1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/uconv/tests/unit/test_utf8_illegals.js Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,136 @@
1.4 +// Tests illegal UTF-8 sequences
1.5 +
1.6 +const Cc = Components.Constructor;
1.7 +const Ci = Components.interfaces;
1.8 +
1.9 +const tests = [
1.10 +{ inStrings: ["%80", // Illegal or incomplete sequences
1.11 + "%8f",
1.12 + "%90",
1.13 + "%9f",
1.14 + "%a0",
1.15 + "%bf",
1.16 + "%c0",
1.17 + "%c1",
1.18 + "%c2",
1.19 + "%df",
1.20 + "%e0",
1.21 + "%e0%a0",
1.22 + "%e0%bf",
1.23 + "%ed%80",
1.24 + "%ed%9f",
1.25 + "%ef",
1.26 + "%ef%bf",
1.27 + "%f0",
1.28 + "%f0%90",
1.29 + "%f0%90%80",
1.30 + "%f0%90%bf",
1.31 + "%f0%bf",
1.32 + "%f0%bf%80",
1.33 + "%f0%bf%bf",
1.34 + "%f4",
1.35 + "%f4%80",
1.36 + "%f4%80%80",
1.37 + "%f4%80%bf",
1.38 + "%f4%8f",
1.39 + "%f4%8f%80",
1.40 + "%f4%8f%bf",
1.41 + "%f5",
1.42 + "%f7",
1.43 + "%f8",
1.44 + "%fb",
1.45 + "%fc",
1.46 + "%fd"],
1.47 + expected: "ABC\ufffdXYZ" },
1.48 +
1.49 +{ inStrings: ["%c0%af", // Illegal bytes in 2-octet
1.50 + "%c1%af"], // sequences
1.51 + expected: "ABC\ufffd\ufffdXYZ" },
1.52 +
1.53 +{ inStrings: ["%e0%80%80", // Illegal bytes in 3-octet
1.54 + "%e0%80%af", // sequences
1.55 + "%e0%9f%bf",
1.56 + // long surrogates
1.57 + "%ed%a0%80", // D800
1.58 + "%ed%ad%bf", // DB7F
1.59 + "%ed%ae%80", // DB80
1.60 + "%ed%af%bf", // DBFF
1.61 + "%ed%b0%80", // DC00
1.62 + "%ed%be%80", // DF80
1.63 + "%ed%bf%bf"], // DFFF
1.64 + expected: "ABC\ufffd\ufffd\ufffdXYZ" },
1.65 +
1.66 +{ inStrings: ["%f0%80%80%80", // Illegal bytes in 4-octet
1.67 + "%f0%80%80%af", // sequences
1.68 + "%f0%8f%bf%bf",
1.69 + "%f4%90%80%80",
1.70 + "%f4%bf%bf%bf",
1.71 + "%f5%80%80%80",
1.72 + "%f7%bf%bf%bf"],
1.73 + expected: "ABC\ufffd\ufffd\ufffd\ufffdXYZ" },
1.74 +
1.75 +{ inStrings: ["%f8%80%80%80%80", // Illegal bytes in 5-octet
1.76 + "%f8%80%80%80%af", // sequences
1.77 + "%fb%bf%bf%bf%bf"],
1.78 + expected: "ABC\ufffd\ufffd\ufffd\ufffd\ufffdXYZ" },
1.79 +
1.80 + // Surrogate pairs
1.81 +{ inStrings: ["%ed%a0%80%ed%b0%80", // D800 DC00
1.82 + "%ed%a0%80%ed%bf%bf", // D800 DFFF
1.83 + "%ed%ad%bf%ed%b0%80", // DB7F DC00
1.84 + "%ed%ad%bf%ed%bf%bf", // DB7F DFFF
1.85 + "%ed%ae%80%ed%b0%80", // DB80 DC00
1.86 + "%ed%ae%80%ed%bf%bf", // DB80 DFFF
1.87 + "%ed%af%bf%ed%b0%80", // DBFF DC00
1.88 + "%ed%ad%bf%ed%bf%bf", // DBFF DFFF
1.89 + "%fc%80%80%80%80%80", // Illegal bytes in 6-octet
1.90 + "%fc%80%80%80%80%af", // sequences
1.91 + "%fd%bf%bf%bf%bf%bf"],
1.92 + expected: "ABC\ufffd\ufffd\ufffd\ufffd\ufffd\ufffdXYZ" },
1.93 +];
1.94 +
1.95 +
1.96 +function testCaseInputStream(inStr, expected)
1.97 +{
1.98 + var dataURI = "data:text/plain; charset=UTF-8,ABC" + inStr + "XYZ"
1.99 + dump(inStr + "==>");
1.100 +
1.101 + var IOService = Cc("@mozilla.org/network/io-service;1",
1.102 + "nsIIOService");
1.103 + var ConverterInputStream =
1.104 + Cc("@mozilla.org/intl/converter-input-stream;1",
1.105 + "nsIConverterInputStream",
1.106 + "init");
1.107 +
1.108 + var ios = new IOService();
1.109 + var channel = ios.newChannel(dataURI, "", null);
1.110 + var testInputStream = channel.open();
1.111 + var testConverter = new ConverterInputStream(testInputStream,
1.112 + "UTF-8",
1.113 + 16,
1.114 + 0xFFFD);
1.115 +
1.116 + if (!(testConverter instanceof Ci.nsIUnicharLineInputStream))
1.117 + throw "not line input stream";
1.118 +
1.119 + var outStr = "";
1.120 + var more;
1.121 + do {
1.122 + // read the line and check for eof
1.123 + var line = {};
1.124 + more = testConverter.readLine(line);
1.125 + outStr += line.value;
1.126 + } while (more);
1.127 +
1.128 + dump(outStr + "; expected=" + expected + "\n");
1.129 + do_check_eq(outStr, expected);
1.130 + do_check_eq(outStr.length, expected.length);
1.131 +}
1.132 +
1.133 +function run_test() {
1.134 + for (var t of tests) {
1.135 + for (var inStr of t.inStrings) {
1.136 + testCaseInputStream(inStr, t.expected);
1.137 + }
1.138 + }
1.139 +}
