idna10.0.0_test.go

Documentation: golang.org/x/text/internal/export/idna

     1  // Copyright 2016 The Go Authors. All rights reserved.
     2  // Use of this source code is governed by a BSD-style
     3  // license that can be found in the LICENSE file.
     4  
     5  //go:build go1.10
     6  
     7  package idna
     8  
     9  import "testing"
    10  
    11  // TestLabelErrors tests strings returned in case of error. All results should
    12  // be identical to the reference implementation and can be verified at
    13  // https://unicode.org/cldr/utility/idna.jsp. The reference implementation,
    14  // however, seems to not display Bidi and ContextJ errors.
    15  //
    16  // In some cases the behavior of browsers is added as a comment. In all cases,
    17  // whenever a resolve search returns an error here, Chrome will treat the input
    18  // string as a search string (including those for Bidi and Context J errors),
    19  // unless noted otherwise.
    20  func TestLabelErrors(t *testing.T) {
    21  	encode := func(s string) string { s, _ = encode(acePrefix, s); return s }
    22  	type kind struct {
    23  		name string
    24  		f    func(string) (string, error)
    25  	}
    26  	punyA := kind{"PunycodeA", punycode.ToASCII}
    27  	resolve := kind{"ResolveA", Lookup.ToASCII}
    28  	display := kind{"ToUnicode", Display.ToUnicode}
    29  	p := New(VerifyDNSLength(true), MapForLookup(), BidiRule())
    30  	lengthU := kind{"CheckLengthU", p.ToUnicode}
    31  	lengthA := kind{"CheckLengthA", p.ToASCII}
    32  	p = New(MapForLookup(), StrictDomainName(false))
    33  	std3 := kind{"STD3", p.ToASCII}
    34  	p = New(MapForLookup(), CheckHyphens(false))
    35  	hyphens := kind{"CheckHyphens", p.ToASCII}
    36  	p = New(MapForLookup(), Transitional(true))
    37  	transitional := kind{"Transitional", p.ToASCII}
    38  	p = New(MapForLookup(), Transitional(false))
    39  	nontransitional := kind{"Nontransitional", p.ToASCII}
    40  
    41  	testCases := []struct {
    42  		kind
    43  		input   string
    44  		want    string
    45  		wantErr string
    46  	}{
    47  		{lengthU, "", "", "A4"}, // From UTS 46 conformance test.
    48  		{lengthA, "", "", "A4"},
    49  
    50  		{lengthU, "xn--", "", "A4"},
    51  		{lengthU, "foo.xn--", "foo.", "A4"}, // TODO: is dropping xn-- correct?
    52  		{lengthU, "xn--.foo", ".foo", "A4"},
    53  		{lengthU, "foo.xn--.bar", "foo..bar", "A4"},
    54  
    55  		{display, "xn--", "", ""},
    56  		{display, "foo.xn--", "foo.", ""}, // TODO: is dropping xn-- correct?
    57  		{display, "xn--.foo", ".foo", ""},
    58  		{display, "foo.xn--.bar", "foo..bar", ""},
    59  
    60  		{lengthA, "a..b", "a..b", "A4"},
    61  		{punyA, ".b", ".b", ""},
    62  		// For backwards compatibility, the Punycode profile does not map runes.
    63  		{punyA, "\u3002b", "xn--b-83t", ""},
    64  		{punyA, "..b", "..b", ""},
    65  
    66  		{lengthA, ".b", ".b", "A4"},
    67  		{lengthA, "\u3002b", ".b", "A4"},
    68  		{lengthA, "..b", "..b", "A4"},
    69  		{lengthA, "b..", "b..", ""},
    70  
    71  		// Sharpened Bidi rules for Unicode 10.0.0. Apply for ALL labels in ANY
    72  		// of the labels is RTL.
    73  		{lengthA, "\ufe05\u3002\u3002\U0002603e\u1ce0", "..xn--t6f5138v", "A4"},
    74  		{lengthA, "FAX\u2a77\U0001d186\u3002\U0001e942\U000e0181\u180c", "", "B6"},
    75  
    76  		{resolve, "a..b", "a..b", ""},
    77  		// Note that leading dots are not stripped. This is to be consistent
    78  		// with the Punycode profile as well as the conformance test.
    79  		{resolve, ".b", ".b", ""},
    80  		{resolve, "\u3002b", ".b", ""},
    81  		{resolve, "..b", "..b", ""},
    82  		{resolve, "b..", "b..", ""},
    83  		{resolve, "\xed", "", "P1"},
    84  
    85  		// Raw punycode
    86  		{punyA, "", "", ""},
    87  		{punyA, "*.foo.com", "*.foo.com", ""},
    88  		{punyA, "Foo.com", "Foo.com", ""},
    89  
    90  		// STD3 rules
    91  		{display, "*.foo.com", "*.foo.com", "P1"},
    92  		{std3, "*.foo.com", "*.foo.com", ""},
    93  
    94  		// Hyphens
    95  		{display, "r3---sn-apo3qvuoxuxbt-j5pe.googlevideo.com", "r3---sn-apo3qvuoxuxbt-j5pe.googlevideo.com", "V2"},
    96  		{hyphens, "r3---sn-apo3qvuoxuxbt-j5pe.googlevideo.com", "r3---sn-apo3qvuoxuxbt-j5pe.googlevideo.com", ""},
    97  		{display, "-label-.com", "-label-.com", "V3"},
    98  		{hyphens, "-label-.com", "-label-.com", ""},
    99  
   100  		// Don't map U+2490 (DIGIT NINE FULL STOP). This is the behavior of
   101  		// Chrome, modern Firefox, Safari, and IE.
   102  		{resolve, "lab⒐be", "xn--labbe-zh9b", "P1"}, // encode("lab⒐be")
   103  		{display, "lab⒐be", "lab⒐be", "P1"},
   104  		{transitional, "plan⒐faß.de", "xn--planfass-c31e.de", "P1"}, // encode("plan⒐fass") + ".de"
   105  		{display, "Plan⒐faß.de", "plan⒐faß.de", "P1"},
   106  
   107  		// Transitional vs Nontransitional processing
   108  		{transitional, "Plan9faß.de", "plan9fass.de", ""},
   109  		{nontransitional, "Plan9faß.de", "xn--plan9fa-6va.de", ""},
   110  
   111  		// Chrome 54.0 recognizes the error and treats this input verbatim as a
   112  		// search string.
   113  		// Safari 10.0 (non-conform spec) decomposes "⒈" and computes the
   114  		// punycode on the result using transitional mapping.
   115  		// Firefox 49.0.1 goes haywire on this string and prints a bunch of what
   116  		// seems to be nested punycode encodings.
   117  		{transitional, "日本⒈co.ßßß.de", "xn--co-wuw5954azlb.ssssss.de", "P1"},
   118  		{display, "日本⒈co.ßßß.de", "日本⒈co.ßßß.de", "P1"},
   119  
   120  		{transitional, "a\u200Cb", "ab", ""},
   121  		{display, "a\u200Cb", "a\u200Cb", "C"},
   122  
   123  		{resolve, encode("a\u200Cb"), encode("a\u200Cb"), "C"},
   124  		{display, "a\u200Cb", "a\u200Cb", "C"},
   125  
   126  		{resolve, "grﻋﺮﺑﻲ.de", "xn--gr-gtd9a1b0g.de", "B"},
   127  		{
   128  			// Notice how the string gets transformed, even with an error.
   129  			// Chrome will use the original string if it finds an error, so not
   130  			// the transformed one.
   131  			display,
   132  			"gr\ufecb\ufeae\ufe91\ufef2.de",
   133  			"gr\u0639\u0631\u0628\u064a.de",
   134  			"B",
   135  		},
   136  
   137  		{resolve, "\u0671.\u03c3\u07dc", "xn--qib.xn--4xa21s", "B"}, // ٱ.σߜ
   138  		{display, "\u0671.\u03c3\u07dc", "\u0671.\u03c3\u07dc", "B"},
   139  
   140  		// normalize input
   141  		{resolve, "a\u0323\u0322", "xn--jta191l", ""}, // ạ̢
   142  		{display, "a\u0323\u0322", "\u1ea1\u0322", ""},
   143  
   144  		// Non-normalized strings are not normalized when they originate from
   145  		// punycode. Despite the error, Chrome, Safari and Firefox will attempt
   146  		// to look up the input punycode.
   147  		{resolve, encode("a\u0323\u0322") + ".com", "xn--a-tdbc.com", "V1"},
   148  		{display, encode("a\u0323\u0322") + ".com", "a\u0323\u0322.com", "V1"},
   149  	}
   150  
   151  	for _, tc := range testCases {
   152  		doTest(t, tc.f, tc.name, tc.input, tc.want, tc.wantErr)
   153  	}
   154  }
   155  
   156  func TestTransitionalDefault(t *testing.T) {
   157  	want := "xn--strae-oqa.de"
   158  	if transitionalLookup {
   159  		want = "strasse.de"
   160  	}
   161  	doTest(t, Lookup.ToASCII, "Lookup", "straße.de", want, "")
   162  }
   163
View as plain text