Lexical pretty-printers for JSON, XML/HTML and JavaScript text. More...

#include "nilorea/n_pretty.h"
#include "nilorea/n_common.h"
#include "nilorea/n_log.h"
#include <ctype.h>
#include <string.h>
#include "cJSON.h"

Include dependency graph for n_pretty.c:

Go to the source code of this file.

Data Structures
struct	JS_STATE
	JS emitter state threaded through the helpers. More...
struct	PRETTY_OUT
	growing output buffer fed by write_and_fit_ex More...

Macros
#define	PRETTY_INDENT_STEP 2
	spaces per indentation level
#define	PRETTY_OUT_PADDING 256
	growth headroom so consecutive short appends share one realloc
#define	PRETTY_TAG_NAME_MAX 31
	longest element name kept for void/raw-text lookups

Functions
static int	_emit_line (PRETTY_OUT o, int depth, const char s, size_t n)
	emit one indented line holding n raw bytes of s
static int	_emit_trimmed (PRETTY_OUT o, int depth, const char s, size_t n)
	emit s trimmed of leading/trailing whitespace as one indented line; empty-after-trim runs are skipped
static const char *	_find_tag_end (const char *p)
	find the '>' closing a tag starting at p, honoring quoted attribute values; returns NULL when the tag never closes
static int	_js_lead (JS_STATE *js)
	emit the indent or the collapsed space before a token, as appropriate
static int	_js_newline (JS_STATE *js)
	terminate the current line unless already at a line start
static int	_js_regex_keyword (const char *s, size_t n)
	whether the n bytes at s are a keyword after which an expression (and therefore a regex literal) can start, e.g.
static size_t	_js_regex_len (const char *s)
	length of the regex literal starting at s (slash to closing slash, escapes and character classes honored, trailing flags included)
static int	_js_regex_possible (char c)
	whether a '/' may start a regex literal after last significant char c (start of input, after an operator, an opener, or a separator)
static size_t	_js_string_len (const char *s)
	length of the string literal starting at s (quote to closing quote, escapes honored, stops at end of line or input on unterminated ones)
static size_t	_js_template_len (const char *s)
	length of the template literal starting at s (backtick to backtick, escapes honored, ${...} interpolations skipped by brace counting)
static int	_js_token (JS_STATE js, const char s, size_t n)
	emit a raw token of n bytes with leading indent/space handling
static int	_jsonl_lead (PRETTY_OUT o, int depth, int at_line_start)
	emit the pending indent of a lenient JSON line, TRUE on success
static int	_jsonl_newline (PRETTY_OUT o, int at_line_start)
	end the current lenient JSON line unless it is still empty, TRUE on success
static int	_name_in_list (const char name, const char *list)
	whether lowercase element name is in a NULL-terminated list
static int	_out_ch (PRETTY_OUT *o, char c)
	append a single character to the output, TRUE on success
static N_STR *	_out_finish (PRETTY_OUT *o)
	hand the buffer over to a new N_STR (no copy); the PRETTY_OUT is emptied
static int	_out_indent (PRETTY_OUT *o, int depth)
	append depth * PRETTY_INDENT_STEP spaces, TRUE on success
static int	_out_mem (PRETTY_OUT o, const char s, size_t n)
	append n bytes of s to the output, TRUE on success
static const char *	_stristr (const char haystack, const char needle)
	case-insensitive search for needle in haystack
static void	_tag_name (const char p, char name, size_t name_size)
	extract the lowercased element name from a tag starting at p ("<name" or "</name")
N_STR *	n_pretty_js (const char *text)
	Conservatively re-indent JavaScript (newlines at top-level braces and semicolons, two-space indent by brace depth).
N_STR *	n_pretty_json (const char *text)
	Reformat a JSON document with cJSON (parse + print).
N_STR *	n_pretty_json_lenient (const char *text)
	Re-indent a JSON document by punctuation alone, without parsing it.
N_STR *	n_pretty_xml (const char *text)
	Re-indent an XML or HTML document (one tag or text run per line).

Variables
static const char *	PRETTY_RAWTEXT_ELEMENTS [] = {"script", "style", NULL}
	HTML elements whose content is raw text (no nested markup).
static const char *	PRETTY_VOID_ELEMENTS []
	HTML elements that never have content or a closing tag.

Detailed Description

Lexical pretty-printers for JSON, XML/HTML and JavaScript text.

Author: Castagnier Mickael

Version: 1.0

Date: 10/07/2026

Definition in file n_pretty.c.

Data Structure Documentation

◆ JS_STATE

struct JS_STATE

JS emitter state threaded through the helpers.

Definition at line 381 of file n_pretty.c.

Collaboration diagram for JS_STATE:

Data Fields
int	at_line_start	1 when nothing has been emitted on the current line yet
int	indent	current indentation depth
int	kw_regex	1 when the last token was a keyword that can precede a regex literal
char	last_sig	last significant code character emitted (0 = none yet)
PRETTY_OUT	o	output buffer
int	pending_space	1 when collapsed whitespace separates the previous and next token

◆ PRETTY_OUT

struct PRETTY_OUT

growing output buffer fed by write_and_fit_ex

Definition at line 40 of file n_pretty.c.

Collaboration diagram for PRETTY_OUT:

Data Fields
char *	buf	buffer, reallocated on growth
NSTRBYTE	size	total allocation in bytes
NSTRBYTE	written	meaningful bytes, excluding the terminating '\0'

Macro Definition Documentation

◆ PRETTY_INDENT_STEP

#define PRETTY_INDENT_STEP 2

spaces per indentation level

Definition at line 52 of file n_pretty.c.

Referenced by _out_indent().

◆ PRETTY_OUT_PADDING

#define PRETTY_OUT_PADDING 256

growth headroom so consecutive short appends share one realloc

Definition at line 50 of file n_pretty.c.

Referenced by _out_mem().

◆ PRETTY_TAG_NAME_MAX

#define PRETTY_TAG_NAME_MAX 31

longest element name kept for void/raw-text lookups

Definition at line 54 of file n_pretty.c.

Referenced by n_pretty_xml().

Function Documentation

◆ _emit_line()

int _emit_line	(	PRETTY_OUT *	o,
		int	depth,
		const char *	s,
		size_t	n )

static

emit one indented line holding n raw bytes of s

Definition at line 269 of file n_pretty.c.

References _out_ch(), _out_indent(), and _out_mem().

Referenced by _emit_trimmed(), and n_pretty_xml().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ _emit_trimmed()

int _emit_trimmed	(	PRETTY_OUT *	o,
		int	depth,
		const char *	s,
		size_t	n )

static

emit s trimmed of leading/trailing whitespace as one indented line; empty-after-trim runs are skipped

Definition at line 277 of file n_pretty.c.

References _emit_line().

Referenced by n_pretty_xml().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ _find_tag_end()

const char * _find_tag_end ( const char * p )

static

find the '>' closing a tag starting at p, honoring quoted attribute values; returns NULL when the tag never closes

Definition at line 242 of file n_pretty.c.

Referenced by n_pretty_xml().

Here is the caller graph for this function:

◆ _js_lead()

int _js_lead ( JS_STATE * js )

static

emit the indent or the collapsed space before a token, as appropriate

Definition at line 397 of file n_pretty.c.

References _out_ch(), _out_indent(), JS_STATE::at_line_start, JS_STATE::indent, JS_STATE::o, and JS_STATE::pending_space.

Referenced by _js_token().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ _js_newline()

int _js_newline ( JS_STATE * js )

static

terminate the current line unless already at a line start

Definition at line 415 of file n_pretty.c.

References _out_ch(), JS_STATE::at_line_start, JS_STATE::o, and JS_STATE::pending_space.

Referenced by n_pretty_js().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ _js_regex_keyword()

int _js_regex_keyword	(	const char *	s,
		size_t	n )

static

whether the n bytes at s are a keyword after which an expression (and therefore a regex literal) can start, e.g.

return /re/

Definition at line 370 of file n_pretty.c.

Referenced by n_pretty_js().

Here is the caller graph for this function:

◆ _js_regex_len()

size_t _js_regex_len ( const char * s )

static

length of the regex literal starting at s (slash to closing slash, escapes and character classes honored, trailing flags included)

Definition at line 466 of file n_pretty.c.

Referenced by n_pretty_js().

Here is the caller graph for this function:

◆ _js_regex_possible()

int _js_regex_possible ( char c )

static

whether a '/' may start a regex literal after last significant char c (start of input, after an operator, an opener, or a separator)

Definition at line 363 of file n_pretty.c.

Referenced by n_pretty_js().

Here is the caller graph for this function:

◆ _js_string_len()

size_t _js_string_len ( const char * s )

static

length of the string literal starting at s (quote to closing quote, escapes honored, stops at end of line or input on unterminated ones)

Definition at line 424 of file n_pretty.c.

Referenced by n_pretty_js().

Here is the caller graph for this function:

◆ _js_template_len()

size_t _js_template_len ( const char * s )

static

length of the template literal starting at s (backtick to backtick, escapes honored, ${...} interpolations skipped by brace counting)

Definition at line 441 of file n_pretty.c.

Referenced by n_pretty_js().

Here is the caller graph for this function:

◆ _js_token()

int _js_token	(	JS_STATE *	js,
		const char *	s,
		size_t	n )

static

emit a raw token of n bytes with leading indent/space handling

Definition at line 409 of file n_pretty.c.

References _js_lead(), _out_mem(), and JS_STATE::o.

Referenced by n_pretty_js().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ _jsonl_lead()

int _jsonl_lead	(	PRETTY_OUT *	o,
		int	depth,
		int *	at_line_start )

static

emit the pending indent of a lenient JSON line, TRUE on success

Definition at line 110 of file n_pretty.c.

References _out_indent().

Referenced by n_pretty_json_lenient().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ _jsonl_newline()

int _jsonl_newline	(	PRETTY_OUT *	o,
		int *	at_line_start )

static

end the current lenient JSON line unless it is still empty, TRUE on success

Definition at line 117 of file n_pretty.c.

References _out_ch().

Referenced by n_pretty_json_lenient().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ _name_in_list()

int _name_in_list	(	const char *	name,
		const char **	list )

static

whether lowercase element name is in a NULL-terminated list

Definition at line 220 of file n_pretty.c.

Referenced by n_pretty_xml().

Here is the caller graph for this function:

◆ _out_ch()

int _out_ch	(	PRETTY_OUT *	o,
		char	c )

static

append a single character to the output, TRUE on success

Definition at line 63 of file n_pretty.c.

References _out_mem().

Referenced by _emit_line(), _js_lead(), _js_newline(), _jsonl_newline(), n_pretty_js(), and n_pretty_json_lenient().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ _out_finish()

N_STR * _out_finish ( PRETTY_OUT * o )

static

hand the buffer over to a new N_STR (no copy); the PRETTY_OUT is emptied

Definition at line 76 of file n_pretty.c.

References PRETTY_OUT::buf, char_to_nstr_nocopy(), Free, PRETTY_OUT::size, and PRETTY_OUT::written.

Referenced by n_pretty_js(), n_pretty_json_lenient(), and n_pretty_xml().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ _out_indent()

int _out_indent	(	PRETTY_OUT *	o,
		int	depth )

static

append depth * PRETTY_INDENT_STEP spaces, TRUE on success

Definition at line 68 of file n_pretty.c.

References _out_mem(), and PRETTY_INDENT_STEP.

Referenced by _emit_line(), _js_lead(), and _jsonl_lead().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ _out_mem()

int _out_mem	(	PRETTY_OUT *	o,
		const char *	s,
		size_t	n )

static

append n bytes of s to the output, TRUE on success

Definition at line 57 of file n_pretty.c.

References PRETTY_OUT::buf, PRETTY_OUT_PADDING, PRETTY_OUT::size, write_and_fit_ex(), and PRETTY_OUT::written.

Referenced by _emit_line(), _js_token(), _out_ch(), _out_indent(), and n_pretty_json_lenient().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ _stristr()

const char * _stristr	(	const char *	haystack,
		const char *	needle )

static

case-insensitive search for needle in haystack

Definition at line 259 of file n_pretty.c.

Referenced by n_pretty_xml().

Here is the caller graph for this function:

◆ _tag_name()

void _tag_name	(	const char *	p,
		char *	name,
		size_t	name_size )

static

extract the lowercased element name from a tag starting at p ("<name" or "</name")

Definition at line 228 of file n_pretty.c.

Referenced by n_pretty_xml().

Here is the caller graph for this function:

Variable Documentation

◆ PRETTY_RAWTEXT_ELEMENTS

const char* PRETTY_RAWTEXT_ELEMENTS[] = {"script", "style", NULL}

static

HTML elements whose content is raw text (no nested markup).

Definition at line 217 of file n_pretty.c.

Referenced by n_pretty_xml().

◆ PRETTY_VOID_ELEMENTS

const char* PRETTY_VOID_ELEMENTS[]

static

Initial value:

= {
    "area", "base", "br", "col", "embed", "hr", "img", "input",
    "link", "meta", "param", "source", "track", "wbr", NULL}

HTML elements that never have content or a closing tag.

Definition at line 212 of file n_pretty.c.

Referenced by n_pretty_xml().

Data Structures

Macros

Functions

Variables

Detailed Description

Data Structure Documentation

◆ JS_STATE

◆ PRETTY_OUT

Macro Definition Documentation

◆ PRETTY_INDENT_STEP

◆ PRETTY_OUT_PADDING

◆ PRETTY_TAG_NAME_MAX

Function Documentation

◆ _emit_line()

◆ _emit_trimmed()

◆ _find_tag_end()

◆ _js_lead()

◆ _js_newline()

◆ _js_regex_keyword()

◆ _js_regex_len()

◆ _js_regex_possible()

◆ _js_string_len()

◆ _js_template_len()

◆ _js_token()

◆ _jsonl_lead()

◆ _jsonl_newline()

◆ _name_in_list()

◆ _out_ch()

◆ _out_finish()

◆ _out_indent()

◆ _out_mem()

◆ _stristr()

◆ _tag_name()

Variable Documentation

◆ PRETTY_RAWTEXT_ELEMENTS

◆ PRETTY_VOID_ELEMENTS