| 1 |
747 |
jeremybenn |
// 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_test
|
| 6 |
|
|
|
| 7 |
|
|
import . "regexp/syntax"
|
| 8 |
|
|
import "testing"
|
| 9 |
|
|
|
| 10 |
|
|
var simplifyTests = []struct {
|
| 11 |
|
|
Regexp string
|
| 12 |
|
|
Simple string
|
| 13 |
|
|
}{
|
| 14 |
|
|
// Already-simple constructs
|
| 15 |
|
|
{`a`, `a`},
|
| 16 |
|
|
{`ab`, `ab`},
|
| 17 |
|
|
{`a|b`, `[a-b]`},
|
| 18 |
|
|
{`ab|cd`, `ab|cd`},
|
| 19 |
|
|
{`(ab)*`, `(ab)*`},
|
| 20 |
|
|
{`(ab)+`, `(ab)+`},
|
| 21 |
|
|
{`(ab)?`, `(ab)?`},
|
| 22 |
|
|
{`.`, `(?s:.)`},
|
| 23 |
|
|
{`^`, `^`},
|
| 24 |
|
|
{`$`, `$`},
|
| 25 |
|
|
{`[ac]`, `[ac]`},
|
| 26 |
|
|
{`[^ac]`, `[^ac]`},
|
| 27 |
|
|
|
| 28 |
|
|
// Posix character classes
|
| 29 |
|
|
{`[[:alnum:]]`, `[0-9A-Za-z]`},
|
| 30 |
|
|
{`[[:alpha:]]`, `[A-Za-z]`},
|
| 31 |
|
|
{`[[:blank:]]`, `[\t ]`},
|
| 32 |
|
|
{`[[:cntrl:]]`, `[\x00-\x1f\x7f]`},
|
| 33 |
|
|
{`[[:digit:]]`, `[0-9]`},
|
| 34 |
|
|
{`[[:graph:]]`, `[!-~]`},
|
| 35 |
|
|
{`[[:lower:]]`, `[a-z]`},
|
| 36 |
|
|
{`[[:print:]]`, `[ -~]`},
|
| 37 |
|
|
{`[[:punct:]]`, "[!-/:-@\\[-`\\{-~]"},
|
| 38 |
|
|
{`[[:space:]]`, `[\t-\r ]`},
|
| 39 |
|
|
{`[[:upper:]]`, `[A-Z]`},
|
| 40 |
|
|
{`[[:xdigit:]]`, `[0-9A-Fa-f]`},
|
| 41 |
|
|
|
| 42 |
|
|
// Perl character classes
|
| 43 |
|
|
{`\d`, `[0-9]`},
|
| 44 |
|
|
{`\s`, `[\t-\n\f-\r ]`},
|
| 45 |
|
|
{`\w`, `[0-9A-Z_a-z]`},
|
| 46 |
|
|
{`\D`, `[^0-9]`},
|
| 47 |
|
|
{`\S`, `[^\t-\n\f-\r ]`},
|
| 48 |
|
|
{`\W`, `[^0-9A-Z_a-z]`},
|
| 49 |
|
|
{`[\d]`, `[0-9]`},
|
| 50 |
|
|
{`[\s]`, `[\t-\n\f-\r ]`},
|
| 51 |
|
|
{`[\w]`, `[0-9A-Z_a-z]`},
|
| 52 |
|
|
{`[\D]`, `[^0-9]`},
|
| 53 |
|
|
{`[\S]`, `[^\t-\n\f-\r ]`},
|
| 54 |
|
|
{`[\W]`, `[^0-9A-Z_a-z]`},
|
| 55 |
|
|
|
| 56 |
|
|
// Posix repetitions
|
| 57 |
|
|
{`a{1}`, `a`},
|
| 58 |
|
|
{`a{2}`, `aa`},
|
| 59 |
|
|
{`a{5}`, `aaaaa`},
|
| 60 |
|
|
{`a{0,1}`, `a?`},
|
| 61 |
|
|
// The next three are illegible because Simplify inserts (?:)
|
| 62 |
|
|
// parens instead of () parens to avoid creating extra
|
| 63 |
|
|
// captured subexpressions. The comments show a version with fewer parens.
|
| 64 |
|
|
{`(a){0,2}`, `(?:(a)(a)?)?`}, // (aa?)?
|
| 65 |
|
|
{`(a){0,4}`, `(?:(a)(?:(a)(?:(a)(a)?)?)?)?`}, // (a(a(aa?)?)?)?
|
| 66 |
|
|
{`(a){2,6}`, `(a)(a)(?:(a)(?:(a)(?:(a)(a)?)?)?)?`}, // aa(a(a(aa?)?)?)?
|
| 67 |
|
|
{`a{0,2}`, `(?:aa?)?`}, // (aa?)?
|
| 68 |
|
|
{`a{0,4}`, `(?:a(?:a(?:aa?)?)?)?`}, // (a(a(aa?)?)?)?
|
| 69 |
|
|
{`a{2,6}`, `aa(?:a(?:a(?:aa?)?)?)?`}, // aa(a(a(aa?)?)?)?
|
| 70 |
|
|
{`a{0,}`, `a*`},
|
| 71 |
|
|
{`a{1,}`, `a+`},
|
| 72 |
|
|
{`a{2,}`, `aa+`},
|
| 73 |
|
|
{`a{5,}`, `aaaaa+`},
|
| 74 |
|
|
|
| 75 |
|
|
// Test that operators simplify their arguments.
|
| 76 |
|
|
{`(?:a{1,}){1,}`, `a+`},
|
| 77 |
|
|
{`(a{1,}b{1,})`, `(a+b+)`},
|
| 78 |
|
|
{`a{1,}|b{1,}`, `a+|b+`},
|
| 79 |
|
|
{`(?:a{1,})*`, `(?:a+)*`},
|
| 80 |
|
|
{`(?:a{1,})+`, `a+`},
|
| 81 |
|
|
{`(?:a{1,})?`, `(?:a+)?`},
|
| 82 |
|
|
{``, `(?:)`},
|
| 83 |
|
|
{`a{0}`, `(?:)`},
|
| 84 |
|
|
|
| 85 |
|
|
// Character class simplification
|
| 86 |
|
|
{`[ab]`, `[a-b]`},
|
| 87 |
|
|
{`[a-za-za-z]`, `[a-z]`},
|
| 88 |
|
|
{`[A-Za-zA-Za-z]`, `[A-Za-z]`},
|
| 89 |
|
|
{`[ABCDEFGH]`, `[A-H]`},
|
| 90 |
|
|
{`[AB-CD-EF-GH]`, `[A-H]`},
|
| 91 |
|
|
{`[W-ZP-XE-R]`, `[E-Z]`},
|
| 92 |
|
|
{`[a-ee-gg-m]`, `[a-m]`},
|
| 93 |
|
|
{`[a-ea-ha-m]`, `[a-m]`},
|
| 94 |
|
|
{`[a-ma-ha-e]`, `[a-m]`},
|
| 95 |
|
|
{`[a-zA-Z0-9 -~]`, `[ -~]`},
|
| 96 |
|
|
|
| 97 |
|
|
// Empty character classes
|
| 98 |
|
|
{`[^[:cntrl:][:^cntrl:]]`, `[^\x00-\x{10FFFF}]`},
|
| 99 |
|
|
|
| 100 |
|
|
// Full character classes
|
| 101 |
|
|
{`[[:cntrl:][:^cntrl:]]`, `(?s:.)`},
|
| 102 |
|
|
|
| 103 |
|
|
// Unicode case folding.
|
| 104 |
|
|
{`(?i)A`, `(?i:A)`},
|
| 105 |
|
|
{`(?i)a`, `(?i:A)`},
|
| 106 |
|
|
{`(?i)[A]`, `(?i:A)`},
|
| 107 |
|
|
{`(?i)[a]`, `(?i:A)`},
|
| 108 |
|
|
{`(?i)K`, `(?i:K)`},
|
| 109 |
|
|
{`(?i)k`, `(?i:K)`},
|
| 110 |
|
|
{`(?i)\x{212a}`, "(?i:K)"},
|
| 111 |
|
|
{`(?i)[K]`, "[Kk\u212A]"},
|
| 112 |
|
|
{`(?i)[k]`, "[Kk\u212A]"},
|
| 113 |
|
|
{`(?i)[\x{212a}]`, "[Kk\u212A]"},
|
| 114 |
|
|
{`(?i)[a-z]`, "[A-Za-z\u017F\u212A]"},
|
| 115 |
|
|
{`(?i)[\x00-\x{FFFD}]`, "[\\x00-\uFFFD]"},
|
| 116 |
|
|
{`(?i)[\x00-\x{10FFFF}]`, `(?s:.)`},
|
| 117 |
|
|
|
| 118 |
|
|
// Empty string as a regular expression.
|
| 119 |
|
|
// The empty string must be preserved inside parens in order
|
| 120 |
|
|
// to make submatches work right, so these tests are less
|
| 121 |
|
|
// interesting than they might otherwise be. String inserts
|
| 122 |
|
|
// explicit (?:) in place of non-parenthesized empty strings,
|
| 123 |
|
|
// to make them easier to spot for other parsers.
|
| 124 |
|
|
{`(a|b|)`, `([a-b]|(?:))`},
|
| 125 |
|
|
{`(|)`, `()`},
|
| 126 |
|
|
{`a()`, `a()`},
|
| 127 |
|
|
{`(()|())`, `(()|())`},
|
| 128 |
|
|
{`(a|)`, `(a|(?:))`},
|
| 129 |
|
|
{`ab()cd()`, `ab()cd()`},
|
| 130 |
|
|
{`()`, `()`},
|
| 131 |
|
|
{`()*`, `()*`},
|
| 132 |
|
|
{`()+`, `()+`},
|
| 133 |
|
|
{`()?`, `()?`},
|
| 134 |
|
|
{`(){0}`, `(?:)`},
|
| 135 |
|
|
{`(){1}`, `()`},
|
| 136 |
|
|
{`(){1,}`, `()+`},
|
| 137 |
|
|
{`(){0,2}`, `(?:()()?)?`},
|
| 138 |
|
|
}
|
| 139 |
|
|
|
| 140 |
|
|
func TestSimplify(t *testing.T) {
|
| 141 |
|
|
for _, tt := range simplifyTests {
|
| 142 |
|
|
re, err := Parse(tt.Regexp, MatchNL|Perl&^OneLine)
|
| 143 |
|
|
if err != nil {
|
| 144 |
|
|
t.Errorf("Parse(%#q) = error %v", tt.Regexp, err)
|
| 145 |
|
|
continue
|
| 146 |
|
|
}
|
| 147 |
|
|
s := re.Simplify().String()
|
| 148 |
|
|
if s != tt.Simple {
|
| 149 |
|
|
t.Errorf("Simplify(%#q) = %#q, want %#q", tt.Regexp, s, tt.Simple)
|
| 150 |
|
|
}
|
| 151 |
|
|
}
|
| 152 |
|
|
}
|
