我跳过了很多圈来得到这个字符串：
"<html>n<head>n<script language="JavaScript"> nn //////////////////////////////////////////////////////////////// n // This [base64 encoder and decoder] was written by Tyler Akins and has been placed in the n // public domain. It would be nice if you left this header intact. n // Base64 code from Tyler Akins -- http://rumkin.com n var keyStr = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/="; nn function encode64(input) { n var output = ""; n var chr1, chr2, chr3; n var enc1, enc2, enc3, enc4; n var i = 0; nn do { n chr1 = input.charCodeAt(i++); n chr2 = input.charCodeAt(i++); n chr3 = input.charCodeAt(i++); nn enc1 = chr1 >> 2; n enc2 = ((chr1 & 3) << 4) | (chr2 >> 4); n enc3 = ((chr2 & 15) << 2) | (chr3 >> 6); n enc4 = chr3 & 63; nn if (isNaN(chr2)) { n enc3 = enc4 = 64; n } else if (isNaN(chr3)) { n enc4 = 64; n } nn output = output + keyStr.charAt(enc1) + keyStr.charAt(enc2) + n keyStr.charAt(enc3) + keyStr.charAt(enc4); n } while (i < input.length); nn return output; n } n function decode64(input) { n var output = ""; n var chr1, chr2, chr3; n var enc1, enc2, enc3, enc4; n var i = 0; nn // remove all characters that are not A-Z, a-z, 0-9, +, /, or = n input = input.replace(/[^A-Za-z0-9\+\/\=]/g, ""); nn do { n enc1 = keyStr.indexOf(input.charAt(i++)); n enc2 = keyStr.indexOf(input.charAt(i++)); n enc3 = keyStr.indexOf(input.charAt(i++)); n enc4 = keyStr.indexOf(input.charAt(i++)); nn chr1 = (enc1 << 2) | (enc2 >> 4); n chr2 = ((enc2 & 15) << 4) | (enc3 >> 2); n chr3 = ((enc3 & 3) << 6) | enc4; nn output = output + String.fromCharCode(chr1); nn if (enc3 != 64) { n output = output + String.fromCharCode(chr2); n } n if (enc4 != 64) { n output = output + String.fromCharCode(chr3); n } n } while (i < input.length); nn return output; n } nn // end of Tyler Akins' code n //////////////////////////////////////////////////////////////// n function escapePluses(s) { n return s.replace(/\+/g, "%2B"); n } n function getFragment(thisuri) { n var pound = thisuri.indexOf("#"); n if (pound == -1) { n return null; n } else { n return thisuri.substr(pound + 1); n } n } n function saveFragment() { n var fragment = getFragment(document.URL); n if (fragment != null) { n var pre_marker = "&aka_frag="; n var g_req = decode64(document.relay.pubcookie_g_req.value); n var header_end = g_req.indexOf(pre_marker) + pre_marker.length; n var req_head = g_req.substr(0,header_end); n var req_foot = g_req.substr(header_end); n if ((req_foot.length > 0) && (req_foot.charAt(0) != '&')) { n req_foot = req_foot.substr(req_foot.indexOf("&")); n } n var new_req = req_head + escapePluses(encode64(fragment)) + req_foot; n document.relay.pubcookie_g_req.value = encode64(new_req); n } n } nn function doStuff() { n saveFragment(); n document.relay.submit(); n } nn// setTimeout('doStuff()', 1000); n</script></head>n<body onLoad="doStuff()">n<form method=post action="https://weblogin.server.com/" name=relay>n<input type=hidden name=pubcookie_g_req value="b25lPXNpYW0tcHJvLnFhLmFrYW1haS5jb20mdHdvPXNpYW0tcHJvLnFhLmFrYW1haS5jb20mdGhyZWU9MSZmb3VyPWE1YSZmaXZlPUdFVCZzaXg9c2lhbS1wcm8ucWEuYWthbWFpLmNvbSZzZXZlbj1MMk52Ym1acFozQmhjbk5sWDNCdmNuUnomZWlnaHQ9JmFrYV9mcmFnPSZob3N0bmFtZT1zaWFtLXByby5xYS5ha2FtYWkuY29tJm5pbmU9MSZmaWxlPSZyZWZlcmVyPShudWxsKSZzZXNzX3JlPTAmcHJlX3Nlc3NfdG9rPS0xNTE4MTQyNjAwJmZsYWc9MA==">n<input type=hidden name=post_stuff value="">n<input type=hidden name=relay_url value="https://siam-pro.qa.server.com/PubCookie.reply">n<noscript>n<p align=center>You do not have Javascript turned on, please click the button to continue.n<p align=center>n<input type=submit name=go value=Continue>n</noscript>n</form>n</html>n"
然后我想在这个正则表达式上匹配该字符串：
<form [^>]*action=(?:\*"([^"]*)\*"|([^" >]*))[^>]* name=relay>(.*?)</form>
这在 rubular 中按预期工作，但在 IRB （1.9.3） 中，我得到以下内容：
1.9.3p448 :147 > data =~/<form [^>]*action=(?:\*"([^"]*)\*"|([^" >]*))[^>]* name=relay>(.*?)</form>/
=> nil
我在这里做错了什么？
Regex 和 HTML/XML 都不是好伙伴。当HTML更改时，您的模式将打破的可能性非常好。解析器可显著降低代码中断的可能性。例如，很容易预测标签中的参数可以更改其顺序：
<form method="post" action="https://weblogin.server.com/" name="relay">
表单可以更改为以下之一：
<form method="post" action="https://weblogin.server.com/" name="relay" >...</form>
<form method="post" action="https://weblogin.server.com/" name="relay1" >...</form>
<form name="relay" method="post" action="https://weblogin.server.com/">...</form>
<form name="relay" method="post" action="https://weblogin.server.com/">...</form >
如果发生其中任何一个，正则表达式会立即被破坏。
解析器不会关心这些更改。
require 'nokogiri'
html = "<html>n<head>n<script language="JavaScript"> nn //////////////////////////////////////////////////////////////// n // This [base64 encoder and decoder] was written by Tyler Akins and has been placed in the n // public domain. It would be nice if you left this header intact. n // Base64 code from Tyler Akins -- http://rumkin.com n var keyStr = "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/="; nn function encode64(input) { n var output = ""; n var chr1, chr2, chr3; n var enc1, enc2, enc3, enc4; n var i = 0; nn do { n chr1 = input.charCodeAt(i++); n chr2 = input.charCodeAt(i++); n chr3 = input.charCodeAt(i++); nn enc1 = chr1 >> 2; n enc2 = ((chr1 & 3) << 4) | (chr2 >> 4); n enc3 = ((chr2 & 15) << 2) | (chr3 >> 6); n enc4 = chr3 & 63; nn if (isNaN(chr2)) { n enc3 = enc4 = 64; n } else if (isNaN(chr3)) { n enc4 = 64; n } nn output = output + keyStr.charAt(enc1) + keyStr.charAt(enc2) + n keyStr.charAt(enc3) + keyStr.charAt(enc4); n } while (i < input.length); nn return output; n } n function decode64(input) { n var output = ""; n var chr1, chr2, chr3; n var enc1, enc2, enc3, enc4; n var i = 0; nn // remove all characters that are not A-Z, a-z, 0-9, +, /, or = n input = input.replace(/[^A-Za-z0-9\+\/\=]/g, ""); nn do { n enc1 = keyStr.indexOf(input.charAt(i++)); n enc2 = keyStr.indexOf(input.charAt(i++)); n enc3 = keyStr.indexOf(input.charAt(i++)); n enc4 = keyStr.indexOf(input.charAt(i++)); nn chr1 = (enc1 << 2) | (enc2 >> 4); n chr2 = ((enc2 & 15) << 4) | (enc3 >> 2); n chr3 = ((enc3 & 3) << 6) | enc4; nn output = output + String.fromCharCode(chr1); nn if (enc3 != 64) { n output = output + String.fromCharCode(chr2); n } n if (enc4 != 64) { n output = output + String.fromCharCode(chr3); n } n } while (i < input.length); nn return output; n } nn // end of Tyler Akins' code n //////////////////////////////////////////////////////////////// n function escapePluses(s) { n return s.replace(/\+/g, "%2B"); n } n function getFragment(thisuri) { n var pound = thisuri.indexOf("#"); n if (pound == -1) { n return null; n } else { n return thisuri.substr(pound + 1); n } n } n function saveFragment() { n var fragment = getFragment(document.URL); n if (fragment != null) { n var pre_marker = "&aka_frag="; n var g_req = decode64(document.relay.pubcookie_g_req.value); n var header_end = g_req.indexOf(pre_marker) + pre_marker.length; n var req_head = g_req.substr(0,header_end); n var req_foot = g_req.substr(header_end); n if ((req_foot.length > 0) && (req_foot.charAt(0) != '&')) { n req_foot = req_foot.substr(req_foot.indexOf("&")); n } n var new_req = req_head + escapePluses(encode64(fragment)) + req_foot; n document.relay.pubcookie_g_req.value = encode64(new_req); n } n } nn function doStuff() { n saveFragment(); n document.relay.submit(); n } nn// setTimeout('doStuff()', 1000); n</script></head>n<body onLoad="doStuff()">n<form method=post action="https://weblogin.server.com/" name=relay>n<input type=hidden name=pubcookie_g_req value="b25lPXNpYW0tcHJvLnFhLmFrYW1haS5jb20mdHdvPXNpYW0tcHJvLnFhLmFrYW1haS5jb20mdGhyZWU9MSZmb3VyPWE1YSZmaXZlPUdFVCZzaXg9c2lhbS1wcm8ucWEuYWthbWFpLmNvbSZzZXZlbj1MMk52Ym1acFozQmhjbk5sWDNCdmNuUnomZWlnaHQ9JmFrYV9mcmFnPSZob3N0bmFtZT1zaWFtLXByby5xYS5ha2FtYWkuY29tJm5pbmU9MSZmaWxlPSZyZWZlcmVyPShudWxsKSZzZXNzX3JlPTAmcHJlX3Nlc3NfdG9rPS0xNTE4MTQyNjAwJmZsYWc9MA==">n<input type=hidden name=post_stuff value="">n<input type=hidden name=relay_url value="https://siam-pro.qa.server.com/PubCookie.reply">n<noscript>n<p align=center>You do not have Javascript turned on, please click the button to continue.n<p align=center>n<input type=submit name=go value=Continue>n</noscript>n</form>n</html>n"
doc = Nokogiri::HTML(html)
form = doc.at('form')
puts form.to_html
# >> <form method="post" action="https://weblogin.server.com/" name="relay">
# >> <input type="hidden" name="pubcookie_g_req" value="b25lPXNpYW0tcHJvLnFhLmFrYW1haS5jb20mdHdvPXNpYW0tcHJvLnFhLmFrYW1haS5jb20mdGhyZWU9MSZmb3VyPWE1YSZmaXZlPUdFVCZzaXg9c2lhbS1wcm8ucWEuYWthbWFpLmNvbSZzZXZlbj1MMk52Ym1acFozQmhjbk5sWDNCdmNuUnomZWlnaHQ9JmFrYV9mcmFnPSZob3N0bmFtZT1zaWFtLXByby5xYS5ha2FtYWkuY29tJm5pbmU9MSZmaWxlPSZyZWZlcmVyPShudWxsKSZzZXNzX3JlPTAmcHJlX3Nlc3NfdG9rPS0xNTE4MTQyNjAwJmZsYWc9MA=="><input type="hidden" name="post_stuff" value=""><input type="hidden" name="relay_url" value="https://siam-pro.qa.server.com/PubCookie.reply"><noscript>
# >> <p align="center">You do not have Javascript turned on, please click the button to continue.
# >> </p>
# >> <p align="center">
# >> <input type="submit" name="go" value="Continue"></p>
# >> </noscript>
# >> </form>
form['action'] # => "https://weblogin.server.com/"
input = form.at('input')
input['name'] # => "pubcookie_g_req"
input['value'] # => "b25lPXNpYW0tcHJvLnFhLmFrYW1haS5jb20mdHdvPXNpYW0tcHJvLnFhLmFrYW1haS5jb20mdGhyZWU9MSZmb3VyPWE1YSZmaXZlPUdFVCZzaXg9c2lhbS1wcm8ucWEuYWthbWFpLmNvbSZzZXZlbj1MMk52Ym1acFozQmhjbk5sWDNCdmNuUnomZWlnaHQ9JmFrYV9mcmFnPSZob3N0bmFtZT1zaWFtLXByby5xYS5ha2FtYWkuY29tJm5pbmU9MSZmaWxlPSZyZWZlcmVyPShudWxsKSZzZXNzX3JlPTAmcHJlX3Nlc3NfdG9rPS0xNTE4MTQyNjAwJmZsYWc9MA=="
Nokogiri 是 Ruby 最受欢迎的 XML/HTML 解析器。它快速，易于使用，并且在我的使用中非常强大。
您需要在此处使用多行正则表达式 - 使用 m 修饰符启用多行匹配，例如
/<form [^>]*action=(?:\*"([^"]*)\*"|([^" >]*))[^>]* name=relay>(.*?)</form>/m