number_precision

`number_precision`

Enumeration of number parsing modes.

Synopsis

Defined in header <boost/json/parse_options.hpp>.

enum class number_precision
    : unsigned char;

Values

Name Description

Name	Description
`imprecise`	Fast, but potentially less precise mode.
`precise`	Slower, but precise mode.
`none`	The fastest mode, that only validates encountered numbers without parsing them.

imprecise

Fast, but potentially less precise mode.

precise

Slower, but precise mode.

none

The fastest mode, that only validates encountered numbers without parsing them.

Description

These values are used to select the way to parse numbers. The default mode used by parsing functions is imprecise. It has the precision not less than 15 decimal places (the number required for IEEE double). This is also the number guaranteed by functions in most C++ standard library implementations.

But those functions often give a more precise result for certain numbers. For cases where such extra precision is needed the library provides precise mode. That extra precision comes with a performance cost, though.

Finally, users might want a to parse numbers with an external function when they implement a custom handler for basic_parser. Since it is wasteful to parse a number only to throw the result away, the library provides a mode to not parse numbers at all. Note: the library still have to validate that the input is a number, in order to catch syntax errors and to know when the number ends.