Source file src/regexp/syntax/parse_test.go

     1  // Copyright 2011 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  package syntax
     6  
     7  import (
     8  	"fmt"
     9  	"strings"
    10  	"testing"
    11  	"unicode"
    12  )
    13  
    14  type parseTest struct {
    15  	Regexp string
    16  	Dump   string
    17  }
    18  
    19  var parseTests = []parseTest{
    20  	// Base cases
    21  	{`a`, `lit{a}`},
    22  	{`a.`, `cat{lit{a}dot{}}`},
    23  	{`a.b`, `cat{lit{a}dot{}lit{b}}`},
    24  	{`ab`, `str{ab}`},
    25  	{`a.b.c`, `cat{lit{a}dot{}lit{b}dot{}lit{c}}`},
    26  	{`abc`, `str{abc}`},
    27  	{`a|^`, `alt{lit{a}bol{}}`},
    28  	{`a|b`, `cc{0x61-0x62}`},
    29  	{`(a)`, `cap{lit{a}}`},
    30  	{`(a)|b`, `alt{cap{lit{a}}lit{b}}`},
    31  	{`a*`, `star{lit{a}}`},
    32  	{`a+`, `plus{lit{a}}`},
    33  	{`a?`, `que{lit{a}}`},
    34  	{`a{2}`, `rep{2,2 lit{a}}`},
    35  	{`a{2,3}`, `rep{2,3 lit{a}}`},
    36  	{`a{2,}`, `rep{2,-1 lit{a}}`},
    37  	{`a*?`, `nstar{lit{a}}`},
    38  	{`a+?`, `nplus{lit{a}}`},
    39  	{`a??`, `nque{lit{a}}`},
    40  	{`a{2}?`, `nrep{2,2 lit{a}}`},
    41  	{`a{2,3}?`, `nrep{2,3 lit{a}}`},
    42  	{`a{2,}?`, `nrep{2,-1 lit{a}}`},
    43  	// Malformed { } are treated as literals.
    44  	{`x{1001`, `str{x{1001}`},
    45  	{`x{9876543210`, `str{x{9876543210}`},
    46  	{`x{9876543210,`, `str{x{9876543210,}`},
    47  	{`x{2,1`, `str{x{2,1}`},
    48  	{`x{1,9876543210`, `str{x{1,9876543210}`},
    49  	{``, `emp{}`},
    50  	{`|`, `emp{}`}, // alt{emp{}emp{}} but got factored
    51  	{`|x|`, `alt{emp{}lit{x}emp{}}`},
    52  	{`.`, `dot{}`},
    53  	{`^`, `bol{}`},
    54  	{`$`, `eol{}`},
    55  	{`\|`, `lit{|}`},
    56  	{`\(`, `lit{(}`},
    57  	{`\)`, `lit{)}`},
    58  	{`\*`, `lit{*}`},
    59  	{`\+`, `lit{+}`},
    60  	{`\?`, `lit{?}`},
    61  	{`{`, `lit{{}`},
    62  	{`}`, `lit{}}`},
    63  	{`\.`, `lit{.}`},
    64  	{`\^`, `lit{^}`},
    65  	{`\$`, `lit{$}`},
    66  	{`\\`, `lit{\}`},
    67  	{`[ace]`, `cc{0x61 0x63 0x65}`},
    68  	{`[abc]`, `cc{0x61-0x63}`},
    69  	{`[a-z]`, `cc{0x61-0x7a}`},
    70  	{`[a]`, `lit{a}`},
    71  	{`\-`, `lit{-}`},
    72  	{`-`, `lit{-}`},
    73  	{`\_`, `lit{_}`},
    74  	{`abc`, `str{abc}`},
    75  	{`abc|def`, `alt{str{abc}str{def}}`},
    76  	{`abc|def|ghi`, `alt{str{abc}str{def}str{ghi}}`},
    77  
    78  	// Posix and Perl extensions
    79  	{`[[:lower:]]`, `cc{0x61-0x7a}`},
    80  	{`[a-z]`, `cc{0x61-0x7a}`},
    81  	{`[^[:lower:]]`, `cc{0x0-0x60 0x7b-0x10ffff}`},
    82  	{`[[:^lower:]]`, `cc{0x0-0x60 0x7b-0x10ffff}`},
    83  	{`(?i)[[:lower:]]`, `cc{0x41-0x5a 0x61-0x7a 0x17f 0x212a}`},
    84  	{`(?i)[a-z]`, `cc{0x41-0x5a 0x61-0x7a 0x17f 0x212a}`},
    85  	{`(?i)[^[:lower:]]`, `cc{0x0-0x40 0x5b-0x60 0x7b-0x17e 0x180-0x2129 0x212b-0x10ffff}`},
    86  	{`(?i)[[:^lower:]]`, `cc{0x0-0x40 0x5b-0x60 0x7b-0x17e 0x180-0x2129 0x212b-0x10ffff}`},
    87  	{`\d`, `cc{0x30-0x39}`},
    88  	{`\D`, `cc{0x0-0x2f 0x3a-0x10ffff}`},
    89  	{`\s`, `cc{0x9-0xa 0xc-0xd 0x20}`},
    90  	{`\S`, `cc{0x0-0x8 0xb 0xe-0x1f 0x21-0x10ffff}`},
    91  	{`\w`, `cc{0x30-0x39 0x41-0x5a 0x5f 0x61-0x7a}`},
    92  	{`\W`, `cc{0x0-0x2f 0x3a-0x40 0x5b-0x5e 0x60 0x7b-0x10ffff}`},
    93  	{`(?i)\w`, `cc{0x30-0x39 0x41-0x5a 0x5f 0x61-0x7a 0x17f 0x212a}`},
    94  	{`(?i)\W`, `cc{0x0-0x2f 0x3a-0x40 0x5b-0x5e 0x60 0x7b-0x17e 0x180-0x2129 0x212b-0x10ffff}`},
    95  	{`[^\\]`, `cc{0x0-0x5b 0x5d-0x10ffff}`},
    96  	//	{ `\C`, `byte{}` },  // probably never
    97  
    98  	// Unicode, negatives, and a double negative.
    99  	{`\p{Braille}`, `cc{0x2800-0x28ff}`},
   100  	{`\P{Braille}`, `cc{0x0-0x27ff 0x2900-0x10ffff}`},
   101  	{`\p{^Braille}`, `cc{0x0-0x27ff 0x2900-0x10ffff}`},
   102  	{`\P{^Braille}`, `cc{0x2800-0x28ff}`},
   103  	{`\pZ`, `cc{0x20 0xa0 0x1680 0x2000-0x200a 0x2028-0x2029 0x202f 0x205f 0x3000}`},
   104  	{`[\p{Braille}]`, `cc{0x2800-0x28ff}`},
   105  	{`[\P{Braille}]`, `cc{0x0-0x27ff 0x2900-0x10ffff}`},
   106  	{`[\p{^Braille}]`, `cc{0x0-0x27ff 0x2900-0x10ffff}`},
   107  	{`[\P{^Braille}]`, `cc{0x2800-0x28ff}`},
   108  	{`[\pZ]`, `cc{0x20 0xa0 0x1680 0x2000-0x200a 0x2028-0x2029 0x202f 0x205f 0x3000}`},
   109  	{`\p{Lu}`, mkCharClass(unicode.IsUpper)},
   110  	{`[\p{Lu}]`, mkCharClass(unicode.IsUpper)},
   111  	{`(?i)[\p{Lu}]`, mkCharClass(isUpperFold)},
   112  	{`\p{Any}`, `dot{}`},
   113  	{`\p{^Any}`, `cc{}`},
   114  
   115  	// Hex, octal.
   116  	{`[\012-\234]\141`, `cat{cc{0xa-0x9c}lit{a}}`},
   117  	{`[\x{41}-\x7a]\x61`, `cat{cc{0x41-0x7a}lit{a}}`},
   118  
   119  	// More interesting regular expressions.
   120  	{`a{,2}`, `str{a{,2}}`},
   121  	{`\.\^\$\\`, `str{.^$\}`},
   122  	{`[a-zABC]`, `cc{0x41-0x43 0x61-0x7a}`},
   123  	{`[^a]`, `cc{0x0-0x60 0x62-0x10ffff}`},
   124  	{`[α-ε☺]`, `cc{0x3b1-0x3b5 0x263a}`}, // utf-8
   125  	{`a*{`, `cat{star{lit{a}}lit{{}}`},
   126  
   127  	// Test precedences
   128  	{`(?:ab)*`, `star{str{ab}}`},
   129  	{`(ab)*`, `star{cap{str{ab}}}`},
   130  	{`ab|cd`, `alt{str{ab}str{cd}}`},
   131  	{`a(b|c)d`, `cat{lit{a}cap{cc{0x62-0x63}}lit{d}}`},
   132  
   133  	// Test flattening.
   134  	{`(?:a)`, `lit{a}`},
   135  	{`(?:ab)(?:cd)`, `str{abcd}`},
   136  	{`(?:a+b+)(?:c+d+)`, `cat{plus{lit{a}}plus{lit{b}}plus{lit{c}}plus{lit{d}}}`},
   137  	{`(?:a+|b+)|(?:c+|d+)`, `alt{plus{lit{a}}plus{lit{b}}plus{lit{c}}plus{lit{d}}}`},
   138  	{`(?:a|b)|(?:c|d)`, `cc{0x61-0x64}`},
   139  	{`a|.`, `dot{}`},
   140  	{`.|a`, `dot{}`},
   141  	{`(?:[abc]|A|Z|hello|world)`, `alt{cc{0x41 0x5a 0x61-0x63}str{hello}str{world}}`},
   142  	{`(?:[abc]|A|Z)`, `cc{0x41 0x5a 0x61-0x63}`},
   143  
   144  	// Test Perl quoted literals
   145  	{`\Q+|*?{[\E`, `str{+|*?{[}`},
   146  	{`\Q+\E+`, `plus{lit{+}}`},
   147  	{`\Qab\E+`, `cat{lit{a}plus{lit{b}}}`},
   148  	{`\Q\\E`, `lit{\}`},
   149  	{`\Q\\\E`, `str{\\}`},
   150  
   151  	// Test Perl \A and \z
   152  	{`(?m)^`, `bol{}`},
   153  	{`(?m)$`, `eol{}`},
   154  	{`(?-m)^`, `bot{}`},
   155  	{`(?-m)$`, `eot{}`},
   156  	{`(?m)\A`, `bot{}`},
   157  	{`(?m)\z`, `eot{\z}`},
   158  	{`(?-m)\A`, `bot{}`},
   159  	{`(?-m)\z`, `eot{\z}`},
   160  
   161  	// Test named captures
   162  	{`(?P<name>a)`, `cap{name:lit{a}}`},
   163  	{`(?<name>a)`, `cap{name:lit{a}}`},
   164  
   165  	// Case-folded literals
   166  	{`[Aa]`, `litfold{A}`},
   167  	{`[\x{100}\x{101}]`, `litfold{Ā}`},
   168  	{`[Δδ]`, `litfold{Δ}`},
   169  
   170  	// Strings
   171  	{`abcde`, `str{abcde}`},
   172  	{`[Aa][Bb]cd`, `cat{strfold{AB}str{cd}}`},
   173  
   174  	// Factoring.
   175  	{`abc|abd|aef|bcx|bcy`, `alt{cat{lit{a}alt{cat{lit{b}cc{0x63-0x64}}str{ef}}}cat{str{bc}cc{0x78-0x79}}}`},
   176  	{`ax+y|ax+z|ay+w`, `cat{lit{a}alt{cat{plus{lit{x}}lit{y}}cat{plus{lit{x}}lit{z}}cat{plus{lit{y}}lit{w}}}}`},
   177  
   178  	// Bug fixes.
   179  	{`(?:.)`, `dot{}`},
   180  	{`(?:x|(?:xa))`, `cat{lit{x}alt{emp{}lit{a}}}`},
   181  	{`(?:.|(?:.a))`, `cat{dot{}alt{emp{}lit{a}}}`},
   182  	{`(?:A(?:A|a))`, `cat{lit{A}litfold{A}}`},
   183  	{`(?:A|a)`, `litfold{A}`},
   184  	{`A|(?:A|a)`, `litfold{A}`},
   185  	{`(?s).`, `dot{}`},
   186  	{`(?-s).`, `dnl{}`},
   187  	{`(?:(?:^).)`, `cat{bol{}dot{}}`},
   188  	{`(?-s)(?:(?:^).)`, `cat{bol{}dnl{}}`},
   189  	{`[\s\S]a`, `cat{cc{0x0-0x10ffff}lit{a}}`},
   190  
   191  	// RE2 prefix_tests
   192  	{`abc|abd`, `cat{str{ab}cc{0x63-0x64}}`},
   193  	{`a(?:b)c|abd`, `cat{str{ab}cc{0x63-0x64}}`},
   194  	{`abc|abd|aef|bcx|bcy`,
   195  		`alt{cat{lit{a}alt{cat{lit{b}cc{0x63-0x64}}str{ef}}}` +
   196  			`cat{str{bc}cc{0x78-0x79}}}`},
   197  	{`abc|x|abd`, `alt{str{abc}lit{x}str{abd}}`},
   198  	{`(?i)abc|ABD`, `cat{strfold{AB}cc{0x43-0x44 0x63-0x64}}`},
   199  	{`[ab]c|[ab]d`, `cat{cc{0x61-0x62}cc{0x63-0x64}}`},
   200  	{`.c|.d`, `cat{dot{}cc{0x63-0x64}}`},
   201  	{`x{2}|x{2}[0-9]`,
   202  		`cat{rep{2,2 lit{x}}alt{emp{}cc{0x30-0x39}}}`},
   203  	{`x{2}y|x{2}[0-9]y`,
   204  		`cat{rep{2,2 lit{x}}alt{lit{y}cat{cc{0x30-0x39}lit{y}}}}`},
   205  	{`a.*?c|a.*?b`,
   206  		`cat{lit{a}alt{cat{nstar{dot{}}lit{c}}cat{nstar{dot{}}lit{b}}}}`},
   207  
   208  	// Valid repetitions.
   209  	{`((((((((((x{2}){2}){2}){2}){2}){2}){2}){2}){2}))`, ``},
   210  	{`((((((((((x{1}){2}){2}){2}){2}){2}){2}){2}){2}){2})`, ``},
   211  
   212  	// Valid nesting.
   213  	{strings.Repeat("(", 999) + strings.Repeat(")", 999), ``},
   214  	{strings.Repeat("(?:", 999) + strings.Repeat(")*", 999), ``},
   215  	{"(" + strings.Repeat("|", 12345) + ")", ``}, // not nested at all
   216  }
   217  
   218  const testFlags = MatchNL | PerlX | UnicodeGroups
   219  
   220  func TestParseSimple(t *testing.T) {
   221  	testParseDump(t, parseTests, testFlags)
   222  }
   223  
   224  var foldcaseTests = []parseTest{
   225  	{`AbCdE`, `strfold{ABCDE}`},
   226  	{`[Aa]`, `litfold{A}`},
   227  	{`a`, `litfold{A}`},
   228  
   229  	// 0x17F is an old English long s (looks like an f) and folds to s.
   230  	// 0x212A is the Kelvin symbol and folds to k.
   231  	{`A[F-g]`, `cat{litfold{A}cc{0x41-0x7a 0x17f 0x212a}}`}, // [Aa][A-z...]
   232  	{`[[:upper:]]`, `cc{0x41-0x5a 0x61-0x7a 0x17f 0x212a}`},
   233  	{`[[:lower:]]`, `cc{0x41-0x5a 0x61-0x7a 0x17f 0x212a}`},
   234  }
   235  
   236  func TestParseFoldCase(t *testing.T) {
   237  	testParseDump(t, foldcaseTests, FoldCase)
   238  }
   239  
   240  var literalTests = []parseTest{
   241  	{"(|)^$.[*+?]{5,10},\\", "str{(|)^$.[*+?]{5,10},\\}"},
   242  }
   243  
   244  func TestParseLiteral(t *testing.T) {
   245  	testParseDump(t, literalTests, Literal)
   246  }
   247  
   248  var matchnlTests = []parseTest{
   249  	{`.`, `dot{}`},
   250  	{"\n", "lit{\n}"},
   251  	{`[^a]`, `cc{0x0-0x60 0x62-0x10ffff}`},
   252  	{`[a\n]`, `cc{0xa 0x61}`},
   253  }
   254  
   255  func TestParseMatchNL(t *testing.T) {
   256  	testParseDump(t, matchnlTests, MatchNL)
   257  }
   258  
   259  var nomatchnlTests = []parseTest{
   260  	{`.`, `dnl{}`},
   261  	{"\n", "lit{\n}"},
   262  	{`[^a]`, `cc{0x0-0x9 0xb-0x60 0x62-0x10ffff}`},
   263  	{`[a\n]`, `cc{0xa 0x61}`},
   264  }
   265  
   266  func TestParseNoMatchNL(t *testing.T) {
   267  	testParseDump(t, nomatchnlTests, 0)
   268  }
   269  
   270  // Test Parse -> Dump.
   271  func testParseDump(t *testing.T, tests []parseTest, flags Flags) {
   272  	for _, tt := range tests {
   273  		re, err := Parse(tt.Regexp, flags)
   274  		if err != nil {
   275  			t.Errorf("Parse(%#q): %v", tt.Regexp, err)
   276  			continue
   277  		}
   278  		if tt.Dump == "" {
   279  			// It parsed. That's all we care about.
   280  			continue
   281  		}
   282  		d := dump(re)
   283  		if d != tt.Dump {
   284  			t.Errorf("Parse(%#q).Dump() = %#q want %#q", tt.Regexp, d, tt.Dump)
   285  		}
   286  	}
   287  }
   288  
   289  // dump prints a string representation of the regexp showing
   290  // the structure explicitly.
   291  func dump(re *Regexp) string {
   292  	var b strings.Builder
   293  	dumpRegexp(&b, re)
   294  	return b.String()
   295  }
   296  
   297  var opNames = []string{
   298  	OpNoMatch:        "no",
   299  	OpEmptyMatch:     "emp",
   300  	OpLiteral:        "lit",
   301  	OpCharClass:      "cc",
   302  	OpAnyCharNotNL:   "dnl",
   303  	OpAnyChar:        "dot",
   304  	OpBeginLine:      "bol",
   305  	OpEndLine:        "eol",
   306  	OpBeginText:      "bot",
   307  	OpEndText:        "eot",
   308  	OpWordBoundary:   "wb",
   309  	OpNoWordBoundary: "nwb",
   310  	OpCapture:        "cap",
   311  	OpStar:           "star",
   312  	OpPlus:           "plus",
   313  	OpQuest:          "que",
   314  	OpRepeat:         "rep",
   315  	OpConcat:         "cat",
   316  	OpAlternate:      "alt",
   317  }
   318  
   319  // dumpRegexp writes an encoding of the syntax tree for the regexp re to b.
   320  // It is used during testing to distinguish between parses that might print
   321  // the same using re's String method.
   322  func dumpRegexp(b *strings.Builder, re *Regexp) {
   323  	if int(re.Op) >= len(opNames) || opNames[re.Op] == "" {
   324  		fmt.Fprintf(b, "op%d", re.Op)
   325  	} else {
   326  		switch re.Op {
   327  		default:
   328  			b.WriteString(opNames[re.Op])
   329  		case OpStar, OpPlus, OpQuest, OpRepeat:
   330  			if re.Flags&NonGreedy != 0 {
   331  				b.WriteByte('n')
   332  			}
   333  			b.WriteString(opNames[re.Op])
   334  		case OpLiteral:
   335  			if len(re.Rune) > 1 {
   336  				b.WriteString("str")
   337  			} else {
   338  				b.WriteString("lit")
   339  			}
   340  			if re.Flags&FoldCase != 0 {
   341  				for _, r := range re.Rune {
   342  					if unicode.SimpleFold(r) != r {
   343  						b.WriteString("fold")
   344  						break
   345  					}
   346  				}
   347  			}
   348  		}
   349  	}
   350  	b.WriteByte('{')
   351  	switch re.Op {
   352  	case OpEndText:
   353  		if re.Flags&WasDollar == 0 {
   354  			b.WriteString(`\z`)
   355  		}
   356  	case OpLiteral:
   357  		for _, r := range re.Rune {
   358  			b.WriteRune(r)
   359  		}
   360  	case OpConcat, OpAlternate:
   361  		for _, sub := range re.Sub {
   362  			dumpRegexp(b, sub)
   363  		}
   364  	case OpStar, OpPlus, OpQuest:
   365  		dumpRegexp(b, re.Sub[0])
   366  	case OpRepeat:
   367  		fmt.Fprintf(b, "%d,%d ", re.Min, re.Max)
   368  		dumpRegexp(b, re.Sub[0])
   369  	case OpCapture:
   370  		if re.Name != "" {
   371  			b.WriteString(re.Name)
   372  			b.WriteByte(':')
   373  		}
   374  		dumpRegexp(b, re.Sub[0])
   375  	case OpCharClass:
   376  		sep := ""
   377  		for i := 0; i < len(re.Rune); i += 2 {
   378  			b.WriteString(sep)
   379  			sep = " "
   380  			lo, hi := re.Rune[i], re.Rune[i+1]
   381  			if lo == hi {
   382  				fmt.Fprintf(b, "%#x", lo)
   383  			} else {
   384  				fmt.Fprintf(b, "%#x-%#x", lo, hi)
   385  			}
   386  		}
   387  	}
   388  	b.WriteByte('}')
   389  }
   390  
   391  func mkCharClass(f func(rune) bool) string {
   392  	re := &Regexp{Op: OpCharClass}
   393  	lo := rune(-1)
   394  	for i := rune(0); i <= unicode.MaxRune; i++ {
   395  		if f(i) {
   396  			if lo < 0 {
   397  				lo = i
   398  			}
   399  		} else {
   400  			if lo >= 0 {
   401  				re.Rune = append(re.Rune, lo, i-1)
   402  				lo = -1
   403  			}
   404  		}
   405  	}
   406  	if lo >= 0 {
   407  		re.Rune = append(re.Rune, lo, unicode.MaxRune)
   408  	}
   409  	return dump(re)
   410  }
   411  
   412  func isUpperFold(r rune) bool {
   413  	if unicode.IsUpper(r) {
   414  		return true
   415  	}
   416  	c := unicode.SimpleFold(r)
   417  	for c != r {
   418  		if unicode.IsUpper(c) {
   419  			return true
   420  		}
   421  		c = unicode.SimpleFold(c)
   422  	}
   423  	return false
   424  }
   425  
   426  func TestFoldConstants(t *testing.T) {
   427  	last := rune(-1)
   428  	for i := rune(0); i <= unicode.MaxRune; i++ {
   429  		if unicode.SimpleFold(i) == i {
   430  			continue
   431  		}
   432  		if last == -1 && minFold != i {
   433  			t.Errorf("minFold=%#U should be %#U", minFold, i)
   434  		}
   435  		last = i
   436  	}
   437  	if maxFold != last {
   438  		t.Errorf("maxFold=%#U should be %#U", maxFold, last)
   439  	}
   440  }
   441  
   442  func TestAppendRangeCollapse(t *testing.T) {
   443  	// AppendRange should collapse each of the new ranges
   444  	// into the earlier ones (it looks back two ranges), so that
   445  	// the slice never grows very large.
   446  	// Note that we are not calling cleanClass.
   447  	var r []rune
   448  	for i := rune('A'); i <= 'Z'; i++ {
   449  		r = appendRange(r, i, i)
   450  		r = appendRange(r, i+'a'-'A', i+'a'-'A')
   451  	}
   452  	if string(r) != "AZaz" {
   453  		t.Errorf("appendRange interlaced A-Z a-z = %s, want AZaz", string(r))
   454  	}
   455  }
   456  
   457  var invalidRegexps = []string{
   458  	`(`,
   459  	`)`,
   460  	`(a`,
   461  	`a)`,
   462  	`(a))`,
   463  	`(a|b|`,
   464  	`a|b|)`,
   465  	`(a|b|))`,
   466  	`(a|b`,
   467  	`a|b)`,
   468  	`(a|b))`,
   469  	`[a-z`,
   470  	`([a-z)`,
   471  	`[a-z)`,
   472  	`([a-z]))`,
   473  	`x{1001}`,
   474  	`x{9876543210}`,
   475  	`x{2,1}`,
   476  	`x{1,9876543210}`,
   477  	"\xff", // Invalid UTF-8
   478  	"[\xff]",
   479  	"[\\\xff]",
   480  	"\\\xff",
   481  	`(?P<name>a`,
   482  	`(?P<name>`,
   483  	`(?P<name`,
   484  	`(?P<x y>a)`,
   485  	`(?P<>a)`,
   486  	`(?<name>a`,
   487  	`(?<name>`,
   488  	`(?<name`,
   489  	`(?<x y>a)`,
   490  	`(?<>a)`,
   491  	`[a-Z]`,
   492  	`(?i)[a-Z]`,
   493  	`\Q\E*`,
   494  	`a{100000}`,  // too much repetition
   495  	`a{100000,}`, // too much repetition
   496  	"((((((((((x{2}){2}){2}){2}){2}){2}){2}){2}){2}){2})",    // too much repetition
   497  	strings.Repeat("(", 1000) + strings.Repeat(")", 1000),    // too deep
   498  	strings.Repeat("(?:", 1000) + strings.Repeat(")*", 1000), // too deep
   499  	"(" + strings.Repeat("(xx?)", 1000) + "){1000}",          // too long
   500  	strings.Repeat("(xx?){1000}", 1000),                      // too long
   501  	strings.Repeat(`\pL`, 27000),                             // too many runes
   502  }
   503  
   504  var onlyPerl = []string{
   505  	`[a-b-c]`,
   506  	`\Qabc\E`,
   507  	`\Q*+?{[\E`,
   508  	`\Q\\E`,
   509  	`\Q\\\E`,
   510  	`\Q\\\\E`,
   511  	`\Q\\\\\E`,
   512  	`(?:a)`,
   513  	`(?P<name>a)`,
   514  }
   515  
   516  var onlyPOSIX = []string{
   517  	"a++",
   518  	"a**",
   519  	"a?*",
   520  	"a+*",
   521  	"a{1}*",
   522  	".{1}{2}.{3}",
   523  }
   524  
   525  func TestParseInvalidRegexps(t *testing.T) {
   526  	for _, regexp := range invalidRegexps {
   527  		if re, err := Parse(regexp, Perl); err == nil {
   528  			t.Errorf("Parse(%#q, Perl) = %s, should have failed", regexp, dump(re))
   529  		}
   530  		if re, err := Parse(regexp, POSIX); err == nil {
   531  			t.Errorf("Parse(%#q, POSIX) = %s, should have failed", regexp, dump(re))
   532  		}
   533  	}
   534  	for _, regexp := range onlyPerl {
   535  		if _, err := Parse(regexp, Perl); err != nil {
   536  			t.Errorf("Parse(%#q, Perl): %v", regexp, err)
   537  		}
   538  		if re, err := Parse(regexp, POSIX); err == nil {
   539  			t.Errorf("Parse(%#q, POSIX) = %s, should have failed", regexp, dump(re))
   540  		}
   541  	}
   542  	for _, regexp := range onlyPOSIX {
   543  		if re, err := Parse(regexp, Perl); err == nil {
   544  			t.Errorf("Parse(%#q, Perl) = %s, should have failed", regexp, dump(re))
   545  		}
   546  		if _, err := Parse(regexp, POSIX); err != nil {
   547  			t.Errorf("Parse(%#q, POSIX): %v", regexp, err)
   548  		}
   549  	}
   550  }
   551  
   552  func TestToStringEquivalentParse(t *testing.T) {
   553  	for _, tt := range parseTests {
   554  		re, err := Parse(tt.Regexp, testFlags)
   555  		if err != nil {
   556  			t.Errorf("Parse(%#q): %v", tt.Regexp, err)
   557  			continue
   558  		}
   559  		if tt.Dump == "" {
   560  			// It parsed. That's all we care about.
   561  			continue
   562  		}
   563  		d := dump(re)
   564  		if d != tt.Dump {
   565  			t.Errorf("Parse(%#q).Dump() = %#q want %#q", tt.Regexp, d, tt.Dump)
   566  			continue
   567  		}
   568  
   569  		s := re.String()
   570  		if s != tt.Regexp {
   571  			// If ToString didn't return the original regexp,
   572  			// it must have found one with fewer parens.
   573  			// Unfortunately we can't check the length here, because
   574  			// ToString produces "\\{" for a literal brace,
   575  			// but "{" is a shorter equivalent in some contexts.
   576  			nre, err := Parse(s, testFlags)
   577  			if err != nil {
   578  				t.Errorf("Parse(%#q.String() = %#q): %v", tt.Regexp, s, err)
   579  				continue
   580  			}
   581  			nd := dump(nre)
   582  			if d != nd {
   583  				t.Errorf("Parse(%#q) -> %#q; %#q vs %#q", tt.Regexp, s, d, nd)
   584  			}
   585  
   586  			ns := nre.String()
   587  			if s != ns {
   588  				t.Errorf("Parse(%#q) -> %#q -> %#q", tt.Regexp, s, ns)
   589  			}
   590  		}
   591  	}
   592  }
   593  
   594  var stringTests = []struct {
   595  	re  string
   596  	out string
   597  }{
   598  	{`x(?i:ab*c|d?e)1`, `x(?i:AB*C|D?E)1`},
   599  	{`x(?i:ab*cd?e)1`, `x(?i:AB*CD?E)1`},
   600  	{`0(?i:ab*c|d?e)1`, `(?i:0(?:AB*C|D?E)1)`},
   601  	{`0(?i:ab*cd?e)1`, `(?i:0AB*CD?E1)`},
   602  	{`x(?i:ab*c|d?e)`, `x(?i:AB*C|D?E)`},
   603  	{`x(?i:ab*cd?e)`, `x(?i:AB*CD?E)`},
   604  	{`0(?i:ab*c|d?e)`, `(?i:0(?:AB*C|D?E))`},
   605  	{`0(?i:ab*cd?e)`, `(?i:0AB*CD?E)`},
   606  	{`(?i:ab*c|d?e)1`, `(?i:(?:AB*C|D?E)1)`},
   607  	{`(?i:ab*cd?e)1`, `(?i:AB*CD?E1)`},
   608  	{`(?i:ab)[123](?i:cd)`, `(?i:AB[1-3]CD)`},
   609  	{`(?i:ab*c|d?e)`, `(?i:AB*C|D?E)`},
   610  	{`[Aa][Bb]`, `(?i:AB)`},
   611  	{`[Aa][Bb]*[Cc]`, `(?i:AB*C)`},
   612  	{`A(?:[Bb][Cc]|[Dd])[Zz]`, `A(?i:(?:BC|D)Z)`},
   613  	{`[Aa](?:[Bb][Cc]|[Dd])Z`, `(?i:A(?:BC|D))Z`},
   614  }
   615  
   616  func TestString(t *testing.T) {
   617  	for _, tt := range stringTests {
   618  		re, err := Parse(tt.re, Perl)
   619  		if err != nil {
   620  			t.Errorf("Parse(%#q): %v", tt.re, err)
   621  			continue
   622  		}
   623  		out := re.String()
   624  		if out != tt.out {
   625  			t.Errorf("Parse(%#q).String() = %#q, want %#q", tt.re, out, tt.out)
   626  		}
   627  	}
   628  }
   629  

View as plain text