mirror of
https://github.com/golang/go
synced 2024-11-22 04:24:39 -07:00
testing: eliminate testing/regexp
Rather than updating the stripped-down regexp implementation embedded in testing, delete it by passing the one function we need from the package main file created by gotest. R=rsc CC=golang-dev https://golang.org/cl/2761043
This commit is contained in:
Rob Pike
parent
1f4d54ea01
commit
4e9cc085d2
@ -147,6 +147,7 @@ importpath=$(gomake -s importpath)
|
||||
echo 'import "./_xtest_"'
|
||||
fi
|
||||
echo 'import "testing"'
|
||||
echo 'import __regexp__ "regexp"' # rename in case tested package is called regexp
|
||||
# test array
|
||||
echo
|
||||
echo 'var tests = []testing.Test{'
|
||||
@ -166,8 +167,8 @@ importpath=$(gomake -s importpath)
|
||||
# body
|
||||
echo
|
||||
echo 'func main() {'
|
||||
echo ' testing.Main(tests)'
|
||||
echo ' testing.RunBenchmarks(benchmarks)'
|
||||
echo ' testing.Main(__regexp__.MatchString, tests)'
|
||||
echo ' testing.RunBenchmarks(__regexp__.MatchString, benchmarks)'
|
||||
echo '}'
|
||||
}>_testmain.go
|
||||
|
||||
|
||||
@ -151,6 +151,7 @@ NOTEST=\
|
||||
rand\
|
||||
runtime/pprof\
|
||||
syscall\
|
||||
testing\
|
||||
testing/iotest\
|
||||
try\
|
||||
../libcgo\
|
||||
|
||||
@ -11,6 +11,7 @@ import (
|
||||
"log"
|
||||
"os"
|
||||
"reflect"
|
||||
"regexp"
|
||||
"testing"
|
||||
)
|
||||
|
||||
@ -89,9 +90,9 @@ func (a test) run(t *testing.T, name string) {
|
||||
}
|
||||
|
||||
func match(t *testing.T, err os.Error, pat string) bool {
|
||||
ok, errstr := testing.MatchString(pat, err.String())
|
||||
if errstr != "" {
|
||||
t.Fatalf("compile regexp %s: %v", pat, errstr)
|
||||
ok, err1 := regexp.MatchString(pat, err.String())
|
||||
if err1 != nil {
|
||||
t.Fatalf("compile regexp %s: %v", pat, err1)
|
||||
}
|
||||
return ok
|
||||
}
|
||||
|
||||
@ -10,6 +10,7 @@ import (
|
||||
"io"
|
||||
"os"
|
||||
"reflect"
|
||||
"regexp"
|
||||
"strings"
|
||||
"testing"
|
||||
"utf8"
|
||||
@ -101,7 +102,7 @@ func (x *Xs) Scan(state ScanState, verb int) os.Error {
|
||||
if err != nil {
|
||||
return err
|
||||
}
|
||||
if !testing.MustCompile("^" + string(verb) + "+$").MatchString(tok) {
|
||||
if !regexp.MustCompile("^" + string(verb) + "+$").MatchString(tok) {
|
||||
return os.ErrorString("syntax error for xs")
|
||||
}
|
||||
*x = Xs(tok)
|
||||
@ -385,7 +386,7 @@ func TestScanf(t *testing.T) {
|
||||
|
||||
func TestScanOverflow(t *testing.T) {
|
||||
// different machines and different types report errors with different strings.
|
||||
re := testing.MustCompile("overflow|too large|out of range|not representable")
|
||||
re := regexp.MustCompile("overflow|too large|out of range|not representable")
|
||||
for _, test := range overflowTests {
|
||||
_, err := Sscan(test.text, test.in)
|
||||
if err == nil {
|
||||
|
||||
@ -23,7 +23,7 @@ func TestTempFile(t *testing.T) {
|
||||
t.Errorf("TempFile(dir, `ioutil_test`) = %v, %v", f, err)
|
||||
}
|
||||
if f != nil {
|
||||
re := testing.MustCompile("^" + regexp.QuoteMeta(dir) + "/ioutil_test[0-9]+$")
|
||||
re := regexp.MustCompile("^" + regexp.QuoteMeta(dir) + "/ioutil_test[0-9]+$")
|
||||
if !re.MatchString(f.Name()) {
|
||||
t.Errorf("TempFile(`"+dir+"`, `ioutil_test`) created bad name %s", f.Name())
|
||||
}
|
||||
|
||||
@ -7,7 +7,6 @@ include ../../Make.inc
|
||||
TARG=testing
|
||||
GOFILES=\
|
||||
benchmark.go\
|
||||
regexp.go\
|
||||
testing.go\
|
||||
|
||||
include ../../Make.pkg
|
||||
|
||||
@ -148,18 +148,18 @@ func (b *B) run() {
|
||||
|
||||
// An internal function but exported because it is cross-package; part of the implementation
|
||||
// of gotest.
|
||||
func RunBenchmarks(benchmarks []Benchmark) {
|
||||
func RunBenchmarks(matchString func(pat, str string) (bool, os.Error), benchmarks []Benchmark) {
|
||||
// If no flag was specified, don't run benchmarks.
|
||||
if len(*matchBenchmarks) == 0 {
|
||||
return
|
||||
}
|
||||
re, err := CompileRegexp(*matchBenchmarks)
|
||||
if err != "" {
|
||||
println("invalid regexp for -benchmarks:", err)
|
||||
os.Exit(1)
|
||||
}
|
||||
for _, Benchmark := range benchmarks {
|
||||
if !re.MatchString(Benchmark.Name) {
|
||||
matched, err := matchString(*matchBenchmarks, Benchmark.Name)
|
||||
if err != nil {
|
||||
println("invalid regexp for -benchmarks:", err)
|
||||
os.Exit(1)
|
||||
}
|
||||
if !matched {
|
||||
continue
|
||||
}
|
||||
b := &B{benchmark: Benchmark}
|
||||
|
||||
@ -1,751 +0,0 @@
|
||||
// Copyright 2009 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.
|
||||
|
||||
// The testing package implements a simple regular expression library.
|
||||
// It is a reduced version of the regular expression package suitable
|
||||
// for use in tests; it avoids many dependencies.
|
||||
//
|
||||
// The syntax of the regular expressions accepted is:
|
||||
//
|
||||
// regexp:
|
||||
// concatenation { '|' concatenation }
|
||||
// concatenation:
|
||||
// { closure }
|
||||
// closure:
|
||||
// term [ '*' | '+' | '?' ]
|
||||
// term:
|
||||
// '^'
|
||||
// '$'
|
||||
// '.'
|
||||
// character
|
||||
// '[' [ '^' ] character-ranges ']'
|
||||
// '(' regexp ')'
|
||||
//
|
||||
|
||||
package testing
|
||||
|
||||
import (
|
||||
"utf8"
|
||||
)
|
||||
|
||||
var debug = false
|
||||
|
||||
// Error codes returned by failures to parse an expression.
|
||||
var (
|
||||
ErrInternal = "internal error"
|
||||
ErrUnmatchedLpar = "unmatched '('"
|
||||
ErrUnmatchedRpar = "unmatched ')'"
|
||||
ErrUnmatchedLbkt = "unmatched '['"
|
||||
ErrUnmatchedRbkt = "unmatched ']'"
|
||||
ErrBadRange = "bad range in character class"
|
||||
ErrExtraneousBackslash = "extraneous backslash"
|
||||
ErrBadClosure = "repeated closure (**, ++, etc.)"
|
||||
ErrBareClosure = "closure applies to nothing"
|
||||
ErrBadBackslash = "illegal backslash escape"
|
||||
)
|
||||
|
||||
// An instruction executed by the NFA
|
||||
type instr interface {
|
||||
kind() int // the type of this instruction: _CHAR, _ANY, etc.
|
||||
next() instr // the instruction to execute after this one
|
||||
setNext(i instr)
|
||||
index() int
|
||||
setIndex(i int)
|
||||
print()
|
||||
}
|
||||
|
||||
// Fields and methods common to all instructions
|
||||
type common struct {
|
||||
_next instr
|
||||
_index int
|
||||
}
|
||||
|
||||
func (c *common) next() instr { return c._next }
|
||||
func (c *common) setNext(i instr) { c._next = i }
|
||||
func (c *common) index() int { return c._index }
|
||||
func (c *common) setIndex(i int) { c._index = i }
|
||||
|
||||
// The representation of a compiled regular expression.
|
||||
// The public interface is entirely through methods.
|
||||
type Regexp struct {
|
||||
expr string // the original expression
|
||||
inst []instr
|
||||
start instr
|
||||
nbra int // number of brackets in expression, for subexpressions
|
||||
}
|
||||
|
||||
const (
|
||||
_START = iota // beginning of program
|
||||
_END // end of program: success
|
||||
_BOT // '^' beginning of text
|
||||
_EOT // '$' end of text
|
||||
_CHAR // 'a' regular character
|
||||
_CHARCLASS // [a-z] character class
|
||||
_ANY // '.' any character including newline
|
||||
_NOTNL // [^\n] special case: any character but newline
|
||||
_BRA // '(' parenthesized expression
|
||||
_EBRA // ')'; end of '(' parenthesized expression
|
||||
_ALT // '|' alternation
|
||||
_NOP // do nothing; makes it easy to link without patching
|
||||
)
|
||||
|
||||
// --- START start of program
|
||||
type _Start struct {
|
||||
common
|
||||
}
|
||||
|
||||
func (start *_Start) kind() int { return _START }
|
||||
func (start *_Start) print() { print("start") }
|
||||
|
||||
// --- END end of program
|
||||
type _End struct {
|
||||
common
|
||||
}
|
||||
|
||||
func (end *_End) kind() int { return _END }
|
||||
func (end *_End) print() { print("end") }
|
||||
|
||||
// --- BOT beginning of text
|
||||
type _Bot struct {
|
||||
common
|
||||
}
|
||||
|
||||
func (bot *_Bot) kind() int { return _BOT }
|
||||
func (bot *_Bot) print() { print("bot") }
|
||||
|
||||
// --- EOT end of text
|
||||
type _Eot struct {
|
||||
common
|
||||
}
|
||||
|
||||
func (eot *_Eot) kind() int { return _EOT }
|
||||
func (eot *_Eot) print() { print("eot") }
|
||||
|
||||
// --- CHAR a regular character
|
||||
type _Char struct {
|
||||
common
|
||||
char int
|
||||
}
|
||||
|
||||
func (char *_Char) kind() int { return _CHAR }
|
||||
func (char *_Char) print() { print("char ", string(char.char)) }
|
||||
|
||||
func newChar(char int) *_Char {
|
||||
c := new(_Char)
|
||||
c.char = char
|
||||
return c
|
||||
}
|
||||
|
||||
// --- CHARCLASS [a-z]
|
||||
|
||||
type _CharClass struct {
|
||||
common
|
||||
char int
|
||||
negate bool // is character class negated? ([^a-z])
|
||||
// stored pairwise: [a-z] is (a,z); x is (x,x):
|
||||
ranges []int
|
||||
}
|
||||
|
||||
func (cclass *_CharClass) kind() int { return _CHARCLASS }
|
||||
|
||||
func (cclass *_CharClass) print() {
|
||||
print("charclass")
|
||||
if cclass.negate {
|
||||
print(" (negated)")
|
||||
}
|
||||
for i := 0; i < len(cclass.ranges); i += 2 {
|
||||
l := cclass.ranges[i]
|
||||
r := cclass.ranges[i+1]
|
||||
if l == r {
|
||||
print(" [", string(l), "]")
|
||||
} else {
|
||||
print(" [", string(l), "-", string(r), "]")
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
func (cclass *_CharClass) addRange(a, b int) {
|
||||
// range is a through b inclusive
|
||||
cclass.ranges = append(cclass.ranges, a, b)
|
||||
}
|
||||
|
||||
func (cclass *_CharClass) matches(c int) bool {
|
||||
for i := 0; i < len(cclass.ranges); i = i + 2 {
|
||||
min := cclass.ranges[i]
|
||||
max := cclass.ranges[i+1]
|
||||
if min <= c && c <= max {
|
||||
return !cclass.negate
|
||||
}
|
||||
}
|
||||
return cclass.negate
|
||||
}
|
||||
|
||||
func newCharClass() *_CharClass {
|
||||
c := new(_CharClass)
|
||||
c.ranges = make([]int, 0, 20)
|
||||
return c
|
||||
}
|
||||
|
||||
// --- ANY any character
|
||||
type _Any struct {
|
||||
common
|
||||
}
|
||||
|
||||
func (any *_Any) kind() int { return _ANY }
|
||||
func (any *_Any) print() { print("any") }
|
||||
|
||||
// --- NOTNL any character but newline
|
||||
type _NotNl struct {
|
||||
common
|
||||
}
|
||||
|
||||
func (notnl *_NotNl) kind() int { return _NOTNL }
|
||||
func (notnl *_NotNl) print() { print("notnl") }
|
||||
|
||||
// --- BRA parenthesized expression
|
||||
type _Bra struct {
|
||||
common
|
||||
n int // subexpression number
|
||||
}
|
||||
|
||||
func (bra *_Bra) kind() int { return _BRA }
|
||||
func (bra *_Bra) print() { print("bra", bra.n) }
|
||||
|
||||
// --- EBRA end of parenthesized expression
|
||||
type _Ebra struct {
|
||||
common
|
||||
n int // subexpression number
|
||||
}
|
||||
|
||||
func (ebra *_Ebra) kind() int { return _EBRA }
|
||||
func (ebra *_Ebra) print() { print("ebra ", ebra.n) }
|
||||
|
||||
// --- ALT alternation
|
||||
type _Alt struct {
|
||||
common
|
||||
left instr // other branch
|
||||
}
|
||||
|
||||
func (alt *_Alt) kind() int { return _ALT }
|
||||
func (alt *_Alt) print() { print("alt(", alt.left.index(), ")") }
|
||||
|
||||
// --- NOP no operation
|
||||
type _Nop struct {
|
||||
common
|
||||
}
|
||||
|
||||
func (nop *_Nop) kind() int { return _NOP }
|
||||
func (nop *_Nop) print() { print("nop") }
|
||||
|
||||
func (re *Regexp) add(i instr) instr {
|
||||
i.setIndex(len(re.inst))
|
||||
re.inst = append(re.inst, i)
|
||||
return i
|
||||
}
|
||||
|
||||
type parser struct {
|
||||
re *Regexp
|
||||
nlpar int // number of unclosed lpars
|
||||
pos int
|
||||
ch int
|
||||
}
|
||||
|
||||
func (p *parser) error(err string) {
|
||||
panic(err)
|
||||
}
|
||||
|
||||
const endOfFile = -1
|
||||
|
||||
func (p *parser) c() int { return p.ch }
|
||||
|
||||
func (p *parser) nextc() int {
|
||||
if p.pos >= len(p.re.expr) {
|
||||
p.ch = endOfFile
|
||||
} else {
|
||||
c, w := utf8.DecodeRuneInString(p.re.expr[p.pos:])
|
||||
p.ch = c
|
||||
p.pos += w
|
||||
}
|
||||
return p.ch
|
||||
}
|
||||
|
||||
func newParser(re *Regexp) *parser {
|
||||
p := new(parser)
|
||||
p.re = re
|
||||
p.nextc() // load p.ch
|
||||
return p
|
||||
}
|
||||
|
||||
func special(c int) bool {
|
||||
s := `\.+*?()|[]^$`
|
||||
for i := 0; i < len(s); i++ {
|
||||
if c == int(s[i]) {
|
||||
return true
|
||||
}
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
func specialcclass(c int) bool {
|
||||
s := `\-[]`
|
||||
for i := 0; i < len(s); i++ {
|
||||
if c == int(s[i]) {
|
||||
return true
|
||||
}
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
func (p *parser) charClass() instr {
|
||||
cc := newCharClass()
|
||||
if p.c() == '^' {
|
||||
cc.negate = true
|
||||
p.nextc()
|
||||
}
|
||||
left := -1
|
||||
for {
|
||||
switch c := p.c(); c {
|
||||
case ']', endOfFile:
|
||||
if left >= 0 {
|
||||
p.error(ErrBadRange)
|
||||
}
|
||||
// Is it [^\n]?
|
||||
if cc.negate && len(cc.ranges) == 2 &&
|
||||
cc.ranges[0] == '\n' && cc.ranges[1] == '\n' {
|
||||
nl := new(_NotNl)
|
||||
p.re.add(nl)
|
||||
return nl
|
||||
}
|
||||
p.re.add(cc)
|
||||
return cc
|
||||
case '-': // do this before backslash processing
|
||||
p.error(ErrBadRange)
|
||||
case '\\':
|
||||
c = p.nextc()
|
||||
switch {
|
||||
case c == endOfFile:
|
||||
p.error(ErrExtraneousBackslash)
|
||||
case c == 'n':
|
||||
c = '\n'
|
||||
case specialcclass(c):
|
||||
// c is as delivered
|
||||
default:
|
||||
p.error(ErrBadBackslash)
|
||||
}
|
||||
fallthrough
|
||||
default:
|
||||
p.nextc()
|
||||
switch {
|
||||
case left < 0: // first of pair
|
||||
if p.c() == '-' { // range
|
||||
p.nextc()
|
||||
left = c
|
||||
} else { // single char
|
||||
cc.addRange(c, c)
|
||||
}
|
||||
case left <= c: // second of pair
|
||||
cc.addRange(left, c)
|
||||
left = -1
|
||||
default:
|
||||
p.error(ErrBadRange)
|
||||
}
|
||||
}
|
||||
}
|
||||
return nil
|
||||
}
|
||||
|
||||
func (p *parser) term() (start, end instr) {
|
||||
switch c := p.c(); c {
|
||||
case '|', endOfFile:
|
||||
return nil, nil
|
||||
case '*', '+':
|
||||
p.error(ErrBareClosure)
|
||||
return
|
||||
case ')':
|
||||
if p.nlpar == 0 {
|
||||
p.error(ErrUnmatchedRpar)
|
||||
}
|
||||
return nil, nil
|
||||
case ']':
|
||||
p.error(ErrUnmatchedRbkt)
|
||||
case '^':
|
||||
p.nextc()
|
||||
start = p.re.add(new(_Bot))
|
||||
return start, start
|
||||
case '$':
|
||||
p.nextc()
|
||||
start = p.re.add(new(_Eot))
|
||||
return start, start
|
||||
case '.':
|
||||
p.nextc()
|
||||
start = p.re.add(new(_Any))
|
||||
return start, start
|
||||
case '[':
|
||||
p.nextc()
|
||||
start = p.charClass()
|
||||
if p.c() != ']' {
|
||||
p.error(ErrUnmatchedLbkt)
|
||||
}
|
||||
p.nextc()
|
||||
return start, start
|
||||
case '(':
|
||||
p.nextc()
|
||||
p.nlpar++
|
||||
p.re.nbra++ // increment first so first subexpr is \1
|
||||
nbra := p.re.nbra
|
||||
start, end = p.regexp()
|
||||
if p.c() != ')' {
|
||||
p.error(ErrUnmatchedLpar)
|
||||
}
|
||||
p.nlpar--
|
||||
p.nextc()
|
||||
bra := new(_Bra)
|
||||
p.re.add(bra)
|
||||
ebra := new(_Ebra)
|
||||
p.re.add(ebra)
|
||||
bra.n = nbra
|
||||
ebra.n = nbra
|
||||
if start == nil {
|
||||
if end == nil {
|
||||
p.error(ErrInternal)
|
||||
}
|
||||
start = ebra
|
||||
} else {
|
||||
end.setNext(ebra)
|
||||
}
|
||||
bra.setNext(start)
|
||||
return bra, ebra
|
||||
case '\\':
|
||||
c = p.nextc()
|
||||
switch {
|
||||
case c == endOfFile:
|
||||
p.error(ErrExtraneousBackslash)
|
||||
return
|
||||
case c == 'n':
|
||||
c = '\n'
|
||||
case special(c):
|
||||
// c is as delivered
|
||||
default:
|
||||
p.error(ErrBadBackslash)
|
||||
}
|
||||
fallthrough
|
||||
default:
|
||||
p.nextc()
|
||||
start = newChar(c)
|
||||
p.re.add(start)
|
||||
return start, start
|
||||
}
|
||||
panic("unreachable")
|
||||
}
|
||||
|
||||
func (p *parser) closure() (start, end instr) {
|
||||
start, end = p.term()
|
||||
if start == nil {
|
||||
return
|
||||
}
|
||||
switch p.c() {
|
||||
case '*':
|
||||
// (start,end)*:
|
||||
alt := new(_Alt)
|
||||
p.re.add(alt)
|
||||
end.setNext(alt) // after end, do alt
|
||||
alt.left = start // alternate brach: return to start
|
||||
start = alt // alt becomes new (start, end)
|
||||
end = alt
|
||||
case '+':
|
||||
// (start,end)+:
|
||||
alt := new(_Alt)
|
||||
p.re.add(alt)
|
||||
end.setNext(alt) // after end, do alt
|
||||
alt.left = start // alternate brach: return to start
|
||||
end = alt // start is unchanged; end is alt
|
||||
case '?':
|
||||
// (start,end)?:
|
||||
alt := new(_Alt)
|
||||
p.re.add(alt)
|
||||
nop := new(_Nop)
|
||||
p.re.add(nop)
|
||||
alt.left = start // alternate branch is start
|
||||
alt.setNext(nop) // follow on to nop
|
||||
end.setNext(nop) // after end, go to nop
|
||||
start = alt // start is now alt
|
||||
end = nop // end is nop pointed to by both branches
|
||||
default:
|
||||
return
|
||||
}
|
||||
switch p.nextc() {
|
||||
case '*', '+', '?':
|
||||
p.error(ErrBadClosure)
|
||||
}
|
||||
return
|
||||
}
|
||||
|
||||
func (p *parser) concatenation() (start, end instr) {
|
||||
for {
|
||||
nstart, nend := p.closure()
|
||||
switch {
|
||||
case nstart == nil: // end of this concatenation
|
||||
if start == nil { // this is the empty string
|
||||
nop := p.re.add(new(_Nop))
|
||||
return nop, nop
|
||||
}
|
||||
return
|
||||
case start == nil: // this is first element of concatenation
|
||||
start, end = nstart, nend
|
||||
default:
|
||||
end.setNext(nstart)
|
||||
end = nend
|
||||
}
|
||||
}
|
||||
panic("unreachable")
|
||||
}
|
||||
|
||||
func (p *parser) regexp() (start, end instr) {
|
||||
start, end = p.concatenation()
|
||||
for {
|
||||
switch p.c() {
|
||||
default:
|
||||
return
|
||||
case '|':
|
||||
p.nextc()
|
||||
nstart, nend := p.concatenation()
|
||||
alt := new(_Alt)
|
||||
p.re.add(alt)
|
||||
alt.left = start
|
||||
alt.setNext(nstart)
|
||||
nop := new(_Nop)
|
||||
p.re.add(nop)
|
||||
end.setNext(nop)
|
||||
nend.setNext(nop)
|
||||
start, end = alt, nop
|
||||
}
|
||||
}
|
||||
panic("unreachable")
|
||||
}
|
||||
|
||||
func unNop(i instr) instr {
|
||||
for i.kind() == _NOP {
|
||||
i = i.next()
|
||||
}
|
||||
return i
|
||||
}
|
||||
|
||||
func (re *Regexp) eliminateNops() {
|
||||
for i := 0; i < len(re.inst); i++ {
|
||||
inst := re.inst[i]
|
||||
if inst.kind() == _END {
|
||||
continue
|
||||
}
|
||||
inst.setNext(unNop(inst.next()))
|
||||
if inst.kind() == _ALT {
|
||||
alt := inst.(*_Alt)
|
||||
alt.left = unNop(alt.left)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
func (re *Regexp) doParse() {
|
||||
p := newParser(re)
|
||||
start := new(_Start)
|
||||
re.add(start)
|
||||
s, e := p.regexp()
|
||||
start.setNext(s)
|
||||
re.start = start
|
||||
e.setNext(re.add(new(_End)))
|
||||
re.eliminateNops()
|
||||
}
|
||||
|
||||
// CompileRegexp parses a regular expression and returns, if successful, a Regexp
|
||||
// object that can be used to match against text.
|
||||
func CompileRegexp(str string) (regexp *Regexp, error string) {
|
||||
regexp = new(Regexp)
|
||||
// doParse will panic if there is a parse error.
|
||||
defer func() {
|
||||
if e := recover(); e != nil {
|
||||
regexp = nil
|
||||
error = e.(string) // Will re-panic if error was not a string, e.g. nil-pointer exception
|
||||
}
|
||||
}()
|
||||
regexp.expr = str
|
||||
regexp.inst = make([]instr, 0, 20)
|
||||
regexp.doParse()
|
||||
return
|
||||
}
|
||||
|
||||
// MustCompileRegexp is like CompileRegexp but panics if the expression cannot be parsed.
|
||||
// It simplifies safe initialization of global variables holding compiled regular
|
||||
// expressions.
|
||||
func MustCompile(str string) *Regexp {
|
||||
regexp, error := CompileRegexp(str)
|
||||
if error != "" {
|
||||
panic(`regexp: compiling "` + str + `": ` + error)
|
||||
}
|
||||
return regexp
|
||||
}
|
||||
|
||||
type state struct {
|
||||
inst instr // next instruction to execute
|
||||
match []int // pairs of bracketing submatches. 0th is start,end
|
||||
}
|
||||
|
||||
// Append new state to to-do list. Leftmost-longest wins so avoid
|
||||
// adding a state that's already active.
|
||||
func addState(s []state, inst instr, match []int) []state {
|
||||
index := inst.index()
|
||||
l := len(s)
|
||||
pos := match[0]
|
||||
// TODO: Once the state is a vector and we can do insert, have inputs always
|
||||
// go in order correctly and this "earlier" test is never necessary,
|
||||
for i := 0; i < l; i++ {
|
||||
if s[i].inst.index() == index && // same instruction
|
||||
s[i].match[0] < pos { // earlier match already going; lefmost wins
|
||||
return s
|
||||
}
|
||||
}
|
||||
if l == cap(s) {
|
||||
s1 := make([]state, 2*l)[0:l]
|
||||
for i := 0; i < l; i++ {
|
||||
s1[i] = s[i]
|
||||
}
|
||||
s = s1
|
||||
}
|
||||
s = s[0 : l+1]
|
||||
s[l].inst = inst
|
||||
s[l].match = match
|
||||
return s
|
||||
}
|
||||
|
||||
// Accepts either string or bytes - the logic is identical either way.
|
||||
// If bytes == nil, scan str.
|
||||
func (re *Regexp) doExecute(str string, bytes []byte, pos int) []int {
|
||||
var s [2][]state // TODO: use a vector when state values (not ptrs) can be vector elements
|
||||
s[0] = make([]state, 10)[0:0]
|
||||
s[1] = make([]state, 10)[0:0]
|
||||
in, out := 0, 1
|
||||
var final state
|
||||
found := false
|
||||
end := len(str)
|
||||
if bytes != nil {
|
||||
end = len(bytes)
|
||||
}
|
||||
for pos <= end {
|
||||
if !found {
|
||||
// prime the pump if we haven't seen a match yet
|
||||
match := make([]int, 2*(re.nbra+1))
|
||||
for i := 0; i < len(match); i++ {
|
||||
match[i] = -1 // no match seen; catches cases like "a(b)?c" on "ac"
|
||||
}
|
||||
match[0] = pos
|
||||
s[out] = addState(s[out], re.start.next(), match)
|
||||
}
|
||||
in, out = out, in // old out state is new in state
|
||||
s[out] = s[out][0:0] // clear out state
|
||||
if len(s[in]) == 0 {
|
||||
// machine has completed
|
||||
break
|
||||
}
|
||||
charwidth := 1
|
||||
c := endOfFile
|
||||
if pos < end {
|
||||
if bytes == nil {
|
||||
c, charwidth = utf8.DecodeRuneInString(str[pos:end])
|
||||
} else {
|
||||
c, charwidth = utf8.DecodeRune(bytes[pos:end])
|
||||
}
|
||||
}
|
||||
for i := 0; i < len(s[in]); i++ {
|
||||
st := s[in][i]
|
||||
switch s[in][i].inst.kind() {
|
||||
case _BOT:
|
||||
if pos == 0 {
|
||||
s[in] = addState(s[in], st.inst.next(), st.match)
|
||||
}
|
||||
case _EOT:
|
||||
if pos == end {
|
||||
s[in] = addState(s[in], st.inst.next(), st.match)
|
||||
}
|
||||
case _CHAR:
|
||||
if c == st.inst.(*_Char).char {
|
||||
s[out] = addState(s[out], st.inst.next(), st.match)
|
||||
}
|
||||
case _CHARCLASS:
|
||||
if st.inst.(*_CharClass).matches(c) {
|
||||
s[out] = addState(s[out], st.inst.next(), st.match)
|
||||
}
|
||||
case _ANY:
|
||||
if c != endOfFile {
|
||||
s[out] = addState(s[out], st.inst.next(), st.match)
|
||||
}
|
||||
case _NOTNL:
|
||||
if c != endOfFile && c != '\n' {
|
||||
s[out] = addState(s[out], st.inst.next(), st.match)
|
||||
}
|
||||
case _BRA:
|
||||
n := st.inst.(*_Bra).n
|
||||
st.match[2*n] = pos
|
||||
s[in] = addState(s[in], st.inst.next(), st.match)
|
||||
case _EBRA:
|
||||
n := st.inst.(*_Ebra).n
|
||||
st.match[2*n+1] = pos
|
||||
s[in] = addState(s[in], st.inst.next(), st.match)
|
||||
case _ALT:
|
||||
s[in] = addState(s[in], st.inst.(*_Alt).left, st.match)
|
||||
// give other branch a copy of this match vector
|
||||
s1 := make([]int, 2*(re.nbra+1))
|
||||
for i := 0; i < len(s1); i++ {
|
||||
s1[i] = st.match[i]
|
||||
}
|
||||
s[in] = addState(s[in], st.inst.next(), s1)
|
||||
case _END:
|
||||
// choose leftmost longest
|
||||
if !found || // first
|
||||
st.match[0] < final.match[0] || // leftmost
|
||||
(st.match[0] == final.match[0] && pos > final.match[1]) { // longest
|
||||
final = st
|
||||
final.match[1] = pos
|
||||
}
|
||||
found = true
|
||||
default:
|
||||
st.inst.print()
|
||||
panic("unknown instruction in execute")
|
||||
}
|
||||
}
|
||||
pos += charwidth
|
||||
}
|
||||
return final.match
|
||||
}
|
||||
|
||||
|
||||
// MatchString returns whether the Regexp matches the string s.
|
||||
// The return value is a boolean: true for match, false for no match.
|
||||
func (re *Regexp) MatchString(s string) bool { return len(re.doExecute(s, nil, 0)) > 0 }
|
||||
|
||||
|
||||
// Match returns whether the Regexp matches the byte slice b.
|
||||
// The return value is a boolean: true for match, false for no match.
|
||||
func (re *Regexp) Match(b []byte) bool { return len(re.doExecute("", b, 0)) > 0 }
|
||||
|
||||
|
||||
// MatchString checks whether a textual regular expression
|
||||
// matches a string. More complicated queries need
|
||||
// to use Compile and the full Regexp interface.
|
||||
func MatchString(pattern string, s string) (matched bool, error string) {
|
||||
re, err := CompileRegexp(pattern)
|
||||
if err != "" {
|
||||
return false, err
|
||||
}
|
||||
return re.MatchString(s), ""
|
||||
}
|
||||
|
||||
|
||||
// Match checks whether a textual regular expression
|
||||
// matches a byte slice. More complicated queries need
|
||||
// to use Compile and the full Regexp interface.
|
||||
func Match(pattern string, b []byte) (matched bool, error string) {
|
||||
re, err := CompileRegexp(pattern)
|
||||
if err != "" {
|
||||
return false, err
|
||||
}
|
||||
return re.Match(b), ""
|
||||
}
|
||||
@ -1,143 +0,0 @@
|
||||
// Copyright 2009 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 testing
|
||||
|
||||
var good_re = []string{
|
||||
``,
|
||||
`.`,
|
||||
`^.$`,
|
||||
`a`,
|
||||
`a*`,
|
||||
`a+`,
|
||||
`a?`,
|
||||
`a|b`,
|
||||
`a*|b*`,
|
||||
`(a*|b)(c*|d)`,
|
||||
`[a-z]`,
|
||||
`[a-abc-c\-\]\[]`,
|
||||
`[a-z]+`,
|
||||
`[]`,
|
||||
`[abc]`,
|
||||
`[^1234]`,
|
||||
`[^\n]`,
|
||||
}
|
||||
|
||||
// TODO: nice to do this with a map
|
||||
type stringError struct {
|
||||
re string
|
||||
err string
|
||||
}
|
||||
|
||||
var bad_re = []stringError{
|
||||
{`*`, ErrBareClosure},
|
||||
{`(abc`, ErrUnmatchedLpar},
|
||||
{`abc)`, ErrUnmatchedRpar},
|
||||
{`x[a-z`, ErrUnmatchedLbkt},
|
||||
{`abc]`, ErrUnmatchedRbkt},
|
||||
{`[z-a]`, ErrBadRange},
|
||||
{`abc\`, ErrExtraneousBackslash},
|
||||
{`a**`, ErrBadClosure},
|
||||
{`a*+`, ErrBadClosure},
|
||||
{`a??`, ErrBadClosure},
|
||||
{`*`, ErrBareClosure},
|
||||
{`\x`, ErrBadBackslash},
|
||||
}
|
||||
|
||||
type vec []int
|
||||
|
||||
type tester struct {
|
||||
re string
|
||||
text string
|
||||
match vec
|
||||
}
|
||||
|
||||
var matches = []tester{
|
||||
{``, "", vec{0, 0}},
|
||||
{`a`, "a", vec{0, 1}},
|
||||
{`x`, "y", vec{}},
|
||||
{`b`, "abc", vec{1, 2}},
|
||||
{`.`, "a", vec{0, 1}},
|
||||
{`.*`, "abcdef", vec{0, 6}},
|
||||
{`^abcd$`, "abcd", vec{0, 4}},
|
||||
{`^bcd'`, "abcdef", vec{}},
|
||||
{`^abcd$`, "abcde", vec{}},
|
||||
{`a+`, "baaab", vec{1, 4}},
|
||||
{`a*`, "baaab", vec{0, 0}},
|
||||
{`[a-z]+`, "abcd", vec{0, 4}},
|
||||
{`[^a-z]+`, "ab1234cd", vec{2, 6}},
|
||||
{`[a\-\]z]+`, "az]-bcz", vec{0, 4}},
|
||||
{`[^\n]+`, "abcd\n", vec{0, 4}},
|
||||
{`[日本語]+`, "日本語日本語", vec{0, 18}},
|
||||
{`()`, "", vec{0, 0, 0, 0}},
|
||||
{`(a)`, "a", vec{0, 1, 0, 1}},
|
||||
{`(.)(.)`, "日a", vec{0, 4, 0, 3, 3, 4}},
|
||||
{`(.*)`, "", vec{0, 0, 0, 0}},
|
||||
{`(.*)`, "abcd", vec{0, 4, 0, 4}},
|
||||
{`(..)(..)`, "abcd", vec{0, 4, 0, 2, 2, 4}},
|
||||
{`(([^xyz]*)(d))`, "abcd", vec{0, 4, 0, 4, 0, 3, 3, 4}},
|
||||
{`((a|b|c)*(d))`, "abcd", vec{0, 4, 0, 4, 2, 3, 3, 4}},
|
||||
{`(((a|b|c)*)(d))`, "abcd", vec{0, 4, 0, 4, 0, 3, 2, 3, 3, 4}},
|
||||
{`a*(|(b))c*`, "aacc", vec{0, 4, 2, 2, -1, -1}},
|
||||
}
|
||||
|
||||
func compileTest(t *T, expr string, error string) *Regexp {
|
||||
re, err := CompileRegexp(expr)
|
||||
if err != error {
|
||||
t.Error("compiling `", expr, "`; unexpected error: ", err)
|
||||
}
|
||||
return re
|
||||
}
|
||||
|
||||
func TestGoodCompile(t *T) {
|
||||
for i := 0; i < len(good_re); i++ {
|
||||
compileTest(t, good_re[i], "")
|
||||
}
|
||||
}
|
||||
|
||||
func TestBadCompile(t *T) {
|
||||
for i := 0; i < len(bad_re); i++ {
|
||||
compileTest(t, bad_re[i].re, bad_re[i].err)
|
||||
}
|
||||
}
|
||||
|
||||
func matchTest(t *T, expr string, str string, match []int) {
|
||||
re := compileTest(t, expr, "")
|
||||
if re == nil {
|
||||
return
|
||||
}
|
||||
m := re.MatchString(str)
|
||||
if m != (len(match) > 0) {
|
||||
t.Error("MatchString failure on `", expr, "` matching `", str, "`:", m, "should be", len(match) > 0)
|
||||
}
|
||||
// now try bytes
|
||||
m = re.Match([]byte(str))
|
||||
if m != (len(match) > 0) {
|
||||
t.Error("Match failure on `", expr, "` matching `", str, "`:", m, "should be", len(match) > 0)
|
||||
}
|
||||
}
|
||||
|
||||
func TestMatch(t *T) {
|
||||
for i := 0; i < len(matches); i++ {
|
||||
test := &matches[i]
|
||||
matchTest(t, test.re, test.text, test.match)
|
||||
}
|
||||
}
|
||||
|
||||
func matchFunctionTest(t *T, expr string, str string, match []int) {
|
||||
m, err := MatchString(expr, str)
|
||||
if err == "" {
|
||||
return
|
||||
}
|
||||
if m != (len(match) > 0) {
|
||||
t.Error("function Match failure on `", expr, "` matching `", str, "`:", m, "should be", len(match) > 0)
|
||||
}
|
||||
}
|
||||
|
||||
func TestMatchFunction(t *T) {
|
||||
for i := 0; i < len(matches); i++ {
|
||||
test := &matches[i]
|
||||
matchFunctionTest(t, test.re, test.text, test.match)
|
||||
}
|
||||
}
|
||||
@ -135,19 +135,19 @@ func tRunner(t *T, test *Test) {
|
||||
|
||||
// An internal function but exported because it is cross-package; part of the implementation
|
||||
// of gotest.
|
||||
func Main(tests []Test) {
|
||||
func Main(matchString func(pat, str string) (bool, os.Error), tests []Test) {
|
||||
flag.Parse()
|
||||
ok := true
|
||||
if len(tests) == 0 {
|
||||
println("testing: warning: no tests to run")
|
||||
}
|
||||
re, err := CompileRegexp(*match)
|
||||
if err != "" {
|
||||
println("invalid regexp for -match:", err)
|
||||
os.Exit(1)
|
||||
}
|
||||
for i := 0; i < len(tests); i++ {
|
||||
if !re.MatchString(tests[i].Name) {
|
||||
matched, err := matchString(*match, tests[i].Name)
|
||||
if err != nil {
|
||||
println("invalid regexp for -match:", err)
|
||||
os.Exit(1)
|
||||
}
|
||||
if !matched {
|
||||
continue
|
||||
}
|
||||
if *chatty {
|
||||
|
||||
Loading…
Reference in New Issue
Block a user