mirror of
https://github.com/rizonesoft/Notepad3.git
synced 2026-06-14 21:09:05 +08:00
250 lines
10 KiB
JSON
250 lines
10 KiB
JSON
// =====================================================================
|
||
// Extended JSON5 / JSON-LD test file for the Notepad3 homebrew JSON lexer
|
||
// (lexilla/lexers_x/LexJSON.cxx).
|
||
//
|
||
// Requires both lexer properties enabled to exercise every branch:
|
||
// lexer.json.allow.comments = 1
|
||
// lexer.json.escape.sequence = 1
|
||
//
|
||
// Each section is annotated with the lexer feature it targets so a failed
|
||
// highlight is easy to attribute to a specific code path.
|
||
// =====================================================================
|
||
|
||
/* ---------------------------------------------------------------
|
||
Stream / block comment. Multiple lines, with stray * and / chars
|
||
inside (separated, so they cannot form a close marker).
|
||
--------------------------------------------------------------- */
|
||
|
||
{
|
||
// ---------- Strings: quoting variants ----------
|
||
doubleQuoted: "plain double-quoted string",
|
||
singleQuoted: 'plain single-quoted string',
|
||
mixedInside1: "contains 'single quotes' inside double",
|
||
mixedInside2: 'contains "double quotes" inside single',
|
||
emptyDouble: "",
|
||
emptySingle: '',
|
||
|
||
// ---------- Strings: line continuation (backslash + EOL) ----------
|
||
lineContLF: "first line\
|
||
second line\
|
||
third line",
|
||
lineContSingle: 'wrap with\
|
||
single quotes too',
|
||
|
||
// ---------- Strings: standard escape sequences ----------
|
||
// (lexer.json.escape.sequence = 1 to highlight these)
|
||
escSimple: "tab=\t nl=\n cr=\r bs=\b ff=\f vt=\v null=\0 slash=\/ bslash=\\ quote=\" apos=\'",
|
||
escUnicode: "äöü € 껮 꿾",
|
||
escSingleStr: 'mix: \té\\done',
|
||
|
||
// ---------- Property names: quoted vs unquoted ----------
|
||
unquotedKey: 'JSON5 allows bare property names',
|
||
"quotedKey": "classic JSON",
|
||
'singleKey': 'single-quoted property name',
|
||
$dollar: 1,
|
||
_underscore: 2,
|
||
mixed123name: 3,
|
||
|
||
// ---------- Numbers: integers, signs, decimals ----------
|
||
zero: 0,
|
||
posInt: 42,
|
||
negInt: -42,
|
||
posSign: +7,
|
||
decimal: 3.14159,
|
||
leadingDot: .5,
|
||
trailingDot: 5. ,
|
||
// Tightened dot-part rule (E1): "1 . 0" is no longer a valid number.
|
||
// Uncomment to verify it now highlights as ERROR (was NUMBER before E1):
|
||
//tightDot: 1 . 0,
|
||
|
||
// ---------- Numbers: exponent / scientific ----------
|
||
exp1: 1e10,
|
||
exp2: 1.5E-3,
|
||
exp3: 2.5e+12,
|
||
expNeg: -6.022e23,
|
||
|
||
// ---------- Numbers: hexadecimal (homebrew-only) ----------
|
||
hexLower: 0xff,
|
||
hexMixed: 0xDeCaF,
|
||
hexBig: 0xCAFEBABE,
|
||
|
||
// ---------- Keywords ----------
|
||
yes: true,
|
||
no: false,
|
||
nothing: null,
|
||
inf: Infinity,
|
||
posInf: +Infinity,
|
||
negInf: -Infinity,
|
||
notNum: NaN,
|
||
|
||
// ---------- Trailing commas (objects + arrays) ----------
|
||
trailingInObj: { a: 1, b: 2, },
|
||
trailingInArr: [ 1, 2, 3, ],
|
||
emptyArr: [],
|
||
emptyObj: {},
|
||
|
||
// ---------- URL hotspots inside strings ----------
|
||
urlHttps: "https://example.com/path?q=1#frag",
|
||
urlHttp: "http://example.com/",
|
||
urlFtp: "ftp://files.example.com/dir/",
|
||
urlSsh: "ssh://git@example.com/repo.git",
|
||
urlGit: "git://example.com/repo.git",
|
||
urlSvn: "svn://svn.example.com/trunk",
|
||
urlMailto: "mailto:user@example.com",
|
||
|
||
// ---------- URL with embedded escape (exercises the URI/LDKEYWORD escape fix) ----------
|
||
// Without the fix the URI style would terminate at '\' and the rest of
|
||
// the string would be mis-styled. With the fix, 'é' is recognized
|
||
// as an escape sequence and the URI style resumes afterwards (well, the
|
||
// remainder is plain string -- the URI ends at the non-URL char).
|
||
urlWithEsc: "https://example.com/café/menu",
|
||
urlWithBackslash:"https://example.com/path\\sub",
|
||
|
||
// ---------- JSON-LD @ keywords ----------
|
||
"@context": "https://schema.org/",
|
||
"@id": "https://example.org/things/1",
|
||
"@type": "Person",
|
||
"@language": "en",
|
||
"@vocab": "https://example.org/vocab#",
|
||
|
||
// ---------- Compact IRIs (single colon, alpha/$/_/- around it) ----------
|
||
"schema:name": "Compact IRI Example",
|
||
"foaf:knows": "another-iri",
|
||
"$compact_-Form:value": "edge characters in prefix",
|
||
|
||
// ---------- Strings that look like compact IRI but aren't (multiple colons -> not highlighted) ----------
|
||
"not:a:compact:iri": "more than one colon disqualifies highlighting",
|
||
"has space:notIRI": "space disqualifies",
|
||
|
||
// ---------- Nested structure with everything mixed ----------
|
||
nested: {
|
||
'array': [
|
||
0xAB,
|
||
.25,
|
||
'item with \'escaped\' single quotes',
|
||
"item with \"escaped\" double quotes",
|
||
{ inner: +Infinity, hex: 0xFEED, },
|
||
],
|
||
'urls': [
|
||
"http://a.example/",
|
||
"https://b.example/path",
|
||
"mailto:noreply@example.com",
|
||
],
|
||
},
|
||
|
||
// ---------- B1: wrong-quote close in ESCAPESEQUENCE ----------
|
||
// To trigger the bug requires (a) lexer.json.escape.sequence = 1 and
|
||
// (b) an escape sequence whose terminating char is immediately followed
|
||
// by the OPPOSITE quote. Pre-fix, the opposite quote falsely closed
|
||
// the string and the rest of the line styled as ERROR. Post-fix, the
|
||
// whole string highlights cleanly.
|
||
b1DoubleAfterU: "before\u1234'after the false-close point",
|
||
b1SingleAfterU: 'before\u5678"after the false-close point',
|
||
b1DoubleAfterX: "before\xAB'after",
|
||
b1SingleAfterX: 'before\xCD"after',
|
||
|
||
// ---------- L1: \xHH hex escape (JSON5 §5.5.4) ----------
|
||
l1Hex: "low=\x00 high=\xFF tilde=\x7E AT=\x40",
|
||
l1HexInSing: 'mix \xAA before \xbb after',
|
||
|
||
// ---------- L2: identity escape (any non-LineTerminator/digit/x/u char) ----------
|
||
// These were ERROR before; JSON5 says \? -> ?, \, -> , etc.
|
||
l2Identity: "\? \, \; \! \% \& \= \[ \] \( \) \{ \}",
|
||
l2InSingle: 'punct: \? \, \; -- still string',
|
||
// Legacy octal (\1..\9) remains forbidden -- uncomment to verify ERROR:
|
||
//l2OctalErr: "\1\2\3",
|
||
|
||
// ---------- L3: bare CR line continuation ----------
|
||
// Cannot be exercised in a CRLF/LF file -- requires CR-only line endings
|
||
// (Edit > Line Endings > Macintosh CR). When that's set, a string ending
|
||
// a line with a single backslash should continue onto the next line just
|
||
// like the LF / CRLF case above.
|
||
|
||
// ---------- L4: U+2028 / U+2029 line continuation ----------
|
||
// The two Unicode LineTerminator chars are below as literals. Each line
|
||
// contains "\<LS>continued" or "\<PS>continued" -- a single STRING that
|
||
// wraps via the JS-style line separator.
|
||
l4LS: "before\
continued",
|
||
l4PS: "before\
continued",
|
||
|
||
// =====================================================================
|
||
// BELOW: extensions covering the strict-ID / hardcoded-literal /
|
||
// signed-NaN / setURL-typo / block-comment-folding lexer changes.
|
||
// =====================================================================
|
||
|
||
// ---------- Hardcoded JSON5 reserved literals: signed NaN ----------
|
||
// Pre-fix the word-list path handled +Infinity / -Infinity but missed
|
||
// +NaN / -NaN (those tokens were never in the keyword list). After the
|
||
// hardcoded-literal change both style as NUMBER.
|
||
posNaN: +NaN,
|
||
negNaN: -NaN,
|
||
|
||
// ---------- Literal-boundary check ----------
|
||
// true / false / null / Infinity / NaN are reserved ONLY when the next
|
||
// character is NOT an identifier-continue char. These keys start with
|
||
// literal text but continue with more identifier characters, so they
|
||
// MUST style as PROPERTYNAME (not KEYWORD / NUMBER).
|
||
nullable: 'starts with literal "null"',
|
||
truer: 'starts with literal "true"',
|
||
falsehood: 'starts with literal "false"',
|
||
Infinityx: 'starts with literal "Infinity"',
|
||
NaNny: 'starts with literal "NaN"',
|
||
|
||
// ---------- Strict Unicode IdentifierName: accepted (UAX #31 ID_Start) ----------
|
||
// Pre-strict accepted any byte >= 0x80 indiscriminately; post-strict
|
||
// requires the actual code point to be a Unicode Letter / $ / _.
|
||
café: 'Latin Extended (precomposed)',
|
||
naïve: 'Latin Extended with diaeresis',
|
||
Σmega: 'Greek capital sigma + ASCII',
|
||
λambda: 'Greek small lambda + ASCII',
|
||
имя: 'Cyrillic letters (Russian "name")',
|
||
名前: 'CJK Han ideographs (Japanese "name")',
|
||
中文: 'CJK Han ideographs (Chinese)',
|
||
|
||
// ---------- IdentifierPart: ZWNJ / ZWJ inside identifier (invisible chars) ----------
|
||
// ZWNJ (U+200C) and ZWJ (U+200D) are valid IdentifierPart per ECMA-262.
|
||
// The bytes are present in the source even though they don't render --
|
||
// inspect with a hex viewer; the unquoted keys below must lex as
|
||
// PROPERTYNAME, not break at the joiner.
|
||
xy_zwnj: 'unquoted key with ZWNJ between x and y',
|
||
xy_zwj: 'unquoted key with ZWJ between x and y',
|
||
|
||
// ---------- Strict Unicode rejection: should style as PARSING ERROR ----------
|
||
// Each leading character is NOT a Unicode Letter, so not ID_Start.
|
||
// Pre-strict accepted these (over-accepting all bytes >= 0x80);
|
||
// post-strict flags them as ERROR.
|
||
|
||
// Emoji 😀 (U+1F600, category So) -- ERROR:
|
||
😀: 'leading emoji must be ERROR',
|
||
|
||
// Math symbol ∞ (U+221E, category Sm) -- ERROR:
|
||
∞: 'leading math symbol must be ERROR',
|
||
|
||
// Currency € (U+20AC, category Sc) -- JSON5 allows ONLY ASCII '$':
|
||
€: 'leading currency symbol must be ERROR',
|
||
|
||
// ---------- URI charset typo fix: ';' inside URL ----------
|
||
// Pre-fix setURL was "...,)," (duplicated ')' with ';' missing), so the
|
||
// URI style terminated at the first ';'. Post-fix ';' is part of the
|
||
// RFC-3986 sub-delim set and the whole URL highlights as URI through
|
||
// the matrix parameters.
|
||
urlSemicolon: "http://example.com/path;jsessionid=abc;v=2",
|
||
|
||
// ---------- Folding: multi-line block comment ----------
|
||
// Fold() now treats a multi-line /* ... */ as a fold region (HEADERFLAG
|
||
// on the opening line). The block below should be foldable; the
|
||
// single-line /* ... */ further down should NOT produce a fold (its
|
||
// enter+exit cancel within the same line).
|
||
|
||
/* This block comment
|
||
spans three lines
|
||
and should produce a fold marker on the opening line. */
|
||
foldDemo: 'after the foldable block comment',
|
||
|
||
/* single-line comment -- should NOT produce a fold marker */
|
||
noFold: 'after the non-foldable comment',
|
||
|
||
// ---------- Trailing item to verify the closing brace styling ----------
|
||
last: 'end-of-test',
|
||
}
|