Declares where a number gets the small gaps that separate its digit groups (GH #192), "2111230496" being displayed as "2 111 230 496". More...

#include <cstddef>
#include <utility>
#include <vector>
#include <wx/string.h>

Include dependency graph for DigitGrouping.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Functions
int	wxm::LocaleDigitGroupSize ()
	How many digits a group has: what the C library's LC_NUMERIC locale says, if that is 3 or 4, otherwise 3.

std::vector< size_t >	wxm::DigitGroupGaps (const wxString &number, size_t minDigits, size_t groupSize=3)
	The indices into number before which a digit-group gap is drawn, ascending.

std::vector< std::vector< size_t > >	wxm::DigitGroupGapsInTokens (const std::vector< std::pair< wxString, bool > > &tokens, size_t minDigits, size_t groupSize=3)
	DigitGroupGaps() for code, where a number may span several tokens.

int	wxm::DigitGroupGapWidth (double fontSize)
	How wide a digit-group gap is, in pixels, for text of the given font size.

Detailed Description

Declares where a number gets the small gaps that separate its digit groups (GH #192), "2111230496" being displayed as "2 111 230 496".

This only decides where the gaps go. They are a display matter only: the text a cell holds, copies, saves and exports never contains them, which is why this works on indices into the text rather than producing a new string.

Function Documentation

◆ DigitGroupGaps()

std::vector< size_t > wxm::DigitGroupGaps	(	const wxString &	number,
		size_t	minDigits,
		size_t	groupSize = `3`
	)

The indices into number before which a digit-group gap is drawn, ascending.

number is a number as Maxima prints it, optionally preceded by a sign (including the unicode minus sign MathParser turns "-" into): "12345", "−3.14159", "1.234567b10", "6.02214076e23".

The digits before the decimal point are grouped from the right, the ones after it from the decimal point on, so the groups always line up with the point: "12 345.678 9".
Nothing is grouped unless the mantissa (the digits before and after the point together) has at least minDigits digits, so a short number like a year (2026) stays as it is by default.
An exponent ("e-10", "b25") and anything else after the mantissa is never grouped, nor is anything that doesn't start with a (signed) digit or point.

◆ DigitGroupGapsInTokens()

std::vector< std::vector< size_t > > wxm::DigitGroupGapsInTokens	(	const std::vector< std::pair< wxString, bool > > &	tokens,
		size_t	minDigits,
		size_t	groupSize = `3`
	)

DigitGroupGaps() for code, where a number may span several tokens.

MaximaTokenizer makes "3.14159" three tokens, "3", "." and "14159", and ".5" two. Grouped one token at a time, the digits after the point would be grouped from the right, like an integer's; this joins such tokens up into the number they form first.

tokens holds each token's text and whether it is a number token. The result has one entry per token: the gaps inside it, as indices into that token's own text. A gap never falls on a token boundary, so each gap belongs to exactly one token.

◆ DigitGroupGapWidth()

int wxm::DigitGroupGapWidth ( double fontSize )

How wide a digit-group gap is, in pixels, for text of the given font size.

5/24 of an em: halfway between a typographic thin space (a sixth of an em, what the gap started out as, which made the groups a bit hard to tell apart) and a quarter of an em (what it was widened to next, which felt a little too wide). Output (TextCell) and code (EditorCell) both use this, so that a number looks the same in either. Never less than one pixel, so the gap doesn't vanish at tiny zoom levels.

◆ LocaleDigitGroupSize()

int wxm::LocaleDigitGroupSize ( )

How many digits a group has: what the C library's LC_NUMERIC locale says, if that is 3 or 4, otherwise 3.