Dependency-free lexical formatters aimed at viewers: they re-indent a document so a human can read it, without validating it. More...

Functions
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).

Detailed Description

Dependency-free lexical formatters aimed at viewers: they re-indent a document so a human can read it, without validating it.

Malformed input is reformatted on a best-effort basis; the formatters never crash on it.

Limits by design (these are viewers, not compilers):

n_pretty_xml() treats HTML like XML with void-element and raw-text (script/style) awareness; whitespace inside text nodes is trimmed at both ends, so whitespace-significant content (pre) is not preserved exactly.
n_pretty_js() is a conservative re-indenter (newlines at braces and semicolons), not a beautifier. Strings, template literals, comments and regex literals are copied verbatim. A '/' is lexed as a regex when it follows an operator, an opener, a separator, or an expression keyword (return, typeof, case, ...); the remaining ambiguity (a regex right after a closing parenthesis, as in if (x) /re/.test(y)) is lexed as division, which can only mis-break lines, never corrupt the text.
n_pretty_json() needs the vendored cJSON (HAVE_CJSON); without it the function logs an error and returns NULL. It parses, so it reformats valid documents only: use n_pretty_json_lenient() for the rest.
n_pretty_json_lenient() re-indents by punctuation without parsing, so it also breaks up a document cJSON rejects (a truncated body, a trailing comma, NDJSON). It cannot report that the input was invalid and it does not normalise anything, so prefer n_pretty_json() first and label what the fallback produces as approximate.

All functions return a freshly allocated N_STR the caller must release with free_nstr(), or NULL when the input is NULL, empty, or does not look like the expected format.

Function Documentation

◆ n_pretty_js()

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).

Parameters

text	The script text. NULL or empty input returns NULL.

Returns: A new indented N_STR (free with free_nstr), or NULL.

Definition at line 489 of file n_pretty.c.

References __n_assert, _js_newline(), _js_regex_keyword(), _js_regex_len(), _js_regex_possible(), _js_string_len(), _js_template_len(), _js_token(), _out_ch(), _out_finish(), JS_STATE::at_line_start, PRETTY_OUT::buf, Free, JS_STATE::indent, JS_STATE::kw_regex, JS_STATE::last_sig, JS_STATE::o, and JS_STATE::pending_space.

Referenced by main().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ n_pretty_json()

N_STR * n_pretty_json ( const char * text )

Reformat a JSON document with cJSON (parse + print).

Parameters

text	The JSON text. NULL, empty, or non-JSON input returns NULL.

Returns: A new indented N_STR (free with free_nstr), or NULL.

Definition at line 89 of file n_pretty.c.

References __n_assert, char_to_nstr_nocopy(), LOG_ERR, and n_log.

Referenced by main().

Here is the call graph for this function:

Here is the caller graph for this function:

◆ n_pretty_json_lenient()

N_STR * n_pretty_json_lenient ( const char * text )

Re-indent a JSON document by punctuation alone, without parsing it.

String literals are copied verbatim, containers open and close a level, and each comma ends a line, so a document cJSON rejects (truncated, trailing comma, concatenated values) still becomes readable.

Parameters