tor-browser

The Tor Browser
git clone https://git.dasho.dev/tor-browser.git
Log | Files | Refs | README | LICENSE

test_serializers_entities_in_attr.js (3788B)

      1 const encoders = {
      2  xml: doc => {
      3    let enc = Cu.createDocumentEncoder("text/xml");
      4    enc.init(doc, "text/xml", Ci.nsIDocumentEncoder.OutputLFLineBreak);
      5    return enc;
      6  },
      7  html: doc => {
      8    let enc = Cu.createDocumentEncoder("text/html");
      9    enc.init(doc, "text/html", Ci.nsIDocumentEncoder.OutputLFLineBreak);
     10    return enc;
     11  },
     12  htmlBasic: doc => {
     13    let enc = Cu.createDocumentEncoder("text/html");
     14    enc.init(
     15      doc,
     16      "text/html",
     17      Ci.nsIDocumentEncoder.OutputEncodeBasicEntities |
     18        Ci.nsIDocumentEncoder.OutputLFLineBreak
     19    );
     20    return enc;
     21  },
     22  xhtml: doc => {
     23    let enc = Cu.createDocumentEncoder("application/xhtml+xml");
     24    enc.init(
     25      doc,
     26      "application/xhtml+xml",
     27      Ci.nsIDocumentEncoder.OutputLFLineBreak
     28    );
     29    return enc;
     30  },
     31 };
     32 
     33 // Which characters should we encode as entities?  It depends on the serializer.
     34 const encodeAll = { html: true, htmlBasic: true, xhtml: true, xml: true };
     35 const encodeHTMLBasic = {
     36  html: false,
     37  htmlBasic: true,
     38  xhtml: false,
     39  xml: false,
     40 };
     41 const encodeXML = { html: false, htmlBasic: false, xhtml: true, xml: true };
     42 const encodeNone = { html: false, htmlBasic: false, xhtml: false, xml: false };
     43 const encodingInfoMap = new Map([
     44  // Basic sanity chars '<', '>', '"', '&' get encoded in all cases.
     45  ["<", encodeAll],
     46  [">", encodeAll],
     47  ['"', encodeAll],
     48  ["&", encodeAll],
     49  // nbsp is only encoded with the HTML encoder when encoding basic entities.
     50  ["\xA0", encodeHTMLBasic],
     51  // Whitespace bits are only encoded in XML.
     52  ["\n", encodeXML],
     53  ["\r", encodeXML],
     54  ["\t", encodeXML],
     55 ]);
     56 
     57 const encodingMap = new Map([
     58  ["<", "&lt;"],
     59  [">", "&gt;"],
     60  ['"', "&quot;"],
     61  ["&", "&amp;"],
     62  ["\xA0", "&nbsp;"],
     63  ["\n", "&#xA;"],
     64  ["\r", "&#xD;"],
     65  ["\t", "&#9;"],
     66 ]);
     67 
     68 function encodingInfoForChar(c) {
     69  var info = encodingInfoMap.get(c);
     70  if (info) {
     71    return info;
     72  }
     73  return encodeNone;
     74 }
     75 
     76 function encodingForChar(c, type) {
     77  var info = encodingInfoForChar(c);
     78  if (!info[type]) {
     79    return c;
     80  }
     81  return encodingMap.get(c);
     82 }
     83 
     84 const doc = new DOMParser().parseFromString("<root></root>", "text/xml");
     85 const root = doc.documentElement;
     86 for (let i = 0; i < 255; ++i) {
     87  let el = doc.createElement("span");
     88  el.setAttribute("x", String.fromCharCode(i));
     89  el.textContent = " ";
     90  root.appendChild(el);
     91 }
     92 for (let type of ["xml", "xhtml", "htmlBasic", "html"]) {
     93  let str = encoders[type](doc).encodeToString();
     94  const prefix = '<root><span x="';
     95  const suffix = '"> </span></root>';
     96  Assert.ok(str.startsWith(prefix), `${type} serialization starts correctly`);
     97  Assert.ok(str.endsWith(suffix), `${type} serialization ends correctly`);
     98  str = str.substring(prefix.length, str.length - suffix.length);
     99  let encodings = str.split('"> </span><span x="');
    100  for (let i = 0; i < 255; ++i) {
    101    let c = String.fromCharCode(i);
    102    Assert.equal(
    103      encodingForChar(c, type),
    104      encodings[i],
    105      `${type} encoding of char ${i} is correct`
    106    );
    107  }
    108 }
    109 
    110 // Verify that we don't escape entities in attributes with JS code.
    111 const docWithJS = new DOMParser().parseFromString("<root></root>", "text/xml");
    112 const rootWithJS = docWithJS.documentElement;
    113 const htmlNS = "http://www.w3.org/1999/xhtml";
    114 const jsCode = "x < 2 || y > 3";
    115 let div = docWithJS.createElementNS(htmlNS, "div");
    116 div.setAttribute("href", `javascript:${jsCode}`);
    117 div.setAttribute("src", `javascript:${jsCode}`);
    118 div.setAttribute("onclick", jsCode);
    119 div.setAttribute("onload", jsCode);
    120 rootWithJS.appendChild(div);
    121 for (let type of ["xml", "xhtml", "htmlBasic", "html"]) {
    122  let str = encoders[type](docWithJS).encodeToString();
    123  Assert.equal(
    124    str.search("&[a-z]+;"),
    125    -1,
    126    `${type} encoding does not escape entities in attributes with JS code`
    127  );
    128 }