WHY THEM  Writes about problem description style in their own reviews.

“Unrelated but this one is computable (maybe the only one?) For example: ```lean abbrev SetTheory.Set.empty_iProd_equiv (X: (∅:Set) → Set) : iProd X ≃ Unit where toFun := fun t ↦ () invFun := fun x ↦ ⟨tuple fun e ↦ False.elim (not_mem_empty _ e.property), by apply tuple_mem_iProd⟩ left_inv := by intro t have h := (mem_iProd _).mp t.property obtain ⟨x, ht⟩ := h ext rw [ht, tuple_inj] ext e have := not_mem_empty e contradiction right_inv := by intro simp ``` Not included because the book has a "why" seemingly for a related question: <img width="1468" height="616" alt="Screenshot 2025-08-07 at 04 16 37" src="https://github.com/user-attachments/assets/edd16e6f-927a-4a37-8bb4-302761ddac53" /> ”

WHY THEM  Writes about async, direct, low ritual in their own reviews.

“tout au singulier = plus de débat et pas de cas par cas par nom de module comme pour des tables sql par exemple”

WHY THEM  Has reviewed this kind of change before.

“The `handleReviewClick` callback deps look correct. One thing — `router` and `transitionRouter` are both in the deps array which is fine, but it would be nice to double-check that `useTransitionRouter` from `next-view-transitions` is stable across renders (i.e. doesn't cause unnecessary re-creation of the callback). AFAIK it should be but worth confirming.”

WHY THEM  Writes about problem description style in their own reviews.

“Wdym ? I want to put all the json result just to make sure i keep every data. This table will be removed in the future when we're sure the payment implementation is ok”

WHY THEM  Has reviewed this kind of change before.

“This looks wrong to me. mb_cstring() doesn't actually increase 'len' at all. It makes sure that there's enough _room_ for 'len', and adds a zero at the end, but the length remains the same. So the "mb.len - 1" is wrong. The length of the string is just "mb.len". Put another way: mb_cstring() is very different from "put_char(0)". mb_cstring makes sure the membuffer contents can be used as a valid C string, *without modifying* the membuffer. So "strlen(membuf)" should always just have been "mb.len". The two are the same (as long as the membuffer doesn't contain NUL characters anywhere else, of course).”

WHY THEM  Writes about code protectiveness and consistency enforcement in their own reviews.

“shouldn't this be gathered with the upper exception? ``` raise ToolException( "No editable documents available. Only .docx documents can be edited." ) ```”