Skip to content

Instantly share code, notes, and snippets.

@rg443a
Created January 10, 2019 05:41
Show Gist options
  • Save rg443a/6f1e166ad53a6570daad4ec0255812c3 to your computer and use it in GitHub Desktop.
Save rg443a/6f1e166ad53a6570daad4ec0255812c3 to your computer and use it in GitHub Desktop.
suffix blacklist (tld filter)
var suffix_black_list=[
"xml",
"atom",
"rss",
"mml",
"txt",
"jad",
"wml",
"htc",
"woff",
"jar ",
"war ",
"ear",
"hqx",
"doc",
"pdf",
"ps ",
"eps ",
"ai",
"rtf",
"m3u8",
"xls",
"eot",
"ppt",
"wmlc",
"kml",
"kmz",
"7z",
"cco",
"jardiff",
"jnlp",
"run",
"pl ",
"pm",
"prc ",
"pdb",
"rar",
"rpm",
"sea",
"swf",
"sit",
"tcl ",
"tk",
"der ",
"pem ",
"crt",
"xpi",
"xspf",
"zip",
"bin ",
"exe ",
"dll",
"deb",
"dmg",
"iso ",
"img",
"msi ",
"msp ",
"msm",
"docx",
"xlsx",
"pptx",
"mid ",
"midi ",
"kar",
"mp3",
"ogg",
"m4a",
"ra",
"3gpp ",
"3gp",
"ts",
"mp4",
"mpeg ",
"mpg",
"mov",
"webm",
"flv",
"m4v",
"mng",
"asx ",
"asf",
"wmv",
"avi",
"epub"
];
var domain_black_list= [
"addthis",
"mozilla",
"mozilla",
"microsoft",
"mozilla",
"google",
"picasaweb",
"yahoo",
"bing",
"youtube",
"twitter",
"linkedin",
"facebook",
"pinterest",
"foursquare",
"instagram",
"last",
"wikipedia",
"baidu",
"academia",
"feedburner",
"galleria",
"socialmediatoday",
"badoo",
"bebo",
"blackplanet",
"classmates",
"fixter",
"friendster",
"gaiaonline",
"habbo",
"hi5",
"hyves",
"ibibo",
"kaixin001",
"meetup",
"millatfacebook",
"multiply",
"myheritage",
"mylife",
"myspace",
"myyearbook",
"nate",
"netlog",
"ning",
"orkut",
"perfspot",
"renren",
"skyrock",
"socialmediatoday",
"sonico",
"soravjain",
"studivz.net",
"tagged",
"tiwitter",
"tuenti",
"wer-kennt-wen",
"zorpia",
"cucumbertown",
"abc",
"advogato",
"alexa",
"alleyinsider",
"athlinks",
"bloomberg",
"business2community",
"buzznet",
"care2",
"caringbridge",
"classmates",
"couchsurfing",
"crainsnewyork",
"cross.tv",
"crunchbase",
"delicious",
"digitalspy",
"dol2day",
"douban",
"elftown",
"eweek",
"facepartyexposed",
"filmaffinity",
"flickr",
"flixster",
"friendster",
"ft",
"gaiaonline",
"gamerdna",
"gather",
"gays",
"geni",
"goodreads",
"goodwizz",
"hispanicprwire",
"hospitalityclub",
"humanresourcesjournal",
"hyves",
"ibibo",
"italki",
"librarything",
"livejournal",
"mediapost",
"mediawiki",
"mobile-ent",
"myheritage",
"myspace",
"nexopia",
"ngopost",
"odnoklassniki",
"orkut",
"patientslikeme",
"poolwo",
"prnewswire",
"readwriteweb",
"reverbnation",
"rnw",
"sinlung",
"socialmediatoday",
"statisticbrain",
"straitstimes",
"studivz.net",
"sulake",
"taringa.net",
"techcrunch",
"techtudo.br",
"telecom.paper",
"thebeijinger",
"travbuddy",
"travellerspoint",
"treehugger",
"trombi",
"usatoday",
"viadeo",
"vox",
"washingtonpost",
"web-strategist",
"weeworld",
"wikidata",
"wikimediafoundation",
"wiser",
"writeaprisoner",
"zibb",
"zh.wikipedia",
"zooppa",
"zanox",
"ovh",
"webadmax",
"digg",
"amazon"
];
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment