correct URL pattern used in text length counter in WebUI (#4968)

2 files changed, 199 insertions, 1 deletions

diff --git a/app/javascript/mastodon/features/compose/util/counter.js b/app/javascript/mastodon/features/compose/util/counter.js
index f0fea1a0e..588a372c6 100644
--- a/app/javascript/mastodon/features/compose/util/counter.js
+++ b/app/javascript/mastodon/features/compose/util/counter.js
@@ -1,7 +1,9 @@
+import { urlRegex } from './url_regex';
+
 const urlPlaceholder = 'xxxxxxxxxxxxxxxxxxxxxxx';
 
 export function countableText(inputText) {
   return inputText
-    .replace(/https?:\/\/\S+/g, urlPlaceholder)
+    .replace(urlRegex, urlPlaceholder)
     .replace(/(?:^|[^\/\w])@(([a-z0-9_]+)@[a-z0-9\.\-]+)/ig, '@$2');
 };
diff --git a/app/javascript/mastodon/features/compose/util/url_regex.js b/app/javascript/mastodon/features/compose/util/url_regex.js
new file mode 100644
index 000000000..c15b9696a
--- /dev/null
+++ b/app/javascript/mastodon/features/compose/util/url_regex.js
@@ -0,0 +1,196 @@
+const regexen = {};
+
+const regexSupplant = function(regex, flags) {
+  flags = flags || '';
+  if (typeof regex !== 'string') {
+    if (regex.global && flags.indexOf('g') < 0) {
+      flags += 'g';
+    }
+    if (regex.ignoreCase && flags.indexOf('i') < 0) {
+      flags += 'i';
+    }
+    if (regex.multiline && flags.indexOf('m') < 0) {
+      flags += 'm';
+    }
+
+    regex = regex.source;
+  }
+  return new RegExp(regex.replace(/#\{(\w+)\}/g, function(match, name) {
+    var newRegex = regexen[name] || '';
+    if (typeof newRegex !== 'string') {
+      newRegex = newRegex.source;
+    }
+    return newRegex;
+  }), flags);
+};
+
+const stringSupplant = function(str, values) {
+  return str.replace(/#\{(\w+)\}/g, function(match, name) {
+    return values[name] || '';
+  });
+};
+
+export const urlRegex = (function() {
+  regexen.spaces_group = /\x09-\x0D\x20\x85\xA0\u1680\u180E\u2000-\u200A\u2028\u2029\u202F\u205F\u3000/;
+  regexen.invalid_chars_group = /\uFFFE\uFEFF\uFFFF\u202A-\u202E/;
+  regexen.punct = /\!'#%&'\(\)*\+,\\\-\.\/:;<=>\?@\[\]\^_{|}~\$/;
+  regexen.validUrlPrecedingChars = regexSupplant(/(?:[^A-Za-z0-9@＠$#＃#{invalid_chars_group}]|^)/);
+  regexen.invalidDomainChars = stringSupplant('#{punct}#{spaces_group}#{invalid_chars_group}', regexen);
+  regexen.validDomainChars = regexSupplant(/[^#{invalidDomainChars}]/);
+  regexen.validSubdomain = regexSupplant(/(?:(?:#{validDomainChars}(?:[_-]|#{validDomainChars})*)?#{validDomainChars}\.)/);
+  regexen.validDomainName = regexSupplant(/(?:(?:#{validDomainChars}(?:-|#{validDomainChars})*)?#{validDomainChars}\.)/);
+  regexen.validGTLD = regexSupplant(RegExp(
+  '(?:(?:' +
+    '삼성|닷컴|닷넷|香格里拉|餐厅|食品|飞利浦|電訊盈科|集团|通販|购物|谷歌|诺基亚|联通|网络|网站|网店|网址|组织机构|移动|珠宝|点看|游戏|淡马锡|机构|書籍|时尚|新闻|政府|' +
+    '政务|手表|手机|我爱你|慈善|微博|广东|工行|家電|娱乐|天主教|大拿|大众汽车|在线|嘉里大酒店|嘉里|商标|商店|商城|公益|公司|八卦|健康|信息|佛山|企业|中文网|中信|世界|' +
+    'ポイント|ファッション|セール|ストア|コム|グーグル|クラウド|みんな|คอม|संगठन|नेट|कॉम|همراه|موقع|موبايلي|كوم|كاثوليك|عرب|شبكة|' +
+    'بيتك|بازار|العليان|ارامكو|اتصالات|ابوظبي|קום|сайт|рус|орг|онлайн|москва|ком|католик|дети|' +
+    'zuerich|zone|zippo|zip|zero|zara|zappos|yun|youtube|you|yokohama|yoga|yodobashi|yandex|yamaxun|' +
+    'yahoo|yachts|xyz|xxx|xperia|xin|xihuan|xfinity|xerox|xbox|wtf|wtc|wow|world|works|work|woodside|' +
+    'wolterskluwer|wme|winners|wine|windows|win|williamhill|wiki|wien|whoswho|weir|weibo|wedding|wed|' +
+    'website|weber|webcam|weatherchannel|weather|watches|watch|warman|wanggou|wang|walter|walmart|' +
+    'wales|vuelos|voyage|voto|voting|vote|volvo|volkswagen|vodka|vlaanderen|vivo|viva|vistaprint|' +
+    'vista|vision|visa|virgin|vip|vin|villas|viking|vig|video|viajes|vet|versicherung|' +
+    'vermögensberatung|vermögensberater|verisign|ventures|vegas|vanguard|vana|vacations|ups|uol|uno|' +
+    'university|unicom|uconnect|ubs|ubank|tvs|tushu|tunes|tui|tube|trv|trust|travelersinsurance|' +
+    'travelers|travelchannel|travel|training|trading|trade|toys|toyota|town|tours|total|toshiba|' +
+    'toray|top|tools|tokyo|today|tmall|tkmaxx|tjx|tjmaxx|tirol|tires|tips|tiffany|tienda|tickets|' +
+    'tiaa|theatre|theater|thd|teva|tennis|temasek|telefonica|telecity|tel|technology|tech|team|tdk|' +
+    'tci|taxi|tax|tattoo|tatar|tatamotors|target|taobao|talk|taipei|tab|systems|symantec|sydney|' +
+    'swiss|swiftcover|swatch|suzuki|surgery|surf|support|supply|supplies|sucks|style|study|studio|' +
+    'stream|store|storage|stockholm|stcgroup|stc|statoil|statefarm|statebank|starhub|star|staples|' +
+    'stada|srt|srl|spreadbetting|spot|spiegel|space|soy|sony|song|solutions|solar|sohu|software|' +
+    'softbank|social|soccer|sncf|smile|smart|sling|skype|sky|skin|ski|site|singles|sina|silk|shriram|' +
+    'showtime|show|shouji|shopping|shop|shoes|shiksha|shia|shell|shaw|sharp|shangrila|sfr|sexy|sex|' +
+    'sew|seven|ses|services|sener|select|seek|security|secure|seat|search|scot|scor|scjohnson|' +
+    'science|schwarz|schule|school|scholarships|schmidt|schaeffler|scb|sca|sbs|sbi|saxo|save|sas|' +
+    'sarl|sapo|sap|sanofi|sandvikcoromant|sandvik|samsung|samsclub|salon|sale|sakura|safety|safe|' +
+    'saarland|ryukyu|rwe|run|ruhr|rugby|rsvp|room|rogers|rodeo|rocks|rocher|rmit|rip|rio|ril|' +
+    'rightathome|ricoh|richardli|rich|rexroth|reviews|review|restaurant|rest|republican|report|' +
+    'repair|rentals|rent|ren|reliance|reit|reisen|reise|rehab|redumbrella|redstone|red|recipes|' +
+    'realty|realtor|realestate|read|raid|radio|racing|qvc|quest|quebec|qpon|pwc|pub|prudential|pru|' +
+    'protection|property|properties|promo|progressive|prof|productions|prod|pro|prime|press|praxi|' +
+    'pramerica|post|porn|politie|poker|pohl|pnc|plus|plumbing|playstation|play|place|pizza|pioneer|' +
+    'pink|ping|pin|pid|pictures|pictet|pics|piaget|physio|photos|photography|photo|phone|philips|phd|' +
+    'pharmacy|pfizer|pet|pccw|pay|passagens|party|parts|partners|pars|paris|panerai|panasonic|' +
+    'pamperedchef|page|ovh|ott|otsuka|osaka|origins|orientexpress|organic|org|orange|oracle|open|ooo|' +
+    'onyourside|online|onl|ong|one|omega|ollo|oldnavy|olayangroup|olayan|okinawa|office|off|observer|' +
+    'obi|nyc|ntt|nrw|nra|nowtv|nowruz|now|norton|northwesternmutual|nokia|nissay|nissan|ninja|nikon|' +
+    'nike|nico|nhk|ngo|nfl|nexus|nextdirect|next|news|newholland|new|neustar|network|netflix|netbank|' +
+    'net|nec|nba|navy|natura|nationwide|name|nagoya|nadex|nab|mutuelle|mutual|museum|mtr|mtpc|mtn|' +
+    'msd|movistar|movie|mov|motorcycles|moto|moscow|mortgage|mormon|mopar|montblanc|monster|money|' +
+    'monash|mom|moi|moe|moda|mobily|mobile|mobi|mma|mls|mlb|mitsubishi|mit|mint|mini|mil|microsoft|' +
+    'miami|metlife|merckmsd|meo|menu|men|memorial|meme|melbourne|meet|media|med|mckinsey|mcdonalds|' +
+    'mcd|mba|mattel|maserati|marshalls|marriott|markets|marketing|market|map|mango|management|man|' +
+    'makeup|maison|maif|madrid|macys|luxury|luxe|lupin|lundbeck|ltda|ltd|lplfinancial|lpl|love|lotto|' +
+    'lotte|london|lol|loft|locus|locker|loans|loan|lixil|living|live|lipsy|link|linde|lincoln|limo|' +
+    'limited|lilly|like|lighting|lifestyle|lifeinsurance|life|lidl|liaison|lgbt|lexus|lego|legal|' +
+    'lefrak|leclerc|lease|lds|lawyer|law|latrobe|latino|lat|lasalle|lanxess|landrover|land|lancome|' +
+    'lancia|lancaster|lamer|lamborghini|ladbrokes|lacaixa|kyoto|kuokgroup|kred|krd|kpn|kpmg|kosher|' +
+    'komatsu|koeln|kiwi|kitchen|kindle|kinder|kim|kia|kfh|kerryproperties|kerrylogistics|kerryhotels|' +
+    'kddi|kaufen|juniper|juegos|jprs|jpmorgan|joy|jot|joburg|jobs|jnj|jmp|jll|jlc|jio|jewelry|jetzt|' +
+    'jeep|jcp|jcb|java|jaguar|iwc|iveco|itv|itau|istanbul|ist|ismaili|iselect|irish|ipiranga|' +
+    'investments|intuit|international|intel|int|insure|insurance|institute|ink|ing|info|infiniti|' +
+    'industries|immobilien|immo|imdb|imamat|ikano|iinet|ifm|ieee|icu|ice|icbc|ibm|hyundai|hyatt|' +
+    'hughes|htc|hsbc|how|house|hotmail|hotels|hoteles|hot|hosting|host|hospital|horse|honeywell|' +
+    'honda|homesense|homes|homegoods|homedepot|holiday|holdings|hockey|hkt|hiv|hitachi|hisamitsu|' +
+    'hiphop|hgtv|hermes|here|helsinki|help|healthcare|health|hdfcbank|hdfc|hbo|haus|hangout|hamburg|' +
+    'hair|guru|guitars|guide|guge|gucci|guardian|group|grocery|gripe|green|gratis|graphics|grainger|' +
+    'gov|got|gop|google|goog|goodyear|goodhands|goo|golf|goldpoint|gold|godaddy|gmx|gmo|gmbh|gmail|' +
+    'globo|global|gle|glass|glade|giving|gives|gifts|gift|ggee|george|genting|gent|gea|gdn|gbiz|' +
+    'garden|gap|games|game|gallup|gallo|gallery|gal|fyi|futbol|furniture|fund|fun|fujixerox|fujitsu|' +
+    'ftr|frontier|frontdoor|frogans|frl|fresenius|free|fox|foundation|forum|forsale|forex|ford|' +
+    'football|foodnetwork|food|foo|fly|flsmidth|flowers|florist|flir|flights|flickr|fitness|fit|' +
+    'fishing|fish|firmdale|firestone|fire|financial|finance|final|film|fido|fidelity|fiat|ferrero|' +
+    'ferrari|feedback|fedex|fast|fashion|farmers|farm|fans|fan|family|faith|fairwinds|fail|fage|' +
+    'extraspace|express|exposed|expert|exchange|everbank|events|eus|eurovision|etisalat|esurance|' +
+    'estate|esq|erni|ericsson|equipment|epson|epost|enterprises|engineering|engineer|energy|emerck|' +
+    'email|education|edu|edeka|eco|eat|earth|dvr|dvag|durban|dupont|duns|dunlop|duck|dubai|dtv|drive|' +
+    'download|dot|doosan|domains|doha|dog|dodge|doctor|docs|dnp|diy|dish|discover|discount|directory|' +
+    'direct|digital|diet|diamonds|dhl|dev|design|desi|dentist|dental|democrat|delta|deloitte|dell|' +
+    'delivery|degree|deals|dealer|deal|dds|dclk|day|datsun|dating|date|data|dance|dad|dabur|cyou|' +
+    'cymru|cuisinella|csc|cruises|cruise|crs|crown|cricket|creditunion|creditcard|credit|courses|' +
+    'coupons|coupon|country|corsica|coop|cool|cookingchannel|cooking|contractors|contact|consulting|' +
+    'construction|condos|comsec|computer|compare|company|community|commbank|comcast|com|cologne|' +
+    'college|coffee|codes|coach|clubmed|club|cloud|clothing|clinique|clinic|click|cleaning|claims|' +
+    'cityeats|city|citic|citi|citadel|cisco|circle|cipriani|church|chrysler|chrome|christmas|chloe|' +
+    'chintai|cheap|chat|chase|channel|chanel|cfd|cfa|cern|ceo|center|ceb|cbs|cbre|cbn|cba|catholic|' +
+    'catering|cat|casino|cash|caseih|case|casa|cartier|cars|careers|career|care|cards|caravan|car|' +
+    'capitalone|capital|capetown|canon|cancerresearch|camp|camera|cam|calvinklein|call|cal|cafe|cab|' +
+    'bzh|buzz|buy|business|builders|build|bugatti|budapest|brussels|brother|broker|broadway|' +
+    'bridgestone|bradesco|box|boutique|bot|boston|bostik|bosch|boots|booking|book|boo|bond|bom|bofa|' +
+    'boehringer|boats|bnpparibas|bnl|bmw|bms|blue|bloomberg|blog|blockbuster|blanco|blackfriday|' +
+    'black|biz|bio|bingo|bing|bike|bid|bible|bharti|bet|bestbuy|best|berlin|bentley|beer|beauty|' +
+    'beats|bcn|bcg|bbva|bbt|bbc|bayern|bauhaus|basketball|baseball|bargains|barefoot|barclays|' +
+    'barclaycard|barcelona|bar|bank|band|bananarepublic|banamex|baidu|baby|azure|axa|aws|avianca|' +
+    'autos|auto|author|auspost|audio|audible|audi|auction|attorney|athleta|associates|asia|asda|arte|' +
+    'art|arpa|army|archi|aramco|arab|aquarelle|apple|app|apartments|aol|anz|anquan|android|analytics|' +
+    'amsterdam|amica|amfam|amex|americanfamily|americanexpress|alstom|alsace|ally|allstate|allfinanz|' +
+    'alipay|alibaba|alfaromeo|akdn|airtel|airforce|airbus|aigo|aig|agency|agakhan|africa|afl|' +
+    'afamilycompany|aetna|aero|aeg|adult|ads|adac|actor|active|aco|accountants|accountant|accenture|' +
+    'academy|abudhabi|abogado|able|abc|abbvie|abbott|abb|abarth|aarp|aaa|onion' +
+  ')(?=[^0-9a-zA-Z@]|$))'));
+  regexen.validCCTLD = regexSupplant(RegExp(
+  '(?:(?:' +
+      '한국|香港|澳門|新加坡|台灣|台湾|中國|中国|გე|ไทย|ලංකා|ഭാരതം|ಭಾರತ|భారత్|சிங்கப்பூர்|இலங்கை|இந்தியா|ଭାରତ|ભારત|ਭਾਰਤ|' +
+      'ভাৰত|ভারত|বাংলা|भारोत|भारतम्|भारत|ڀارت|پاکستان|مليسيا|مصر|قطر|فلسطين|عمان|عراق|سورية|سودان|تونس|' +
+      'بھارت|بارت|ایران|امارات|المغرب|السعودية|الجزائر|الاردن|հայ|қаз|укр|срб|рф|мон|мкд|ею|бел|бг|ελ|' +
+      'zw|zm|za|yt|ye|ws|wf|vu|vn|vi|vg|ve|vc|va|uz|uy|us|um|uk|ug|ua|tz|tw|tv|tt|tr|tp|to|tn|tm|tl|tk|' +
+      'tj|th|tg|tf|td|tc|sz|sy|sx|sv|su|st|ss|sr|so|sn|sm|sl|sk|sj|si|sh|sg|se|sd|sc|sb|sa|rw|ru|rs|ro|' +
+      're|qa|py|pw|pt|ps|pr|pn|pm|pl|pk|ph|pg|pf|pe|pa|om|nz|nu|nr|np|no|nl|ni|ng|nf|ne|nc|na|mz|my|mx|' +
+      'mw|mv|mu|mt|ms|mr|mq|mp|mo|mn|mm|ml|mk|mh|mg|mf|me|md|mc|ma|ly|lv|lu|lt|ls|lr|lk|li|lc|lb|la|kz|' +
+      'ky|kw|kr|kp|kn|km|ki|kh|kg|ke|jp|jo|jm|je|it|is|ir|iq|io|in|im|il|ie|id|hu|ht|hr|hn|hm|hk|gy|gw|' +
+      'gu|gt|gs|gr|gq|gp|gn|gm|gl|gi|gh|gg|gf|ge|gd|gb|ga|fr|fo|fm|fk|fj|fi|eu|et|es|er|eh|eg|ee|ec|dz|' +
+      'do|dm|dk|dj|de|cz|cy|cx|cw|cv|cu|cr|co|cn|cm|cl|ck|ci|ch|cg|cf|cd|cc|ca|bz|by|bw|bv|bt|bs|br|bq|' +
+      'bo|bn|bm|bl|bj|bi|bh|bg|bf|be|bd|bb|ba|az|ax|aw|au|at|as|ar|aq|ao|an|am|al|ai|ag|af|ae|ad|ac' +
+  ')(?=[^0-9a-zA-Z@]|$))'));
+  regexen.validPunycode = /(?:xn--[0-9a-z]+)/;
+  regexen.validSpecialCCTLD = /(?:(?:co|tv)(?=[^0-9a-zA-Z@]|$))/;
+  regexen.validDomain = regexSupplant(/(?:#{validSubdomain}*#{validDomainName}(?:#{validGTLD}|#{validCCTLD}|#{validPunycode}))/);
+  regexen.validPortNumber = /[0-9]+/;
+  regexen.pd = /\u002d\u058a\u05be\u1400\u1806\u2010-\u2015\u2e17\u2e1a\u2e3a\u2e40\u301c\u3030\u30a0\ufe31\ufe58\ufe63\uff0d/;
+  regexen.validGeneralUrlPathChars = regexSupplant(/[^#{spaces_group}\(\)\?]/i);
+  // Allow URL paths to contain up to two nested levels of balanced parens
+  //  1. Used in Wikipedia URLs like /Primer_(film)
+  //  2. Used in IIS sessions like /S(dfd346)/
+  //  3. Used in Rdio URLs like /track/We_Up_(Album_Version_(Edited))/
+  regexen.validUrlBalancedParens = regexSupplant(
+    '\\('                                   +
+      '(?:'                                 +
+        '#{validGeneralUrlPathChars}+'      +
+        '|'                                 +
+        // allow one nested level of balanced parentheses
+        '(?:'                               +
+          '#{validGeneralUrlPathChars}*'    +
+          '\\('                             +
+            '#{validGeneralUrlPathChars}+'  +
+          '\\)'                             +
+          '#{validGeneralUrlPathChars}*'    +
+        ')'                                 +
+      ')'                                   +
+    '\\)'
+  , 'i');
+  // Valid end-of-path chracters (so /foo. does not gobble the period).
+  // 1. Allow =&# for empty URL parameters and other URL-join artifacts
+  regexen.validUrlPathEndingChars = regexSupplant(/[^#{spaces_group}\(\)\?!\*';:=\,\.\$%\[\]#{pd}_~&\|@]|(?:#{validUrlBalancedParens})/i);
+  // Allow @ in a url, but only in the middle. Catch things like http://example.com/@user/
+  regexen.validUrlPath = regexSupplant('(?:' +
+    '(?:' +
+      '#{validGeneralUrlPathChars}*' +
+        '(?:#{validUrlBalancedParens}#{validGeneralUrlPathChars}*)*' +
+        '#{validUrlPathEndingChars}'+
+      ')|(?:@#{validGeneralUrlPathChars}+\/)'+
+    ')', 'i');
+  regexen.validUrlQueryChars = /[a-z0-9!?\*'@\(\);:&=\+\$\/%#\[\]\-_\.,~|]/i;
+  regexen.validUrlQueryEndingChars = /[a-z0-9_&=#\/]/i;
+  regexen.validUrl = regexSupplant(
+    '('                                                          + // $1 URL
+      '(https?:\\/\\/)'                                          + // $2 Protocol
+      '(#{validDomain})'                                         + // $3 Domain(s)
+      '(?::(#{validPortNumber}))?'                               + // $4 Port number (optional)
+      '(\\/#{validUrlPath}*)?'                                   + // $5 URL Path
+      '(\\?#{validUrlQueryChars}*#{validUrlQueryEndingChars})?'  + // $6 Query String
+    ')'
+  , 'gi');
+  return regexen.validUrl;
+}());