diff --git a/docs/development/V1-2_ANATOMICAL_SHEETS.md b/docs/development/V1-2_ANATOMICAL_SHEETS.md index 49d86083..5c375ae3 100644 --- a/docs/development/V1-2_ANATOMICAL_SHEETS.md +++ b/docs/development/V1-2_ANATOMICAL_SHEETS.md @@ -1567,8 +1567,9 @@ of the document's JSON text in one form, the same for every layer: cap + 1, multi-byte and astral characters, a dense document, numbers that JavaScript prints with an exponent); the core's size tests of V1-2.2 reuse them. -- The numbers' printed lengths are summed first, one number at a time, and - the whole text is built only once they fit the cap: a small payload of +- The numbers' printed lengths are summed first (one linear walk, each number + added in turn), and the whole text is built only once they fit the cap: a + small payload of numbers with huge exponents (`1e131071` prints 131 072 digits) answers `document.too_large`, never an out-of-memory error.