Currently the parser creates IQuantity instances for values that are then converted to primitives, i.e. line numbers and byte code indexes. Checking the field id and then reading them with a PrimitiveReader will speed up parsing of recordings with many stack trace entries.