sascha
/
difftastic
mirror of https://github.com/Wilfred/difftastic/
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Support xml-model & xml-stylesheet instructions

pull/504/merge
ObserverOfTime 2023-08-05 18:19:42 +07:00
parent 1b9749175d
commit 9e8ce35817
No known key found for this signature in database
GPG Key ID: 8A2DEA1DBAEBCA9E
12 changed files with 7829 additions and 6802 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
.github/workflows/ci.yml vendored
Unescape Escape View File

@ -19,5 +19,6 @@ jobs:
- run: npm install - run: npm install
name: Install dependencies name: Install dependencies
- run: >- - run: >-
npm test-dtd && npm test-xml npm run test-dtd &&
npm run test-xml
name: Run tests name: Run tests

32
common/index.js
Unescape Escape View File

@ -26,16 +26,29 @@ module.exports.NAME_CHAR = '[\-.0-9\xB7' +
'\u0300-\u036F\u203F-\u2040' + '\u0300-\u036F\u203F-\u2040' +
module.exports.NAME_START_CHAR.substring(1); module.exports.NAME_START_CHAR.substring(1);
/** @param {"'" | ''} q */ /**
* @param {"'" | ''} q
* @see {@link https://www.w3.org/TR/xml/#NT-PubidChar}
*/
module.exports.pubid_char = (q) => module.exports.pubid_char = (q) =>
new RegExp(`[\x20\x0D\x0Aa-zA-Z0-9\-${q}()+,./:=?;!*#@$_%]*`); new RegExp(`[ \\r\\na-zA-Z0-9\\-${q}()+,./:=?;!*#@$_%]*`);
/** /**
* @param {GrammarSymbols<any>} $ * @param {GrammarSymbols<any>} $
* @param {'"' | "'"} q * @param {'"' | "'"} q
*/ */
module.exports.att_value = ($, q) => module.exports.att_value = ($, q) =>
seq(q, repeat(choice(new RegExp(`[^<&${q}]`), $._Reference)), q); seq(
q,
field(
'content',
repeat(choice(
new RegExp(`[^<&${q}]`),
$._Reference
))
),
q
);
/** /**
* @param {GrammarSymbols<any>} $ * @param {GrammarSymbols<any>} $
@ -44,11 +57,14 @@ module.exports.att_value = ($, q) =>
module.exports.entity_value = ($, q) => module.exports.entity_value = ($, q) =>
seq( seq(
q, q,
repeat(choice( field(
new RegExp(`[^<%&${q}]`), 'content',
$.PEReference, repeat(choice(
$._Reference new RegExp(`[^<%&${q}]`),
)), $.PEReference,
$._Reference
))
),
q q
); );

18
tree-sitter-dtd/grammar.js
Unescape Escape View File

@ -2,7 +2,7 @@
* @file Tree-sitter grammar definition for DTD * @file Tree-sitter grammar definition for DTD
* @author ObserverOfTime * @author ObserverOfTime
* @license MIT * @license MIT
* @see {@link https://www.w3.org/TR/xml/|W3C standard} * @see {@link https://www.w3.org/TR/xml/|XML standard}
*/ */
const c = require('../common'); const c = require('../common');
@ -44,7 +44,8 @@ module.exports = grammar({
O($._S), O($._S),
'[', '[',
repeat($._extSubsetDecl), repeat($._extSubsetDecl),
']]>' ']',
']>'
), ),
_markupdecl: $ => choice( _markupdecl: $ => choice(
@ -265,7 +266,7 @@ module.exports = grammar({
_Char: _ => /[\u0001-\uD7FF\uE000-\uFFFD\u{10000}-\u{10FFFF}]/u, _Char: _ => /[\u0001-\uD7FF\uE000-\uFFFD\u{10000}-\u{10FFFF}]/u,
_S: _ => /[\x20\x09\x0D\x0A]+/, _S: _ => /[ \t\r\n]+/,
Name: _ => new RegExp(`${c.NAME_START_CHAR}${c.NAME_CHAR}*`, 'u'), Name: _ => new RegExp(`${c.NAME_START_CHAR}${c.NAME_CHAR}*`, 'u'),
@ -332,20 +333,17 @@ module.exports = grammar({
EncName: _ => /[A-Za-z][A-Za-z0-9._\-]*/, EncName: _ => /[A-Za-z][A-Za-z0-9._\-]*/,
// TODO: parse attributes
PI: $ => seq( PI: $ => seq(
'<?', '<?',
// FIXME: disallow /xml/i // FIXME: disallow /^xml$/i
alias($.Name, $.PITarget), alias($.Name, $.PITarget),
O(seq($._S, /([^?]|\?[^>])*/)), O(seq($._S, /([^?]|\?[^>])*/)),
'?>' '?>'
), ),
Comment: _ => token(seq( Comment: _ => token(
'<!--', seq('<!--', /([^-]|-[^-])*/, '-->')
/([^-]|-[^-])*/, ),
'-->'
)),
_Eq: $ => seq(O($._S), '=', O($._S)) _Eq: $ => seq(O($._S), '=', O($._S))
} }

140
tree-sitter-dtd/src/grammar.json generated
Unescape Escape View File

@ -104,7 +104,11 @@
}, },
{ {
"type": "STRING", "type": "STRING",
"value": "]]>" "value": "]"
},
{
"type": "STRING",
"value": "]>"
} }
] ]
}, },
@ -1129,23 +1133,27 @@
"value": "\"" "value": "\""
}, },
{ {
"type": "REPEAT", "type": "FIELD",
"name": "content",
"content": { "content": {
"type": "CHOICE", "type": "REPEAT",
"members": [ "content": {
{ "type": "CHOICE",
"type": "PATTERN", "members": [
"value": "[^<%&\"]" {
}, "type": "PATTERN",
{ "value": "[^<%&\"]"
"type": "SYMBOL", },
"name": "PEReference" {
}, "type": "SYMBOL",
{ "name": "PEReference"
"type": "SYMBOL", },
"name": "_Reference" {
} "type": "SYMBOL",
] "name": "_Reference"
}
]
}
} }
}, },
{ {
@ -1162,23 +1170,27 @@
"value": "'" "value": "'"
}, },
{ {
"type": "REPEAT", "type": "FIELD",
"name": "content",
"content": { "content": {
"type": "CHOICE", "type": "REPEAT",
"members": [ "content": {
{ "type": "CHOICE",
"type": "PATTERN", "members": [
"value": "[^<%&']" {
}, "type": "PATTERN",
{ "value": "[^<%&']"
"type": "SYMBOL", },
"name": "PEReference" {
}, "type": "SYMBOL",
{ "name": "PEReference"
"type": "SYMBOL", },
"name": "_Reference" {
} "type": "SYMBOL",
] "name": "_Reference"
}
]
}
} }
}, },
{ {
@ -1305,7 +1317,7 @@
}, },
"_S": { "_S": {
"type": "PATTERN", "type": "PATTERN",
"value": "[\\x20\\x09\\x0D\\x0A]+" "value": "[ \\t\\r\\n]+"
}, },
"Name": { "Name": {
"type": "PATTERN", "type": "PATTERN",
@ -1395,19 +1407,23 @@
"value": "\"" "value": "\""
}, },
{ {
"type": "REPEAT", "type": "FIELD",
"name": "content",
"content": { "content": {
"type": "CHOICE", "type": "REPEAT",
"members": [ "content": {
{ "type": "CHOICE",
"type": "PATTERN", "members": [
"value": "[^<&\"]" {
}, "type": "PATTERN",
{ "value": "[^<&\"]"
"type": "SYMBOL", },
"name": "_Reference" {
} "type": "SYMBOL",
] "name": "_Reference"
}
]
}
} }
}, },
{ {
@ -1424,19 +1440,23 @@
"value": "'" "value": "'"
}, },
{ {
"type": "REPEAT", "type": "FIELD",
"name": "content",
"content": { "content": {
"type": "CHOICE", "type": "REPEAT",
"members": [ "content": {
{ "type": "CHOICE",
"type": "PATTERN", "members": [
"value": "[^<&']" {
}, "type": "PATTERN",
{ "value": "[^<&']"
"type": "SYMBOL", },
"name": "_Reference" {
} "type": "SYMBOL",
] "name": "_Reference"
}
]
}
} }
}, },
{ {
@ -1605,7 +1625,7 @@
"name": "content", "name": "content",
"content": { "content": {
"type": "PATTERN", "type": "PATTERN",
"value": "[ \\r\\na-zA-Z0-9-'()+,./:=?;!*#@$_%]*" "value": "[ \\r\\na-zA-Z0-9\\-'()+,./:=?;!*#@$_%]*"
} }
}, },
{ {
@ -1634,7 +1654,7 @@
"name": "content", "name": "content",
"content": { "content": {
"type": "PATTERN", "type": "PATTERN",
"value": "[ \\r\\na-zA-Z0-9-()+,./:=?;!*#@$_%]*" "value": "[ \\r\\na-zA-Z0-9\\-()+,./:=?;!*#@$_%]*"
} }
}, },
{ {

56
tree-sitter-dtd/src/node-types.json generated
Unescape Escape View File

@ -119,16 +119,17 @@
{ {
"type": "AttValue", "type": "AttValue",
"named": true, "named": true,
"fields": {}, "fields": {
"children": { "content": {
"multiple": true, "multiple": true,
"required": false, "required": false,
"types": [ "types": [
{ {
"type": "_Reference", "type": "_Reference",
"named": true "named": true
} }
] ]
}
} }
}, },
{ {
@ -192,20 +193,21 @@
{ {
"type": "EntityValue", "type": "EntityValue",
"named": true, "named": true,
"fields": {}, "fields": {
"children": { "content": {
"multiple": true, "multiple": true,
"required": false, "required": false,
"types": [ "types": [
{ {
"type": "PEReference", "type": "PEReference",
"named": true "named": true
}, },
{ {
"type": "_Reference", "type": "_Reference",
"named": true "named": true
} }
] ]
}
} }
}, },
{ {
@ -746,7 +748,11 @@
"named": false "named": false
}, },
{ {
"type": "]]>", "type": "]",
"named": false
},
{
"type": "]>",
"named": false "named": false
}, },
{ {

4403
tree-sitter-dtd/src/parser.c generated
View File

File diff suppressed because it is too large Load Diff

57
tree-sitter-xml/grammar.js
Unescape Escape View File

@ -2,7 +2,9 @@
* @file Tree-sitter grammar definition for XML * @file Tree-sitter grammar definition for XML
* @author ObserverOfTime * @author ObserverOfTime
* @license MIT * @license MIT
* @see {@link https://www.w3.org/TR/xml/|W3C standard} * @see {@link https://www.w3.org/TR/xml/|XML standard}
* @see {@link https://www.w3.org/TR/xml-model/|XML model}
* @see {@link https://www.w3.org/TR/xml-stylesheet/|XML stylesheet}
*/ */
const c = require('../common'); const c = require('../common');
@ -20,23 +22,34 @@ module.exports = grammar(DTD, {
], ],
rules: { rules: {
document: $ => seq( document: $ => prec(2, seq(
optional($._S),
optional($.prolog), optional($.prolog),
repeat($._Misc),
field('root', $.element), field('root', $.element),
repeat($._Misc), repeat($._Misc),
), )),
prolog: $ => prec.right(choice( prolog: $ => prec.right(choice(
$.XMLDecl, seq(
$.XMLDecl,
repeat($._Misc),
),
seq( seq(
O($.XMLDecl), O($.XMLDecl),
repeat($._Misc), repeat($._Misc),
$.doctypedecl $.doctypedecl,
) repeat($._Misc),
),
repeat1($._Misc)
)), )),
_Misc: $ => choice($.PI, $.Comment, $._S), _Misc: $ => choice(
$.PI,
$.StyleSheetPI,
$.XmlModelPI,
$.Comment,
$._S
),
XMLDecl: $ => seq( XMLDecl: $ => seq(
'<?', '<?',
@ -56,7 +69,8 @@ module.exports = grammar(DTD, {
), ),
doctypedecl: $ => seq( doctypedecl: $ => seq(
'<!DOCTYPE', '<!',
'DOCTYPE',
$._S, $._S,
$.Name, $.Name,
O(seq($._S, $.ExternalID)), O(seq($._S, $.ExternalID)),
@ -118,6 +132,29 @@ module.exports = grammar(DTD, {
seq('<![CDATA[', $.CData, ']]>') seq('<![CDATA[', $.CData, ']]>')
), ),
CData: _ => /([^\]]|][^\]]|]][^>])*/ CData: _ => /([^\]]|][^\]]|]][^>])*/,
StyleSheetPI: $ => seq(
'<?',
'xml-stylesheet',
c.rseq($._S, $.PseudoAtt),
O($._S),
'?>'
),
XmlModelPI: $ => seq(
'<?',
'xml-model',
c.rseq($._S, $.PseudoAtt),
O($._S),
'?>'
),
PseudoAtt: $ => seq($.Name, $._Eq, $.PseudoAttValue),
PseudoAttValue: $ => choice(
c.att_value($, '"'),
c.att_value($, "'")
)
} }
}); });

619
tree-sitter-xml/src/grammar.json generated
Unescape Escape View File

@ -3,43 +3,52 @@
"word": "Name", "word": "Name",
"rules": { "rules": {
"document": { "document": {
"type": "SEQ", "type": "PREC",
"members": [ "value": 2,
{ "content": {
"type": "CHOICE", "type": "SEQ",
"members": [ "members": [
{ {
"type": "CHOICE",
"members": [
{
"type": "SYMBOL",
"name": "_S"
},
{
"type": "BLANK"
}
]
},
{
"type": "CHOICE",
"members": [
{
"type": "SYMBOL",
"name": "prolog"
},
{
"type": "BLANK"
}
]
},
{
"type": "FIELD",
"name": "root",
"content": {
"type": "SYMBOL", "type": "SYMBOL",
"name": "prolog" "name": "element"
}, }
{ },
"type": "BLANK" {
"type": "REPEAT",
"content": {
"type": "SYMBOL",
"name": "_Misc"
} }
]
},
{
"type": "REPEAT",
"content": {
"type": "SYMBOL",
"name": "_Misc"
}
},
{
"type": "FIELD",
"name": "root",
"content": {
"type": "SYMBOL",
"name": "element"
}
},
{
"type": "REPEAT",
"content": {
"type": "SYMBOL",
"name": "_Misc"
} }
} ]
] }
}, },
"_extSubsetDecl": { "_extSubsetDecl": {
"type": "CHOICE", "type": "CHOICE",
@ -119,7 +128,11 @@
}, },
{ {
"type": "STRING", "type": "STRING",
"value": "]]>" "value": "]"
},
{
"type": "STRING",
"value": "]>"
} }
] ]
}, },
@ -170,7 +183,11 @@
"members": [ "members": [
{ {
"type": "STRING", "type": "STRING",
"value": "<!ELEMENT" "value": "<!"
},
{
"type": "STRING",
"value": "ELEMENT"
}, },
{ {
"type": "SYMBOL", "type": "SYMBOL",
@ -402,24 +419,28 @@
"type": "SEQ", "type": "SEQ",
"members": [ "members": [
{ {
"type": "CHOICE", "type": "SYMBOL",
"members": [ "name": "_choice"
{
"type": "SYMBOL",
"name": "_choice"
},
{
"type": "SYMBOL",
"name": "_seq"
}
]
}, },
{ {
"type": "CHOICE", "type": "CHOICE",
"members": [ "members": [
{ {
"type": "SYMBOL", "type": "CHOICE",
"name": "_occurences" "members": [
{
"type": "STRING",
"value": "?"
},
{
"type": "STRING",
"value": "*"
},
{
"type": "STRING",
"value": "+"
}
]
}, },
{ {
"type": "BLANK" "type": "BLANK"
@ -438,17 +459,8 @@
"type": "CHOICE", "type": "CHOICE",
"members": [ "members": [
{ {
"type": "CHOICE", "type": "SYMBOL",
"members": [ "name": "Name"
{
"type": "SYMBOL",
"name": "Name"
},
{
"type": "SYMBOL",
"name": "PEReference"
}
]
}, },
{ {
"type": "SYMBOL", "type": "SYMBOL",
@ -456,7 +468,7 @@
}, },
{ {
"type": "SYMBOL", "type": "SYMBOL",
"name": "_seq" "name": "PEReference"
} }
] ]
}, },
@ -464,8 +476,21 @@
"type": "CHOICE", "type": "CHOICE",
"members": [ "members": [
{ {
"type": "SYMBOL", "type": "CHOICE",
"name": "_occurences" "members": [
{
"type": "STRING",
"value": "?"
},
{
"type": "STRING",
"value": "*"
},
{
"type": "STRING",
"value": "+"
}
]
}, },
{ {
"type": "BLANK" "type": "BLANK"
@ -499,7 +524,7 @@
"name": "_cp" "name": "_cp"
}, },
{ {
"type": "REPEAT1", "type": "REPEAT",
"content": { "content": {
"type": "SEQ", "type": "SEQ",
"members": [ "members": [
@ -515,79 +540,19 @@
} }
] ]
}, },
{
"type": "STRING",
"value": "|"
},
{ {
"type": "CHOICE", "type": "CHOICE",
"members": [ "members": [
{ {
"type": "SYMBOL", "type": "STRING",
"name": "_S" "value": "|"
},
{
"type": "BLANK"
}
]
},
{
"type": "SYMBOL",
"name": "_cp"
}
]
}
},
{
"type": "STRING",
"value": ")"
}
]
},
"_seq": {
"type": "SEQ",
"members": [
{
"type": "STRING",
"value": "("
},
{
"type": "CHOICE",
"members": [
{
"type": "SYMBOL",
"name": "_S"
},
{
"type": "BLANK"
}
]
},
{
"type": "SYMBOL",
"name": "_cp"
},
{
"type": "REPEAT",
"content": {
"type": "SEQ",
"members": [
{
"type": "CHOICE",
"members": [
{
"type": "SYMBOL",
"name": "_S"
}, },
{ {
"type": "BLANK" "type": "STRING",
"value": ","
} }
] ]
}, },
{
"type": "STRING",
"value": ","
},
{ {
"type": "CHOICE", "type": "CHOICE",
"members": [ "members": [
@ -618,7 +583,11 @@
"members": [ "members": [
{ {
"type": "STRING", "type": "STRING",
"value": "<!ATTLIST" "value": "<!"
},
{
"type": "STRING",
"value": "ATTLIST"
}, },
{ {
"type": "SYMBOL", "type": "SYMBOL",
@ -1040,7 +1009,11 @@
"members": [ "members": [
{ {
"type": "STRING", "type": "STRING",
"value": "<!ENTITY" "value": "<!"
},
{
"type": "STRING",
"value": "ENTITY"
}, },
{ {
"type": "SYMBOL", "type": "SYMBOL",
@ -1116,7 +1089,11 @@
"members": [ "members": [
{ {
"type": "STRING", "type": "STRING",
"value": "<!ENTITY" "value": "<!"
},
{
"type": "STRING",
"value": "ENTITY"
}, },
{ {
"type": "SYMBOL", "type": "SYMBOL",
@ -1180,23 +1157,27 @@
"value": "\"" "value": "\""
}, },
{ {
"type": "REPEAT", "type": "FIELD",
"name": "content",
"content": { "content": {
"type": "CHOICE", "type": "REPEAT",
"members": [ "content": {
{ "type": "CHOICE",
"type": "PATTERN", "members": [
"value": "[^<%&\"]" {
}, "type": "PATTERN",
{ "value": "[^<%&\"]"
"type": "SYMBOL", },
"name": "PEReference" {
}, "type": "SYMBOL",
{ "name": "PEReference"
"type": "SYMBOL", },
"name": "_Reference" {
} "type": "SYMBOL",
] "name": "_Reference"
}
]
}
} }
}, },
{ {
@ -1213,23 +1194,27 @@
"value": "'" "value": "'"
}, },
{ {
"type": "REPEAT", "type": "FIELD",
"name": "content",
"content": { "content": {
"type": "CHOICE", "type": "REPEAT",
"members": [ "content": {
{ "type": "CHOICE",
"type": "PATTERN", "members": [
"value": "[^<%&']" {
}, "type": "PATTERN",
{ "value": "[^<%&']"
"type": "SYMBOL", },
"name": "PEReference" {
}, "type": "SYMBOL",
{ "name": "PEReference"
"type": "SYMBOL", },
"name": "_Reference" {
} "type": "SYMBOL",
] "name": "_Reference"
}
]
}
} }
}, },
{ {
@ -1275,7 +1260,11 @@
"members": [ "members": [
{ {
"type": "STRING", "type": "STRING",
"value": "<!NOTATION" "value": "<!"
},
{
"type": "STRING",
"value": "NOTATION"
}, },
{ {
"type": "SYMBOL", "type": "SYMBOL",
@ -1346,30 +1335,13 @@
} }
] ]
}, },
"_occurences": {
"type": "CHOICE",
"members": [
{
"type": "STRING",
"value": "?"
},
{
"type": "STRING",
"value": "*"
},
{
"type": "STRING",
"value": "+"
}
]
},
"_Char": { "_Char": {
"type": "PATTERN", "type": "PATTERN",
"value": "[\\u0001-\\uD7FF\\uE000-\\uFFFD\\u{10000}-\\u{10FFFF}]" "value": "[\\u0001-\\uD7FF\\uE000-\\uFFFD\\u{10000}-\\u{10FFFF}]"
}, },
"_S": { "_S": {
"type": "PATTERN", "type": "PATTERN",
"value": "[\\x20\\x09\\x0D\\x0A]+" "value": "[ \\t\\r\\n]+"
}, },
"Name": { "Name": {
"type": "PATTERN", "type": "PATTERN",
@ -1459,19 +1431,23 @@
"value": "\"" "value": "\""
}, },
{ {
"type": "REPEAT", "type": "FIELD",
"name": "content",
"content": { "content": {
"type": "CHOICE", "type": "REPEAT",
"members": [ "content": {
{ "type": "CHOICE",
"type": "PATTERN", "members": [
"value": "[^<&\"]" {
}, "type": "PATTERN",
{ "value": "[^<&\"]"
"type": "SYMBOL", },
"name": "_Reference" {
} "type": "SYMBOL",
] "name": "_Reference"
}
]
}
} }
}, },
{ {
@ -1488,19 +1464,23 @@
"value": "'" "value": "'"
}, },
{ {
"type": "REPEAT", "type": "FIELD",
"name": "content",
"content": { "content": {
"type": "CHOICE", "type": "REPEAT",
"members": [ "content": {
{ "type": "CHOICE",
"type": "PATTERN", "members": [
"value": "[^<&']" {
}, "type": "PATTERN",
{ "value": "[^<&']"
"type": "SYMBOL", },
"name": "_Reference" {
} "type": "SYMBOL",
] "name": "_Reference"
}
]
}
} }
}, },
{ {
@ -1669,7 +1649,7 @@
"name": "content", "name": "content",
"content": { "content": {
"type": "PATTERN", "type": "PATTERN",
"value": "[ \\r\\na-zA-Z0-9-'()+,./:=?;!*#@$_%]*" "value": "[ \\r\\na-zA-Z0-9\\-'()+,./:=?;!*#@$_%]*"
} }
}, },
{ {
@ -1698,7 +1678,7 @@
"name": "content", "name": "content",
"content": { "content": {
"type": "PATTERN", "type": "PATTERN",
"value": "[ \\r\\na-zA-Z0-9-()+,./:=?;!*#@$_%]*" "value": "[ \\r\\na-zA-Z0-9\\-()+,./:=?;!*#@$_%]*"
} }
}, },
{ {
@ -1994,8 +1974,20 @@
"type": "CHOICE", "type": "CHOICE",
"members": [ "members": [
{ {
"type": "SYMBOL", "type": "SEQ",
"name": "XMLDecl" "members": [
{
"type": "SYMBOL",
"name": "XMLDecl"
},
{
"type": "REPEAT",
"content": {
"type": "SYMBOL",
"name": "_Misc"
}
}
]
}, },
{ {
"type": "SEQ", "type": "SEQ",
@ -2022,8 +2014,22 @@
{ {
"type": "SYMBOL", "type": "SYMBOL",
"name": "doctypedecl" "name": "doctypedecl"
},
{
"type": "REPEAT",
"content": {
"type": "SYMBOL",
"name": "_Misc"
}
} }
] ]
},
{
"type": "REPEAT1",
"content": {
"type": "SYMBOL",
"name": "_Misc"
}
} }
] ]
} }
@ -2035,6 +2041,14 @@
"type": "SYMBOL", "type": "SYMBOL",
"name": "PI" "name": "PI"
}, },
{
"type": "SYMBOL",
"name": "StyleSheetPI"
},
{
"type": "SYMBOL",
"name": "XmlModelPI"
},
{ {
"type": "SYMBOL", "type": "SYMBOL",
"name": "Comment" "name": "Comment"
@ -2124,7 +2138,11 @@
"members": [ "members": [
{ {
"type": "STRING", "type": "STRING",
"value": "<!DOCTYPE" "value": "<!"
},
{
"type": "STRING",
"value": "DOCTYPE"
}, },
{ {
"type": "SYMBOL", "type": "SYMBOL",
@ -2468,6 +2486,184 @@
"CData": { "CData": {
"type": "PATTERN", "type": "PATTERN",
"value": "([^\\]]|][^\\]]|]][^>])*" "value": "([^\\]]|][^\\]]|]][^>])*"
},
"StyleSheetPI": {
"type": "SEQ",
"members": [
{
"type": "STRING",
"value": "<?"
},
{
"type": "STRING",
"value": "xml-stylesheet"
},
{
"type": "REPEAT",
"content": {
"type": "SEQ",
"members": [
{
"type": "SYMBOL",
"name": "_S"
},
{
"type": "SYMBOL",
"name": "PseudoAtt"
}
]
}
},
{
"type": "CHOICE",
"members": [
{
"type": "SYMBOL",
"name": "_S"
},
{
"type": "BLANK"
}
]
},
{
"type": "STRING",
"value": "?>"
}
]
},
"XmlModelPI": {
"type": "SEQ",
"members": [
{
"type": "STRING",
"value": "<?"
},
{
"type": "STRING",
"value": "xml-model"
},
{
"type": "REPEAT",
"content": {
"type": "SEQ",
"members": [
{
"type": "SYMBOL",
"name": "_S"
},
{
"type": "SYMBOL",
"name": "PseudoAtt"
}
]
}
},
{
"type": "CHOICE",
"members": [
{
"type": "SYMBOL",
"name": "_S"
},
{
"type": "BLANK"
}
]
},
{
"type": "STRING",
"value": "?>"
}
]
},
"PseudoAtt": {
"type": "SEQ",
"members": [
{
"type": "SYMBOL",
"name": "Name"
},
{
"type": "SYMBOL",
"name": "_Eq"
},
{
"type": "SYMBOL",
"name": "PseudoAttValue"
}
]
},
"PseudoAttValue": {
"type": "CHOICE",
"members": [
{
"type": "SEQ",
"members": [
{
"type": "STRING",
"value": "\""
},
{
"type": "FIELD",
"name": "content",
"content": {
"type": "REPEAT",
"content": {
"type": "CHOICE",
"members": [
{
"type": "PATTERN",
"value": "[^<&\"]"
},
{
"type": "SYMBOL",
"name": "_Reference"
}
]
}
}
},
{
"type": "STRING",
"value": "\""
}
]
},
{
"type": "SEQ",
"members": [
{
"type": "STRING",
"value": "'"
},
{
"type": "FIELD",
"name": "content",
"content": {
"type": "REPEAT",
"content": {
"type": "CHOICE",
"members": [
{
"type": "PATTERN",
"value": "[^<&']"
},
{
"type": "SYMBOL",
"name": "_Reference"
}
]
}
}
},
{
"type": "STRING",
"value": "'"
}
]
}
]
} }
}, },
"extras": [], "extras": [],
@ -2475,7 +2671,6 @@
"precedences": [], "precedences": [],
"externals": [], "externals": [],
"inline": [ "inline": [
"_occurences",
"_extSubsetDecl", "_extSubsetDecl",
"conditionalSect" "conditionalSect"
], ],

179
tree-sitter-xml/src/node-types.json generated
Unescape Escape View File

@ -119,16 +119,17 @@
{ {
"type": "AttValue", "type": "AttValue",
"named": true, "named": true,
"fields": {}, "fields": {
"children": { "content": {
"multiple": true, "multiple": true,
"required": false, "required": false,
"types": [ "types": [
{ {
"type": "_Reference", "type": "_Reference",
"named": true "named": true
} }
] ]
}
} }
}, },
{ {
@ -260,20 +261,21 @@
{ {
"type": "EntityValue", "type": "EntityValue",
"named": true, "named": true,
"fields": {}, "fields": {
"children": { "content": {
"multiple": true, "multiple": true,
"required": false, "required": false,
"types": [ "types": [
{ {
"type": "PEReference", "type": "PEReference",
"named": true "named": true
}, },
{ {
"type": "_Reference", "type": "_Reference",
"named": true "named": true
} }
] ]
}
} }
}, },
{ {
@ -478,6 +480,41 @@
] ]
} }
}, },
{
"type": "PseudoAtt",
"named": true,
"fields": {},
"children": {
"multiple": true,
"required": true,
"types": [
{
"type": "Name",
"named": true
},
{
"type": "PseudoAttValue",
"named": true
}
]
}
},
{
"type": "PseudoAttValue",
"named": true,
"fields": {
"content": {
"multiple": true,
"required": false,
"types": [
{
"type": "_Reference",
"named": true
}
]
}
}
},
{ {
"type": "PubidLiteral", "type": "PubidLiteral",
"named": true, "named": true,
@ -521,6 +558,21 @@
] ]
} }
}, },
{
"type": "StyleSheetPI",
"named": true,
"fields": {},
"children": {
"multiple": true,
"required": false,
"types": [
{
"type": "PseudoAtt",
"named": true
}
]
}
},
{ {
"type": "SystemLiteral", "type": "SystemLiteral",
"named": true, "named": true,
@ -545,6 +597,21 @@
] ]
} }
}, },
{
"type": "XmlModelPI",
"named": true,
"fields": {},
"children": {
"multiple": true,
"required": false,
"types": [
{
"type": "PseudoAtt",
"named": true
}
]
}
},
{ {
"type": "children", "type": "children",
"named": true, "named": true,
@ -676,6 +743,14 @@
"type": "PI", "type": "PI",
"named": true "named": true
}, },
{
"type": "StyleSheetPI",
"named": true
},
{
"type": "XmlModelPI",
"named": true
},
{ {
"type": "prolog", "type": "prolog",
"named": true "named": true
@ -739,7 +814,7 @@
"fields": {}, "fields": {},
"children": { "children": {
"multiple": true, "multiple": true,
"required": true, "required": false,
"types": [ "types": [
{ {
"type": "Comment", "type": "Comment",
@ -749,10 +824,18 @@
"type": "PI", "type": "PI",
"named": true "named": true
}, },
{
"type": "StyleSheetPI",
"named": true
},
{ {
"type": "XMLDecl", "type": "XMLDecl",
"named": true "named": true
}, },
{
"type": "XmlModelPI",
"named": true
},
{ {
"type": "doctypedecl", "type": "doctypedecl",
"named": true "named": true
@ -833,23 +916,7 @@
"named": false "named": false
}, },
{ {
"type": "<!ATTLIST", "type": "<!",
"named": false
},
{
"type": "<!DOCTYPE",
"named": false
},
{
"type": "<!ELEMENT",
"named": false
},
{
"type": "<!ENTITY",
"named": false
},
{
"type": "<!NOTATION",
"named": false "named": false
}, },
{ {
@ -888,6 +955,10 @@
"type": "ANY", "type": "ANY",
"named": false "named": false
}, },
{
"type": "ATTLIST",
"named": false
},
{ {
"type": "CData", "type": "CData",
"named": true "named": true
@ -900,10 +971,22 @@
"type": "Comment", "type": "Comment",
"named": true "named": true
}, },
{
"type": "DOCTYPE",
"named": false
},
{
"type": "ELEMENT",
"named": false
},
{ {
"type": "EMPTY", "type": "EMPTY",
"named": false "named": false
}, },
{
"type": "ENTITY",
"named": false
},
{ {
"type": "EncName", "type": "EncName",
"named": true "named": true
@ -964,6 +1047,10 @@
"type": "]", "type": "]",
"named": false "named": false
}, },
{
"type": "]>",
"named": false
},
{ {
"type": "]]>", "type": "]]>",
"named": false "named": false
@ -988,6 +1075,14 @@
"type": "xml", "type": "xml",
"named": false "named": false
}, },
{
"type": "xml-model",
"named": false
},
{
"type": "xml-stylesheet",
"named": false
},
{ {
"type": "yes", "type": "yes",
"named": false "named": false

9002
tree-sitter-xml/src/parser.c generated
View File

File diff suppressed because it is too large Load Diff

119
tree-sitter-xml/test/corpus/extra.txt
Unescape Escape View File

@ -0,0 +1,119 @@
================================================================================
XML Model
================================================================================
<?xml version="1.0"?>
<?xml-model href="http://www.docbook.org/xml/5.0/rng/docbook.rng"?>
<?xml-model href="http://www.docbook.org/xml/5.0/xsd/docbook.xsd"?>
<book xmlns="http://docbook.org/ns/docbook">
…
</book>
--------------------------------------------------------------------------------
(document
(prolog
(XMLDecl
(VersionNum))
(XmlModelPI
(PseudoAtt
(Name)
(PseudoAttValue)))
(XmlModelPI
(PseudoAtt
(Name)
(PseudoAttValue))))
(element
(STag
(Name)
(Attribute
(Name)
(AttValue)))
(content
(CharData))
(ETag
(Name))))
================================================================================
XML Stylesheet
================================================================================
<?xml-stylesheet href="common.css"?>
<?xml-stylesheet href="default.css" title="Default style"?>
<?xml-stylesheet alternate="yes" href="alt.css" title="Alternative style"?>
<?xml-stylesheet href="single-col.css" media="all and (max-width: 30em)"?>
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<title>Example with xml-stylesheet processing instructions</title>
</head>
<body>
...
</body>
</html>
--------------------------------------------------------------------------------
(document
(prolog
(StyleSheetPI
(PseudoAtt
(Name)
(PseudoAttValue)))
(StyleSheetPI
(PseudoAtt
(Name)
(PseudoAttValue))
(PseudoAtt
(Name)
(PseudoAttValue)))
(StyleSheetPI
(PseudoAtt
(Name)
(PseudoAttValue))
(PseudoAtt
(Name)
(PseudoAttValue))
(PseudoAtt
(Name)
(PseudoAttValue)))
(StyleSheetPI
(PseudoAtt
(Name)
(PseudoAttValue))
(PseudoAtt
(Name)
(PseudoAttValue))))
(element
(STag
(Name)
(Attribute
(Name)
(AttValue)))
(content
(CharData)
(element
(STag
(Name))
(content
(CharData)
(element
(STag
(Name))
(content
(CharData))
(ETag
(Name)))
(CharData))
(ETag
(Name)))
(CharData)
(element
(STag
(Name))
(content
(CharData))
(ETag
(Name)))
(CharData))
(ETag
(Name))))

3
tree-sitter-xml/test/corpus/standard.txt
Unescape Escape View File

@ -39,7 +39,8 @@ Comment
-------------------------------------------------------------------------------- --------------------------------------------------------------------------------
(document (document
(Comment) (prolog
(Comment))
(element (element
(EmptyElemTag (EmptyElemTag
(Name)))) (Name))))