Explode is incorrectly spaced when namespaces are open-ed

In a context where a namespace has been open-ed, it appears the column width for the theorem row is based on a fully qualified name for the theorem (even though such name is not displayed), which throws off the spacing for the whole table. See below.

Minimal example:

import Mathlib.Tactic.Explode

namespace LongNamedNamespace
axiom TheTheorem (x: Type): x = x
end LongNamedNamespace

section
open LongNamedNamespace
theorem Theorem2 (x: Type) : x = x := by
  apply TheTheorem;
/-
Theorem2 : ∀ (x : Type), x = x

0│   │ x                             ├ Type
1│   │ TheTheorem │ x = x
2│0,1│ ∀I                            │ ∀ (x : Type), x = x


-/
#explode Theorem2
end


/-
Theorem2 : ∀ (x : Type), x = x

0│   │ x                             ├ Type
1│   │ LongNamedNamespace.TheTheorem │ x = x
2│0,1│ ∀I                            │ ∀ (x : Type), x = x


-/
#explode Theorem2

Start by running the minimal example using Mathlib.Tactic.Explode and compare the two #explode outputs with and without open LongNamedNamespace. Trace how the theorem row width is chosen; done means the open-namespace output uses the displayed theorem name and no longer adds spacing for the qualified name.