Suppose that $A \subseteq \mathbb{R}$, and consider the Lebesgue outer measure $m: \mathcal{P}(\mathbb{R}) \to [0,\infty]$ on $\mathbb{R}$. Denote by $A^o$, $\overline{A}$ and $\partial A$ the interior, closure and boundary of $A$, respectively.
Suppose that $m(A^o) = m(\overline{A})$. Is it then true that $m(\partial A) = 0$?
This is certainly true if $m(A) < \infty$. Since open sets and closed sets are Lebesgue measurable, we have that $A^o$ and $\overline{A}$ are Lebesgue measurable, and hence so is $\partial A = \overline{A} \setminus A^o$. Therefore, we can write $m(\overline{A}) = m(A^o) + m(\partial A)$. By the monotonicity of the Lebesgue outer measure, we also have that $m(A^o) < \infty$. Hence, we can cancel to conclude. But what about the case $m(A) = \infty$?
If we take for example $A = \bigcup_{n \geq 1} (n - 1, n)$, then $m(A^0) = m(\overline{A}) = \infty$, but $m(\partial A) = m(\mathbb{N}) = 0$. Are there any counter examples?