Public-Mirrors
/
markdown-it
mirror of https://github.com/markdown-it/markdown-it.git
1
0
Fork 0
Code Issues Releases Activity
Browse Source

Reworked html parse in block lexer + added `html` option

pull/14/head
Vitaly Puzrin 10 years ago
parent
e6341a4812
commit
a8e6bab747
8 changed files with 315 additions and 46 deletions
Whitespace
Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL
Unified View
Diff Options
Show Stats Download Patch File Download Diff File
  1. 3
      bin/specsplit.js
  2. 58
      lib/common/html_blocks.js
  3. 48
      lib/common/html_re.js
  4. 172
      lib/common/url_schemas.js
  5. 2
      lib/lexer_block.js
  6. 75
      lib/lexer_block/htmlblock.js
  7. 2
      lib/lexer_block/paragraph.js
  8. 1
      test/stmd.js

3
bin/specsplit.js
View File

@ -58,7 +58,8 @@ function readFile(filename, encoding, callback) {
readFile(options.spec, 'utf8', function (error, input) { readFile(options.spec, 'utf8', function (error, input) {
var good = [], bad = [], var good = [], bad = [],
markdown = new Remarkable({ markdown = new Remarkable({
xhtml: true, // write <hr /> instead of <hr> and so on html: true,
xhtml: true,
codeLangPrefix: 'language-' codeLangPrefix: 'language-'
}); });

58
lib/common/html_blocks.js
View File

@ -0,0 +1,58 @@
// List of valid html blocks names, accorting to stmd spec
// http://jgm.github.io/stmd/spec.html#html-blocks
'use strict';
module.exports = [
'article',
'aside',
'button',
'blockquote',
'body',
'canvas',
'caption',
'col',
'colgroup',
'dd',
'div',
'dl',
'dt',
'embed',
'fieldset',
'figcaption',
'figure',
'footer',
'form',
'h1',
'h2',
'h3',
'h4',
'h5',
'h6',
'header',
'hgroup',
'hr',
'iframe',
'li',
'map',
'object',
'ol',
'output',
'p',
'pre',
'progress',
'script',
'section',
'style',
'table',
'tbody',
'td',
'textarea',
'tfoot',
'th',
'tr',
'thead',
'ul',
'video'
];

48
lib/lexer_block/html.js → lib/common/html_re.js
View File

@ -1,12 +1,8 @@
// HTML block // Regexps to match html elements
'use strict'; 'use strict';
var isEmpty = require('../helpers').isEmpty;
var getLines = require('../helpers').getLines;
function replace(regex, options) { function replace(regex, options) {
regex = regex.source; regex = regex.source;
options = options || ''; options = options || '';
@ -16,14 +12,14 @@ function replace(regex, options) {
return new RegExp(regex, options); return new RegExp(regex, options);
} }
val = val.source || val; val = val.source || val;
val = val.replace(/(^|[^\[])\^/g, '$1'); //val = val.replace(/(^|[^\[])\^/g, '$1');
regex = regex.replace(name, val); regex = regex.replace(name, val);
return self; return self;
}; };
} }
var attr_name = /[a-zA-Z_:][a-zA-Z0-9:._-]*/; var attr_name = /[a-zA-Z_:][a-zA-Z0-9:._-]*/;
var unquoted = /[^"'=<>`\x00-\x20]+/; var unquoted = /[^"'=<>`\x00-\x20]+/;
var single_quoted = /'[^']*'/; var single_quoted = /'[^']*'/;
@ -51,7 +47,7 @@ var processing = /<[?].*?[?]>/;
var declaration = /<![A-Z]\s+[^>]*>/; var declaration = /<![A-Z]\s+[^>]*>/;
var cdata = /<!\[CDATA\[([^\]]+|\][^\]]|\]\][^>])*\]\]>/; var cdata = /<!\[CDATA\[([^\]]+|\][^\]]|\]\][^>])*\]\]>/;
var html_tag = replace(/^(?:open_tag|close_tag|comment|processing|declaration|cdata)/, 'i') var HTML_TAG_RE = replace(/^(?:open_tag|close_tag|comment|processing|declaration|cdata)/, 'i')
('open_tag', open_tag) ('open_tag', open_tag)
('close_tag', close_tag) ('close_tag', close_tag)
('comment', comment) ('comment', comment)
@ -61,38 +57,4 @@ var html_tag = replace(/^(?:open_tag|close_tag|comment|processing|declaration|cd
(); ();
module.exports = function html(state, startLine, endLine, silent) { module.exports.HTML_TAG_RE = HTML_TAG_RE;
var nextLine,
pos = state.bMarks[startLine],
max = state.eMarks[startLine],
shift = state.tShift[startLine];
pos += shift;
if (pos + 3 >= max ||
shift > 3 ||
state.blkLevel > 0) { return false; }
if (state.src.charCodeAt(pos) !== 0x3C/* < */) { return false; }
// TODO: (?) optimize check.
nextLine = startLine + 1;
while (nextLine < state.lineMax && !isEmpty(state, nextLine)) {
nextLine++;
}
if (!html_tag.test(state.src.slice(pos, state.eMarks[nextLine - 1]).replace(/\n/g,' '))) {
return false;
}
if (silent) { return true; }
state.tokens.push({
type: 'html',
content: getLines(state, startLine, nextLine, 0, true)
});
state.line = nextLine;
return true;
};

172
lib/common/url_schemas.js
View File

@ -0,0 +1,172 @@
// List of valid url schemas, accorting to stmd spec
// http://jgm.github.io/stmd/spec.html#autolinks
'use strict';
module.exports = [
'coap',
'doi',
'javascript',
'aaa',
'aaas',
'about',
'acap',
'cap',
'cid',
'crid',
'data',
'dav',
'dict',
'dns',
'file',
'ftp',
'geo',
'go',
'gopher',
'h323',
'http',
'https',
'iax',
'icap',
'im',
'imap',
'info',
'ipp',
'iris',
'iris.beep',
'iris.xpc',
'iris.xpcs',
'iris.lwz',
'ldap',
'mailto',
'mid',
'msrp',
'msrps',
'mtqp',
'mupdate',
'news',
'nfs',
'ni',
'nih',
'nntp',
'opaquelocktoken',
'pop',
'pres',
'rtsp',
'service',
'session',
'shttp',
'sieve',
'sip',
'sips',
'sms',
'snmp',
'soap.beep',
'soap.beeps',
'tag',
'tel',
'telnet',
'tftp',
'thismessage',
'tn3270',
'tip',
'tv',
'urn',
'vemmi',
'ws',
'wss',
'xcon',
'xcon-userid',
'xmlrpc.beep',
'xmlrpc.beeps',
'xmpp',
'z39.50r',
'z39.50s',
'adiumxtra',
'afp',
'afs',
'aim',
'apt',
'attachment',
'aw',
'beshare',
'bitcoin',
'bolo',
'callto',
'chrome',
'chrome-extension',
'com-eventbrite-attendee',
'content',
'cvs',
'dlna-playsingle',
'dlna-playcontainer',
'dtn',
'dvb',
'ed2k',
'facetime',
'feed',
'finger',
'fish',
'gg',
'git',
'gizmoproject',
'gtalk',
'hcp',
'icon',
'ipn',
'irc',
'irc6',
'ircs',
'itms',
'jar',
'jms',
'keyparc',
'lastfm',
'ldaps',
'magnet',
'maps',
'market',
'message',
'mms',
'ms-help',
'msnim',
'mumble',
'mvn',
'notes',
'oid',
'palm',
'paparazzi',
'platform',
'proxy',
'psyc',
'query',
'res',
'resource',
'rmi',
'rsync',
'rtmp',
'secondlife',
'sftp',
'sgn',
'skype',
'smb',
'soldat',
'spotify',
'ssh',
'steam',
'svn',
'teamspeak',
'things',
'udp',
'unreal',
'ut2004',
'ventrilo',
'view-source',
'webcal',
'wtai',
'wyciwyg',
'xfire',
'xri',
'ymsgr'
];

2
lib/lexer_block.js
View File

@ -19,7 +19,7 @@ rules.push(require('./lexer_block/hr'));
rules.push(require('./lexer_block/list')); rules.push(require('./lexer_block/list'));
rules.push(require('./lexer_block/heading')); rules.push(require('./lexer_block/heading'));
rules.push(require('./lexer_block/lheading')); rules.push(require('./lexer_block/lheading'));
rules.push(require('./lexer_block/html')); rules.push(require('./lexer_block/htmlblock'));
rules.push(require('./lexer_block/table')); rules.push(require('./lexer_block/table'));
rules.push(require('./lexer_block/paragraph')); rules.push(require('./lexer_block/paragraph'));

75
lib/lexer_block/htmlblock.js
View File

@ -0,0 +1,75 @@
// HTML block
'use strict';
var isEmpty = require('../helpers').isEmpty;
var getLines = require('../helpers').getLines;
var block_names = require('../common/html_blocks');
var HTML_TAG_OPEN_RE = /^<([a-zA-Z]{1,15})[\s\/>]/;
var HTML_TAG_CLOSE_RE = /^<\/([a-zA-Z]{1,15})[\s>]/;
function isLetter(ch) {
/*eslint no-bitwise:0*/
var lc = ch | 0x20; // to lower case
return (lc >= 0x61/* a */) && (lc <= 0x7a/* z */);
}
module.exports = function htmlblock(state, startLine, endLine, silent) {
var ch, match, nextLine,
pos = state.bMarks[startLine],
max = state.eMarks[startLine],
shift = state.tShift[startLine];
pos += shift;
if (!state.options.html) { return false; }
if (shift > 3 || pos + 2 >= max || state.blkLevel > 0) { return false; }
if (state.src.charCodeAt(pos) !== 0x3C/* < */) { return false; }
ch = state.src.charCodeAt(pos + 1);
if (ch === 0x21/* ! */ || ch === 0x3F/* ? */) {
// Directive start / comment start / processing instruction start
if (silent) { return true; }
} else if (ch === 0x2F/* / */ || isLetter(ch)) {
// Probably start or end of tag
if (ch === 0x2F/* \ */) {
// closing tag
match = state.src.slice(pos, max).match(HTML_TAG_CLOSE_RE);
if (!match) { return false; }
} else {
// opening tag
match = state.src.slice(pos, max).match(HTML_TAG_OPEN_RE);
if (!match) { return false; }
}
// Make sure tag name is valid
if (block_names.indexOf(match[1]) < 0) { return false; }
if (silent) { return true; }
} else {
return false;
}
// If we are here - we detected HTML block.
// Let's roll down till empty line (block end).
nextLine = startLine + 1;
while (nextLine < state.lineMax && !isEmpty(state, nextLine)) {
nextLine++;
}
state.tokens.push({
type: 'html',
content: getLines(state, startLine, nextLine, 0, true)
});
state.line = nextLine;
return true;
};

2
lib/lexer_block/paragraph.js
View File

@ -28,7 +28,7 @@ module.exports = function paragraph(state, startLine/*, endLine*/) {
// setex header can't interrupt paragraph // setex header can't interrupt paragraph
// if (rules_named.lheading(state, nextLine, endLine, true)) { break; } // if (rules_named.lheading(state, nextLine, endLine, true)) { break; }
if (rules_named.blockquote(state, nextLine, endLine, true)) { break; } if (rules_named.blockquote(state, nextLine, endLine, true)) { break; }
if (rules_named.html(state, nextLine, endLine, true)) { break; } if (rules_named.htmlblock(state, nextLine, endLine, true)) { break; }
if (rules_named.table(state, nextLine, endLine, true)) { break; } if (rules_named.table(state, nextLine, endLine, true)) { break; }
//if (rules_named.tag(state, nextLine, endLine, true)) { break; } //if (rules_named.tag(state, nextLine, endLine, true)) { break; }
//if (rules_named.def(state, nextLine, endLine, true)) { break; } //if (rules_named.def(state, nextLine, endLine, true)) { break; }

1
test/stmd.js
View File

@ -11,6 +11,7 @@ var Remarked = require('../');
describe('stmd', function () { describe('stmd', function () {
var md = new Remarked({ var md = new Remarked({
html: true,
xhtml: true, xhtml: true,
codeLangPrefix: 'language-' codeLangPrefix: 'language-'
}); });

Write Preview
Loading…
Cancel
Save