Regular TextRanges are very expensive. In that cases where text ranges value can be predicted should use computed text ranges. For example:
{ "name": "John Doe" }
where "name" is the whole text range in KeyDeclarationIn the previous case we shouldn't create a new text range just for getting its "pure" value (name itself), but we know exactly that "name" is computed from name and two other quotes. It is easy to calculate.
Regular TextRanges are very expensive. In that cases where text ranges value can be predicted should use computed text ranges. For example:
{ "name": "John Doe" } where "name" is the whole text range in KeyDeclarationIn the previous case we shouldn't create a new text range just for getting its "pure" value (name itself), but we know exactly that "name" is computed from name and two other quotes. It is easy to calculate.