...
Run Format

Source file src/regexp/exec_test.go

Documentation: regexp

  // Copyright 2010 The Go Authors. All rights reserved.
  // Use of this source code is governed by a BSD-style
  // license that can be found in the LICENSE file.
  
  package regexp
  
  import (
  	"bufio"
  	"compress/bzip2"
  	"fmt"
  	"internal/testenv"
  	"io"
  	"os"
  	"path/filepath"
  	"regexp/syntax"
  	"strconv"
  	"strings"
  	"testing"
  	"unicode/utf8"
  )
  
  // TestRE2 tests this package's regexp API against test cases
  // considered during RE2's exhaustive tests, which run all possible
  // regexps over a given set of atoms and operators, up to a given
  // complexity, over all possible strings over a given alphabet,
  // up to a given size. Rather than try to link with RE2, we read a
  // log file containing the test cases and the expected matches.
  // The log file, re2-exhaustive.txt, is generated by running 'make log'
  // in the open source RE2 distribution https://github.com/google/re2/.
  //
  // The test file format is a sequence of stanzas like:
  //
  //	strings
  //	"abc"
  //	"123x"
  //	regexps
  //	"[a-z]+"
  //	0-3;0-3
  //	-;-
  //	"([0-9])([0-9])([0-9])"
  //	-;-
  //	-;0-3 0-1 1-2 2-3
  //
  // The stanza begins by defining a set of strings, quoted
  // using Go double-quote syntax, one per line. Then the
  // regexps section gives a sequence of regexps to run on
  // the strings. In the block that follows a regexp, each line
  // gives the semicolon-separated match results of running
  // the regexp on the corresponding string.
  // Each match result is either a single -, meaning no match, or a
  // space-separated sequence of pairs giving the match and
  // submatch indices. An unmatched subexpression formats
  // its pair as a single - (not illustrated above).  For now
  // each regexp run produces two match results, one for a
  // ``full match'' that restricts the regexp to matching the entire
  // string or nothing, and one for a ``partial match'' that gives
  // the leftmost first match found in the string.
  //
  // Lines beginning with # are comments. Lines beginning with
  // a capital letter are test names printed during RE2's test suite
  // and are echoed into t but otherwise ignored.
  //
  // At time of writing, re2-exhaustive.txt is 59 MB but compresses to 385 kB,
  // so we store re2-exhaustive.txt.bz2 in the repository and decompress it on the fly.
  //
  func TestRE2Search(t *testing.T) {
  	testRE2(t, "testdata/re2-search.txt")
  }
  
  func testRE2(t *testing.T, file string) {
  	f, err := os.Open(file)
  	if err != nil {
  		t.Fatal(err)
  	}
  	defer f.Close()
  	var txt io.Reader
  	if strings.HasSuffix(file, ".bz2") {
  		z := bzip2.NewReader(f)
  		txt = z
  		file = file[:len(file)-len(".bz2")] // for error messages
  	} else {
  		txt = f
  	}
  	lineno := 0
  	scanner := bufio.NewScanner(txt)
  	var (
  		str       []string
  		input     []string
  		inStrings bool
  		re        *Regexp
  		refull    *Regexp
  		nfail     int
  		ncase     int
  	)
  	for lineno := 1; scanner.Scan(); lineno++ {
  		line := scanner.Text()
  		switch {
  		case line == "":
  			t.Fatalf("%s:%d: unexpected blank line", file, lineno)
  		case line[0] == '#':
  			continue
  		case 'A' <= line[0] && line[0] <= 'Z':
  			// Test name.
  			t.Logf("%s\n", line)
  			continue
  		case line == "strings":
  			str = str[:0]
  			inStrings = true
  		case line == "regexps":
  			inStrings = false
  		case line[0] == '"':
  			q, err := strconv.Unquote(line)
  			if err != nil {
  				// Fatal because we'll get out of sync.
  				t.Fatalf("%s:%d: unquote %s: %v", file, lineno, line, err)
  			}
  			if inStrings {
  				str = append(str, q)
  				continue
  			}
  			// Is a regexp.
  			if len(input) != 0 {
  				t.Fatalf("%s:%d: out of sync: have %d strings left before %#q", file, lineno, len(input), q)
  			}
  			re, err = tryCompile(q)
  			if err != nil {
  				if err.Error() == "error parsing regexp: invalid escape sequence: `\\C`" {
  					// We don't and likely never will support \C; keep going.
  					continue
  				}
  				t.Errorf("%s:%d: compile %#q: %v", file, lineno, q, err)
  				if nfail++; nfail >= 100 {
  					t.Fatalf("stopping after %d errors", nfail)
  				}
  				continue
  			}
  			full := `\A(?:` + q + `)\z`
  			refull, err = tryCompile(full)
  			if err != nil {
  				// Fatal because q worked, so this should always work.
  				t.Fatalf("%s:%d: compile full %#q: %v", file, lineno, full, err)
  			}
  			input = str
  		case line[0] == '-' || '0' <= line[0] && line[0] <= '9':
  			// A sequence of match results.
  			ncase++
  			if re == nil {
  				// Failed to compile: skip results.
  				continue
  			}
  			if len(input) == 0 {
  				t.Fatalf("%s:%d: out of sync: no input remaining", file, lineno)
  			}
  			var text string
  			text, input = input[0], input[1:]
  			if !isSingleBytes(text) && strings.Contains(re.String(), `\B`) {
  				// RE2's \B considers every byte position,
  				// so it sees 'not word boundary' in the
  				// middle of UTF-8 sequences. This package
  				// only considers the positions between runes,
  				// so it disagrees. Skip those cases.
  				continue
  			}
  			res := strings.Split(line, ";")
  			if len(res) != len(run) {
  				t.Fatalf("%s:%d: have %d test results, want %d", file, lineno, len(res), len(run))
  			}
  			for i := range res {
  				have, suffix := run[i](re, refull, text)
  				want := parseResult(t, file, lineno, res[i])
  				if !same(have, want) {
  					t.Errorf("%s:%d: %#q%s.FindSubmatchIndex(%#q) = %v, want %v", file, lineno, re, suffix, text, have, want)
  					if nfail++; nfail >= 100 {
  						t.Fatalf("stopping after %d errors", nfail)
  					}
  					continue
  				}
  				b, suffix := match[i](re, refull, text)
  				if b != (want != nil) {
  					t.Errorf("%s:%d: %#q%s.MatchString(%#q) = %v, want %v", file, lineno, re, suffix, text, b, !b)
  					if nfail++; nfail >= 100 {
  						t.Fatalf("stopping after %d errors", nfail)
  					}
  					continue
  				}
  			}
  
  		default:
  			t.Fatalf("%s:%d: out of sync: %s\n", file, lineno, line)
  		}
  	}
  	if err := scanner.Err(); err != nil {
  		t.Fatalf("%s:%d: %v", file, lineno, err)
  	}
  	if len(input) != 0 {
  		t.Fatalf("%s:%d: out of sync: have %d strings left at EOF", file, lineno, len(input))
  	}
  	t.Logf("%d cases tested", ncase)
  }
  
  var run = []func(*Regexp, *Regexp, string) ([]int, string){
  	runFull,
  	runPartial,
  	runFullLongest,
  	runPartialLongest,
  }
  
  func runFull(re, refull *Regexp, text string) ([]int, string) {
  	refull.longest = false
  	return refull.FindStringSubmatchIndex(text), "[full]"
  }
  
  func runPartial(re, refull *Regexp, text string) ([]int, string) {
  	re.longest = false
  	return re.FindStringSubmatchIndex(text), ""
  }
  
  func runFullLongest(re, refull *Regexp, text string) ([]int, string) {
  	refull.longest = true
  	return refull.FindStringSubmatchIndex(text), "[full,longest]"
  }
  
  func runPartialLongest(re, refull *Regexp, text string) ([]int, string) {
  	re.longest = true
  	return re.FindStringSubmatchIndex(text), "[longest]"
  }
  
  var match = []func(*Regexp, *Regexp, string) (bool, string){
  	matchFull,
  	matchPartial,
  	matchFullLongest,
  	matchPartialLongest,
  }
  
  func matchFull(re, refull *Regexp, text string) (bool, string) {
  	refull.longest = false
  	return refull.MatchString(text), "[full]"
  }
  
  func matchPartial(re, refull *Regexp, text string) (bool, string) {
  	re.longest = false
  	return re.MatchString(text), ""
  }
  
  func matchFullLongest(re, refull *Regexp, text string) (bool, string) {
  	refull.longest = true
  	return refull.MatchString(text), "[full,longest]"
  }
  
  func matchPartialLongest(re, refull *Regexp, text string) (bool, string) {
  	re.longest = true
  	return re.MatchString(text), "[longest]"
  }
  
  func isSingleBytes(s string) bool {
  	for _, c := range s {
  		if c >= utf8.RuneSelf {
  			return false
  		}
  	}
  	return true
  }
  
  func tryCompile(s string) (re *Regexp, err error) {
  	// Protect against panic during Compile.
  	defer func() {
  		if r := recover(); r != nil {
  			err = fmt.Errorf("panic: %v", r)
  		}
  	}()
  	return Compile(s)
  }
  
  func parseResult(t *testing.T, file string, lineno int, res string) []int {
  	// A single - indicates no match.
  	if res == "-" {
  		return nil
  	}
  	// Otherwise, a space-separated list of pairs.
  	n := 1
  	for j := 0; j < len(res); j++ {
  		if res[j] == ' ' {
  			n++
  		}
  	}
  	out := make([]int, 2*n)
  	i := 0
  	n = 0
  	for j := 0; j <= len(res); j++ {
  		if j == len(res) || res[j] == ' ' {
  			// Process a single pair.  - means no submatch.
  			pair := res[i:j]
  			if pair == "-" {
  				out[n] = -1
  				out[n+1] = -1
  			} else {
  				k := strings.Index(pair, "-")
  				if k < 0 {
  					t.Fatalf("%s:%d: invalid pair %s", file, lineno, pair)
  				}
  				lo, err1 := strconv.Atoi(pair[:k])
  				hi, err2 := strconv.Atoi(pair[k+1:])
  				if err1 != nil || err2 != nil || lo > hi {
  					t.Fatalf("%s:%d: invalid pair %s", file, lineno, pair)
  				}
  				out[n] = lo
  				out[n+1] = hi
  			}
  			n += 2
  			i = j + 1
  		}
  	}
  	return out
  }
  
  func same(x, y []int) bool {
  	if len(x) != len(y) {
  		return false
  	}
  	for i, xi := range x {
  		if xi != y[i] {
  			return false
  		}
  	}
  	return true
  }
  
  // TestFowler runs this package's regexp API against the
  // POSIX regular expression tests collected by Glenn Fowler
  // at http://www2.research.att.com/~astopen/testregex/testregex.html.
  func TestFowler(t *testing.T) {
  	files, err := filepath.Glob("testdata/*.dat")
  	if err != nil {
  		t.Fatal(err)
  	}
  	for _, file := range files {
  		t.Log(file)
  		testFowler(t, file)
  	}
  }
  
  var notab = MustCompilePOSIX(`[^\t]+`)
  
  func testFowler(t *testing.T, file string) {
  	f, err := os.Open(file)
  	if err != nil {
  		t.Error(err)
  		return
  	}
  	defer f.Close()
  	b := bufio.NewReader(f)
  	lineno := 0
  	lastRegexp := ""
  Reading:
  	for {
  		lineno++
  		line, err := b.ReadString('\n')
  		if err != nil {
  			if err != io.EOF {
  				t.Errorf("%s:%d: %v", file, lineno, err)
  			}
  			break Reading
  		}
  
  		// http://www2.research.att.com/~astopen/man/man1/testregex.html
  		//
  		// INPUT FORMAT
  		//   Input lines may be blank, a comment beginning with #, or a test
  		//   specification. A specification is five fields separated by one
  		//   or more tabs. NULL denotes the empty string and NIL denotes the
  		//   0 pointer.
  		if line[0] == '#' || line[0] == '\n' {
  			continue Reading
  		}
  		line = line[:len(line)-1]
  		field := notab.FindAllString(line, -1)
  		for i, f := range field {
  			if f == "NULL" {
  				field[i] = ""
  			}
  			if f == "NIL" {
  				t.Logf("%s:%d: skip: %s", file, lineno, line)
  				continue Reading
  			}
  		}
  		if len(field) == 0 {
  			continue Reading
  		}
  
  		//   Field 1: the regex(3) flags to apply, one character per REG_feature
  		//   flag. The test is skipped if REG_feature is not supported by the
  		//   implementation. If the first character is not [BEASKLP] then the
  		//   specification is a global control line. One or more of [BEASKLP] may be
  		//   specified; the test will be repeated for each mode.
  		//
  		//     B 	basic			BRE	(grep, ed, sed)
  		//     E 	REG_EXTENDED		ERE	(egrep)
  		//     A	REG_AUGMENTED		ARE	(egrep with negation)
  		//     S	REG_SHELL		SRE	(sh glob)
  		//     K	REG_SHELL|REG_AUGMENTED	KRE	(ksh glob)
  		//     L	REG_LITERAL		LRE	(fgrep)
  		//
  		//     a	REG_LEFT|REG_RIGHT	implicit ^...$
  		//     b	REG_NOTBOL		lhs does not match ^
  		//     c	REG_COMMENT		ignore space and #...\n
  		//     d	REG_SHELL_DOT		explicit leading . match
  		//     e	REG_NOTEOL		rhs does not match $
  		//     f	REG_MULTIPLE		multiple \n separated patterns
  		//     g	FNM_LEADING_DIR		testfnmatch only -- match until /
  		//     h	REG_MULTIREF		multiple digit backref
  		//     i	REG_ICASE		ignore case
  		//     j	REG_SPAN		. matches \n
  		//     k	REG_ESCAPE		\ to escape [...] delimiter
  		//     l	REG_LEFT		implicit ^...
  		//     m	REG_MINIMAL		minimal match
  		//     n	REG_NEWLINE		explicit \n match
  		//     o	REG_ENCLOSED		(|&) magic inside [@|&](...)
  		//     p	REG_SHELL_PATH		explicit / match
  		//     q	REG_DELIMITED		delimited pattern
  		//     r	REG_RIGHT		implicit ...$
  		//     s	REG_SHELL_ESCAPED	\ not special
  		//     t	REG_MUSTDELIM		all delimiters must be specified
  		//     u	standard unspecified behavior -- errors not counted
  		//     v	REG_CLASS_ESCAPE	\ special inside [...]
  		//     w	REG_NOSUB		no subexpression match array
  		//     x	REG_LENIENT		let some errors slide
  		//     y	REG_LEFT		regexec() implicit ^...
  		//     z	REG_NULL		NULL subexpressions ok
  		//     $	                        expand C \c escapes in fields 2 and 3
  		//     /	                        field 2 is a regsubcomp() expression
  		//     =	                        field 3 is a regdecomp() expression
  		//
  		//   Field 1 control lines:
  		//
  		//     C		set LC_COLLATE and LC_CTYPE to locale in field 2
  		//
  		//     ?test ...	output field 5 if passed and != EXPECTED, silent otherwise
  		//     &test ...	output field 5 if current and previous passed
  		//     |test ...	output field 5 if current passed and previous failed
  		//     ; ...	output field 2 if previous failed
  		//     {test ...	skip if failed until }
  		//     }		end of skip
  		//
  		//     : comment		comment copied as output NOTE
  		//     :comment:test	:comment: ignored
  		//     N[OTE] comment	comment copied as output NOTE
  		//     T[EST] comment	comment
  		//
  		//     number		use number for nmatch (20 by default)
  		flag := field[0]
  		switch flag[0] {
  		case '?', '&', '|', ';', '{', '}':
  			// Ignore all the control operators.
  			// Just run everything.
  			flag = flag[1:]
  			if flag == "" {
  				continue Reading
  			}
  		case ':':
  			i := strings.Index(flag[1:], ":")
  			if i < 0 {
  				t.Logf("skip: %s", line)
  				continue Reading
  			}
  			flag = flag[1+i+1:]
  		case 'C', 'N', 'T', '0', '1', '2', '3', '4', '5', '6', '7', '8', '9':
  			t.Logf("skip: %s", line)
  			continue Reading
  		}
  
  		// Can check field count now that we've handled the myriad comment formats.
  		if len(field) < 4 {
  			t.Errorf("%s:%d: too few fields: %s", file, lineno, line)
  			continue Reading
  		}
  
  		// Expand C escapes (a.k.a. Go escapes).
  		if strings.Contains(flag, "$") {
  			f := `"` + field[1] + `"`
  			if field[1], err = strconv.Unquote(f); err != nil {
  				t.Errorf("%s:%d: cannot unquote %s", file, lineno, f)
  			}
  			f = `"` + field[2] + `"`
  			if field[2], err = strconv.Unquote(f); err != nil {
  				t.Errorf("%s:%d: cannot unquote %s", file, lineno, f)
  			}
  		}
  
  		//   Field 2: the regular expression pattern; SAME uses the pattern from
  		//     the previous specification.
  		//
  		if field[1] == "SAME" {
  			field[1] = lastRegexp
  		}
  		lastRegexp = field[1]
  
  		//   Field 3: the string to match.
  		text := field[2]
  
  		//   Field 4: the test outcome...
  		ok, shouldCompile, shouldMatch, pos := parseFowlerResult(field[3])
  		if !ok {
  			t.Errorf("%s:%d: cannot parse result %#q", file, lineno, field[3])
  			continue Reading
  		}
  
  		//   Field 5: optional comment appended to the report.
  
  	Testing:
  		// Run test once for each specified capital letter mode that we support.
  		for _, c := range flag {
  			pattern := field[1]
  			syn := syntax.POSIX | syntax.ClassNL
  			switch c {
  			default:
  				continue Testing
  			case 'E':
  				// extended regexp (what we support)
  			case 'L':
  				// literal
  				pattern = QuoteMeta(pattern)
  			}
  
  			for _, c := range flag {
  				switch c {
  				case 'i':
  					syn |= syntax.FoldCase
  				}
  			}
  
  			re, err := compile(pattern, syn, true)
  			if err != nil {
  				if shouldCompile {
  					t.Errorf("%s:%d: %#q did not compile", file, lineno, pattern)
  				}
  				continue Testing
  			}
  			if !shouldCompile {
  				t.Errorf("%s:%d: %#q should not compile", file, lineno, pattern)
  				continue Testing
  			}
  			match := re.MatchString(text)
  			if match != shouldMatch {
  				t.Errorf("%s:%d: %#q.Match(%#q) = %v, want %v", file, lineno, pattern, text, match, shouldMatch)
  				continue Testing
  			}
  			have := re.FindStringSubmatchIndex(text)
  			if (len(have) > 0) != match {
  				t.Errorf("%s:%d: %#q.Match(%#q) = %v, but %#q.FindSubmatchIndex(%#q) = %v", file, lineno, pattern, text, match, pattern, text, have)
  				continue Testing
  			}
  			if len(have) > len(pos) {
  				have = have[:len(pos)]
  			}
  			if !same(have, pos) {
  				t.Errorf("%s:%d: %#q.FindSubmatchIndex(%#q) = %v, want %v", file, lineno, pattern, text, have, pos)
  			}
  		}
  	}
  }
  
  func parseFowlerResult(s string) (ok, compiled, matched bool, pos []int) {
  	//   Field 4: the test outcome. This is either one of the posix error
  	//     codes (with REG_ omitted) or the match array, a list of (m,n)
  	//     entries with m and n being first and last+1 positions in the
  	//     field 3 string, or NULL if REG_NOSUB is in effect and success
  	//     is expected. BADPAT is acceptable in place of any regcomp(3)
  	//     error code. The match[] array is initialized to (-2,-2) before
  	//     each test. All array elements from 0 to nmatch-1 must be specified
  	//     in the outcome. Unspecified endpoints (offset -1) are denoted by ?.
  	//     Unset endpoints (offset -2) are denoted by X. {x}(o:n) denotes a
  	//     matched (?{...}) expression, where x is the text enclosed by {...},
  	//     o is the expression ordinal counting from 1, and n is the length of
  	//     the unmatched portion of the subject string. If x starts with a
  	//     number then that is the return value of re_execf(), otherwise 0 is
  	//     returned.
  	switch {
  	case s == "":
  		// Match with no position information.
  		ok = true
  		compiled = true
  		matched = true
  		return
  	case s == "NOMATCH":
  		// Match failure.
  		ok = true
  		compiled = true
  		matched = false
  		return
  	case 'A' <= s[0] && s[0] <= 'Z':
  		// All the other error codes are compile errors.
  		ok = true
  		compiled = false
  		return
  	}
  	compiled = true
  
  	var x []int
  	for s != "" {
  		var end byte = ')'
  		if len(x)%2 == 0 {
  			if s[0] != '(' {
  				ok = false
  				return
  			}
  			s = s[1:]
  			end = ','
  		}
  		i := 0
  		for i < len(s) && s[i] != end {
  			i++
  		}
  		if i == 0 || i == len(s) {
  			ok = false
  			return
  		}
  		var v = -1
  		var err error
  		if s[:i] != "?" {
  			v, err = strconv.Atoi(s[:i])
  			if err != nil {
  				ok = false
  				return
  			}
  		}
  		x = append(x, v)
  		s = s[i+1:]
  	}
  	if len(x)%2 != 0 {
  		ok = false
  		return
  	}
  	ok = true
  	matched = true
  	pos = x
  	return
  }
  
  var text []byte
  
  func makeText(n int) []byte {
  	if len(text) >= n {
  		return text[:n]
  	}
  	text = make([]byte, n)
  	x := ^uint32(0)
  	for i := range text {
  		x += x
  		x ^= 1
  		if int32(x) < 0 {
  			x ^= 0x88888eef
  		}
  		if x%31 == 0 {
  			text[i] = '\n'
  		} else {
  			text[i] = byte(x%(0x7E+1-0x20) + 0x20)
  		}
  	}
  	return text
  }
  
  func BenchmarkMatch(b *testing.B) {
  	isRaceBuilder := strings.HasSuffix(testenv.Builder(), "-race")
  
  	for _, data := range benchData {
  		r := MustCompile(data.re)
  		for _, size := range benchSizes {
  			if isRaceBuilder && size.n > 1<<10 {
  				continue
  			}
  			t := makeText(size.n)
  			b.Run(data.name+"/"+size.name, func(b *testing.B) {
  				b.SetBytes(int64(size.n))
  				for i := 0; i < b.N; i++ {
  					if r.Match(t) {
  						b.Fatal("match!")
  					}
  				}
  			})
  		}
  	}
  }
  
  func BenchmarkMatch_onepass_regex(b *testing.B) {
  	isRaceBuilder := strings.HasSuffix(testenv.Builder(), "-race")
  	r := MustCompile(`(?s)\A.*\z`)
  	if r.get().op == notOnePass {
  		b.Fatalf("want onepass regex, but %q is not onepass", r)
  	}
  	for _, size := range benchSizes {
  		if isRaceBuilder && size.n > 1<<10 {
  			continue
  		}
  		t := makeText(size.n)
  		bs := make([][]byte, len(t))
  		for i, s := range t {
  			bs[i] = []byte{s}
  		}
  		b.Run(size.name, func(b *testing.B) {
  			b.SetBytes(int64(size.n))
  			b.ReportAllocs()
  			for i := 0; i < b.N; i++ {
  				for _, byts := range bs {
  					if !r.Match(byts) {
  						b.Fatal("not match!")
  					}
  				}
  			}
  		})
  	}
  }
  
  var benchData = []struct{ name, re string }{
  	{"Easy0", "ABCDEFGHIJKLMNOPQRSTUVWXYZ$"},
  	{"Easy0i", "(?i)ABCDEFGHIJklmnopqrstuvwxyz$"},
  	{"Easy1", "A[AB]B[BC]C[CD]D[DE]E[EF]F[FG]G[GH]H[HI]I[IJ]J$"},
  	{"Medium", "[XYZ]ABCDEFGHIJKLMNOPQRSTUVWXYZ$"},
  	{"Hard", "[ -~]*ABCDEFGHIJKLMNOPQRSTUVWXYZ$"},
  	{"Hard1", "ABCD|CDEF|EFGH|GHIJ|IJKL|KLMN|MNOP|OPQR|QRST|STUV|UVWX|WXYZ"},
  }
  
  var benchSizes = []struct {
  	name string
  	n    int
  }{
  	{"32", 32},
  	{"1K", 1 << 10},
  	{"32K", 32 << 10},
  	{"1M", 1 << 20},
  	{"32M", 32 << 20},
  }
  
  func TestLongest(t *testing.T) {
  	re, err := Compile(`a(|b)`)
  	if err != nil {
  		t.Fatal(err)
  	}
  	if g, w := re.FindString("ab"), "a"; g != w {
  		t.Errorf("first match was %q, want %q", g, w)
  	}
  	re.Longest()
  	if g, w := re.FindString("ab"), "ab"; g != w {
  		t.Errorf("longest match was %q, want %q", g, w)
  	}
  }
  
  // TestProgramTooLongForBacktrack tests that a regex which is too long
  // for the backtracker still executes properly.
  func TestProgramTooLongForBacktrack(t *testing.T) {
  	longRegex := MustCompile(`(one|two|three|four|five|six|seven|eight|nine|ten|eleven|twelve|thirteen|fourteen|fifteen|sixteen|seventeen|eighteen|nineteen|twenty|twentyone|twentytwo|twentythree|twentyfour|twentyfive|twentysix|twentyseven|twentyeight|twentynine|thirty|thirtyone|thirtytwo|thirtythree|thirtyfour|thirtyfive|thirtysix|thirtyseven|thirtyeight|thirtynine|forty|fortyone|fortytwo|fortythree|fortyfour|fortyfive|fortysix|fortyseven|fortyeight|fortynine|fifty|fiftyone|fiftytwo|fiftythree|fiftyfour|fiftyfive|fiftysix|fiftyseven|fiftyeight|fiftynine|sixty|sixtyone|sixtytwo|sixtythree|sixtyfour|sixtyfive|sixtysix|sixtyseven|sixtyeight|sixtynine|seventy|seventyone|seventytwo|seventythree|seventyfour|seventyfive|seventysix|seventyseven|seventyeight|seventynine|eighty|eightyone|eightytwo|eightythree|eightyfour|eightyfive|eightysix|eightyseven|eightyeight|eightynine|ninety|ninetyone|ninetytwo|ninetythree|ninetyfour|ninetyfive|ninetysix|ninetyseven|ninetyeight|ninetynine|onehundred)`)
  	if !longRegex.MatchString("two") {
  		t.Errorf("longRegex.MatchString(\"two\") was false, want true")
  	}
  	if longRegex.MatchString("xxx") {
  		t.Errorf("longRegex.MatchString(\"xxx\") was true, want false")
  	}
  }
  

View as plain text