhds (F x))) ∧
    (∀ n, RiemannIntegrableOn (f n) (Icc 0 1)) ∧
    ¬ RiemannIntegrableOn F (Icc 0 1) := by

∧ binds tighter than the preceding binders, so ∀ n, ∀ x ∈ Icc 0 1 scopes over
the entire four-way conjunction rather than over |f n x| ≤ M alone. The three
later conjuncts mention neither n nor that x — the second rebinds its own
x, the third rebinds its own n, shadowing the outer one — so each is
re-asserted once per (n, x) pair.

Since ℕ and Icc 0 1 are both nonempty this is equivalent to the intended
statement, so nothing becomes provable or unprovable here. The cost is only that
a reader has to notice that equivalence before they can see what the exercise is
asking, and the shadowed ∀ n invites a misreading of the third conjunct.

The fix parenthesises the boundedness hypothesis as its own conjunct, which is
how Exercise 1.2.2' immediately below already writes exactly the same hypothesis:

(∃ M, ∀ n, ∀ x ∈ Set.Icc 0 1, |f n x| ≤ M) → ...

so the two neighbouring statements now agree on the shape of the shared
hypothesis.

Independent of #649, which touches Exercise 1.2.2' a few lines below; the changed
lines are disjoint, so the two merge in either order.

No Lean toolchain build was run for this — relying on the build CI on this PR.