qbit/go
1
0
Fork 0
mirror of https://github.com/golang/go synced 2024-11-26 06:27:58 -07:00
Code Releases Activity

remove some incorrect capitalizations.

Browse Source 
add commentary.

R=rsc
DELTA=231  (41 added, 24 deleted, 166 changed)
OCL=25724
CL=25726
This commit is contained in:
Rob Pike 2009-03-04 21:06:38 -08:00
parent 811dd2584c
commit a0feb32567
2 changed files with 193 additions and 176 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
src/lib/regexp/all_test.go
View File

@ -85,7 +85,7 @@ var matches = []tester {
tester{ `a*(|(b))c*`, "aacc", vec{0,4, 2,2, -1,-1} }, tester{ `a*(|(b))c*`, "aacc", vec{0,4, 2,2, -1,-1} },
} }
func compileTest(t *testing.T, expr string, error *os.Error) regexp.Regexp { func compileTest(t *testing.T, expr string, error *os.Error) *regexp.Regexp {
re, err := regexp.Compile(expr); re, err := regexp.Compile(expr);
if err != error { if err != error {
t.Error("compiling `", expr, "`; unexpected error: ", err.String()); t.Error("compiling `", expr, "`; unexpected error: ", err.String());

367
src/lib/regexp/regexp.go
View File

@ -2,8 +2,25 @@
// Use of this source code is governed by a BSD-style // Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file. // license that can be found in the LICENSE file.
// Regular expression library. // Package regexp implements a simple regular expression library.
//
// The syntax of the regular expressions accepted is:
//
// regexp:
// concatenation { '|' concatenation }
// concatenation:
// { closure }
//
// closure:
// term [ '*' | '+' | '?' ]
// term:
// '^'
// '$'
// '.'
// character
// '[' [ '^' ] character-ranges ']'
// '(' regexp ')'
//
package regexp package regexp
import ( import (
@ -14,7 +31,7 @@ import (
var debug = false; var debug = false;
// Error codes returned by faliures to parse an expression.
var ErrInternal = os.NewError("internal error"); var ErrInternal = os.NewError("internal error");
var ErrUnmatchedLpar = os.NewError("unmatched '('"); var ErrUnmatchedLpar = os.NewError("unmatched '('");
var ErrUnmatchedRpar = os.NewError("unmatched ')'"); var ErrUnmatchedRpar = os.NewError("unmatched ')'");
@ -28,28 +45,30 @@ var ErrBadBackslash = os.NewError("illegal backslash escape");
// An instruction executed by the NFA // An instruction executed by the NFA
type instr interface { type instr interface {
Type() int; // the type of this instruction: _CHAR, _ANY, etc. kind() int; // the type of this instruction: _CHAR, _ANY, etc.
Next() instr; // the instruction to execute after this one next() instr; // the instruction to execute after this one
SetNext(i instr); setNext(i instr);
Index() int; index() int;
SetIndex(i int); setIndex(i int);
Print(); print();
} }
// Fields and methods common to all instructions // Fields and methods common to all instructions
type common struct { type common struct {
next instr; _next instr;
index int; _index int;
} }
func (c *common) Next() instr { return c.next } func (c *common) next() instr { return c._next }
func (c *common) SetNext(i instr) { c.next = i } func (c *common) setNext(i instr) { c._next = i }
func (c *common) Index() int { return c.index } func (c *common) index() int { return c._index }
func (c *common) SetIndex(i int) { c.index = i } func (c *common) setIndex(i int) { c._index = i }
type _RE struct { // The representation of a compiled regular expression.
// The public interface is entirely through methods.
type Regexp struct {
expr string; // the original expression expr string; // the original expression
ch chan<- *_RE; // reply channel when we're done ch chan<- *Regexp; // reply channel when we're done
error *os.Error; // compile- or run-time error; nil if OK error *os.Error; // compile- or run-time error; nil if OK
inst *array.Array; inst *array.Array;
start instr; start instr;
@ -76,32 +95,32 @@ type _Start struct {
common common
} }
func (start *_Start) Type() int { return _START } func (start *_Start) kind() int { return _START }
func (start *_Start) Print() { print("start") } func (start *_Start) print() { print("start") }
// --- END end of program // --- END end of program
type _End struct { type _End struct {
common common
} }
func (end *_End) Type() int { return _END } func (end *_End) kind() int { return _END }
func (end *_End) Print() { print("end") } func (end *_End) print() { print("end") }
// --- BOT beginning of text // --- BOT beginning of text
type _Bot struct { type _Bot struct {
common common
} }
func (bot *_Bot) Type() int { return _BOT } func (bot *_Bot) kind() int { return _BOT }
func (bot *_Bot) Print() { print("bot") } func (bot *_Bot) print() { print("bot") }
// --- EOT end of text // --- EOT end of text
type _Eot struct { type _Eot struct {
common common
} }
func (eot *_Eot) Type() int { return _EOT } func (eot *_Eot) kind() int { return _EOT }
func (eot *_Eot) Print() { print("eot") } func (eot *_Eot) print() { print("eot") }
// --- CHAR a regular character // --- CHAR a regular character
type _Char struct { type _Char struct {
@ -109,8 +128,8 @@ type _Char struct {
char int; char int;
} }
func (char *_Char) Type() int { return _CHAR } func (char *_Char) kind() int { return _CHAR }
func (char *_Char) Print() { print("char ", string(char.char)) } func (char *_Char) print() { print("char ", string(char.char)) }
func newChar(char int) *_Char { func newChar(char int) *_Char {
c := new(_Char); c := new(_Char);
@ -128,9 +147,9 @@ type _CharClass struct {
ranges *array.IntArray; ranges *array.IntArray;
} }
func (cclass *_CharClass) Type() int { return _CHARCLASS } func (cclass *_CharClass) kind() int { return _CHARCLASS }
func (cclass *_CharClass) Print() { func (cclass *_CharClass) print() {
print("charclass"); print("charclass");
if cclass.negate { if cclass.negate {
print(" (negated)"); print(" (negated)");
@ -146,13 +165,13 @@ func (cclass *_CharClass) Print() {
} }
} }
func (cclass *_CharClass) AddRange(a, b int) { func (cclass *_CharClass) addRange(a, b int) {
// range is a through b inclusive // range is a through b inclusive
cclass.ranges.Push(a); cclass.ranges.Push(a);
cclass.ranges.Push(b); cclass.ranges.Push(b);
} }
func (cclass *_CharClass) Matches(c int) bool { func (cclass *_CharClass) matches(c int) bool {
for i := 0; i < cclass.ranges.Len(); i = i+2 { for i := 0; i < cclass.ranges.Len(); i = i+2 {
min := cclass.ranges.At(i); min := cclass.ranges.At(i);
max := cclass.ranges.At(i+1); max := cclass.ranges.At(i+1);
@ -174,8 +193,8 @@ type _Any struct {
common common
} }
func (any *_Any) Type() int { return _ANY } func (any *_Any) kind() int { return _ANY }
func (any *_Any) Print() { print("any") } func (any *_Any) print() { print("any") }
// --- BRA parenthesized expression // --- BRA parenthesized expression
type _Bra struct { type _Bra struct {
@ -183,8 +202,8 @@ type _Bra struct {
n int; // subexpression number n int; // subexpression number
} }
func (bra *_Bra) Type() int { return _BRA } func (bra *_Bra) kind() int { return _BRA }
func (bra *_Bra) Print() { print("bra", bra.n); } func (bra *_Bra) print() { print("bra", bra.n); }
// --- EBRA end of parenthesized expression // --- EBRA end of parenthesized expression
type _Ebra struct { type _Ebra struct {
@ -192,8 +211,8 @@ type _Ebra struct {
n int; // subexpression number n int; // subexpression number
} }
func (ebra *_Ebra) Type() int { return _EBRA } func (ebra *_Ebra) kind() int { return _EBRA }
func (ebra *_Ebra) Print() { print("ebra ", ebra.n); } func (ebra *_Ebra) print() { print("ebra ", ebra.n); }
// --- ALT alternation // --- ALT alternation
type _Alt struct { type _Alt struct {
@ -201,32 +220,32 @@ type _Alt struct {
left instr; // other branch left instr; // other branch
} }
func (alt *_Alt) Type() int { return _ALT } func (alt *_Alt) kind() int { return _ALT }
func (alt *_Alt) Print() { print("alt(", alt.left.Index(), ")"); } func (alt *_Alt) print() { print("alt(", alt.left.index(), ")"); }
// --- NOP no operation // --- NOP no operation
type _Nop struct { type _Nop struct {
common common
} }
func (nop *_Nop) Type() int { return _NOP } func (nop *_Nop) kind() int { return _NOP }
func (nop *_Nop) Print() { print("nop") } func (nop *_Nop) print() { print("nop") }
// report error and exit compiling/executing goroutine // report error and exit compiling/executing goroutine
func (re *_RE) Error(err *os.Error) { func (re *Regexp) setError(err *os.Error) {
re.error = err; re.error = err;
re.ch <- re; re.ch <- re;
sys.Goexit(); sys.Goexit();
} }
func (re *_RE) Add(i instr) instr { func (re *Regexp) add(i instr) instr {
i.SetIndex(re.inst.Len()); i.setIndex(re.inst.Len());
re.inst.Push(i); re.inst.Push(i);
return i; return i;
} }
type parser struct { type parser struct {
re *_RE; re *Regexp;
nlpar int; // number of unclosed lpars nlpar int; // number of unclosed lpars
pos int; pos int;
ch int; ch int;
@ -249,33 +268,14 @@ func (p *parser) nextc() int {
return p.ch; return p.ch;
} }
func newParser(re *_RE) *parser { func newParser(re *Regexp) *parser {
p := new(parser); p := new(parser);
p.re = re; p.re = re;
p.nextc(); // load p.ch p.nextc(); // load p.ch
return p; return p;
} }
/* func (p *parser) regexp() (start, end instr)
Grammar:
regexp:
concatenation { '|' concatenation }
concatenation:
{ closure }
closure:
term [ '*' | '+' | '?' ]
term:
'^'
'$'
'.'
character
'[' [ '^' ] character-ranges ']'
'(' regexp ')'
*/
func (p *parser) Regexp() (start, end instr)
var iNULL instr var iNULL instr
@ -299,9 +299,9 @@ func specialcclass(c int) bool {
return false return false
} }
func (p *parser) CharClass() instr { func (p *parser) charClass() instr {
cc := newCharClass(); cc := newCharClass();
p.re.Add(cc); p.re.add(cc);
if p.c() == '^' { if p.c() == '^' {
cc.negate = true; cc.negate = true;
p.nextc(); p.nextc();
@ -311,22 +311,22 @@ func (p *parser) CharClass() instr {
switch c := p.c(); c { switch c := p.c(); c {
case ']', endOfFile: case ']', endOfFile:
if left >= 0 { if left >= 0 {
p.re.Error(ErrBadRange); p.re.setError(ErrBadRange);
} }
return cc; return cc;
case '-': // do this before backslash processing case '-': // do this before backslash processing
p.re.Error(ErrBadRange); p.re.setError(ErrBadRange);
case '\\': case '\\':
c = p.nextc(); c = p.nextc();
switch { switch {
case c == endOfFile: case c == endOfFile:
p.re.Error(ErrExtraneousBackslash); p.re.setError(ErrExtraneousBackslash);
case c == 'n': case c == 'n':
c = '\n'; c = '\n';
case specialcclass(c): case specialcclass(c):
// c is as delivered // c is as delivered
default: default:
p.re.Error(ErrBadBackslash); p.re.setError(ErrBadBackslash);
} }
fallthrough; fallthrough;
default: default:
@ -337,49 +337,49 @@ func (p *parser) CharClass() instr {
p.nextc(); p.nextc();
left = c; left = c;
} else { // single char } else { // single char
cc.AddRange(c, c); cc.addRange(c, c);
} }
case left <= c: // second of pair case left <= c: // second of pair
cc.AddRange(left, c); cc.addRange(left, c);
left = -1; left = -1;
default: default:
p.re.Error(ErrBadRange); p.re.setError(ErrBadRange);
} }
} }
} }
return iNULL return iNULL
} }
func (p *parser) Term() (start, end instr) { func (p *parser) term() (start, end instr) {
switch c := p.c(); c { switch c := p.c(); c {
case '|', endOfFile: case '|', endOfFile:
return iNULL, iNULL; return iNULL, iNULL;
case '*', '+': case '*', '+':
p.re.Error(ErrBareClosure); p.re.setError(ErrBareClosure);
case ')': case ')':
if p.nlpar == 0 { if p.nlpar == 0 {
p.re.Error(ErrUnmatchedRpar); p.re.setError(ErrUnmatchedRpar);
} }
return iNULL, iNULL; return iNULL, iNULL;
case ']': case ']':
p.re.Error(ErrUnmatchedRbkt); p.re.setError(ErrUnmatchedRbkt);
case '^': case '^':
p.nextc(); p.nextc();
start = p.re.Add(new(_Bot)); start = p.re.add(new(_Bot));
return start, start; return start, start;
case '$': case '$':
p.nextc(); p.nextc();
start = p.re.Add(new(_Eot)); start = p.re.add(new(_Eot));
return start, start; return start, start;
case '.': case '.':
p.nextc(); p.nextc();
start = p.re.Add(new(_Any)); start = p.re.add(new(_Any));
return start, start; return start, start;
case '[': case '[':
p.nextc(); p.nextc();
start = p.CharClass(); start = p.charClass();
if p.c() != ']' { if p.c() != ']' {
p.re.Error(ErrUnmatchedLbkt); p.re.setError(ErrUnmatchedLbkt);
} }
p.nextc(); p.nextc();
return start, start; return start, start;
@ -388,50 +388,52 @@ func (p *parser) Term() (start, end instr) {
p.nlpar++; p.nlpar++;
p.re.nbra++; // increment first so first subexpr is \1 p.re.nbra++; // increment first so first subexpr is \1
nbra := p.re.nbra; nbra := p.re.nbra;
start, end = p.Regexp(); start, end = p.regexp();
if p.c() != ')' { if p.c() != ')' {
p.re.Error(ErrUnmatchedLpar); p.re.setError(ErrUnmatchedLpar);
} }
p.nlpar--; p.nlpar--;
p.nextc(); p.nextc();
bra := new(_Bra); bra := new(_Bra);
p.re.Add(bra); p.re.add(bra);
ebra := new(_Ebra); ebra := new(_Ebra);
p.re.Add(ebra); p.re.add(ebra);
bra.n = nbra; bra.n = nbra;
ebra.n = nbra; ebra.n = nbra;
if start == iNULL { if start == iNULL {
if end == iNULL { p.re.Error(ErrInternal) } if end == iNULL {
p.re.setError(ErrInternal)
}
start = ebra start = ebra
} else { } else {
end.SetNext(ebra); end.setNext(ebra);
} }
bra.SetNext(start); bra.setNext(start);
return bra, ebra; return bra, ebra;
case '\\': case '\\':
c = p.nextc(); c = p.nextc();
switch { switch {
case c == endOfFile: case c == endOfFile:
p.re.Error(ErrExtraneousBackslash); p.re.setError(ErrExtraneousBackslash);
case c == 'n': case c == 'n':
c = '\n'; c = '\n';
case special(c): case special(c):
// c is as delivered // c is as delivered
default: default:
p.re.Error(ErrBadBackslash); p.re.setError(ErrBadBackslash);
} }
fallthrough; fallthrough;
default: default:
p.nextc(); p.nextc();
start = newChar(c); start = newChar(c);
p.re.Add(start); p.re.add(start);
return start, start return start, start
} }
panic("unreachable"); panic("unreachable");
} }
func (p *parser) Closure() (start, end instr) { func (p *parser) closure() (start, end instr) {
start, end = p.Term(); start, end = p.term();
if start == iNULL { if start == iNULL {
return return
} }
@ -439,27 +441,27 @@ func (p *parser) Closure() (start, end instr) {
case '*': case '*':
// (start,end)*: // (start,end)*:
alt := new(_Alt); alt := new(_Alt);
p.re.Add(alt); p.re.add(alt);
end.SetNext(alt); // after end, do alt end.setNext(alt); // after end, do alt
alt.left = start; // alternate brach: return to start alt.left = start; // alternate brach: return to start
start = alt; // alt becomes new (start, end) start = alt; // alt becomes new (start, end)
end = alt; end = alt;
case '+': case '+':
// (start,end)+: // (start,end)+:
alt := new(_Alt); alt := new(_Alt);
p.re.Add(alt); p.re.add(alt);
end.SetNext(alt); // after end, do alt end.setNext(alt); // after end, do alt
alt.left = start; // alternate brach: return to start alt.left = start; // alternate brach: return to start
end = alt; // start is unchanged; end is alt end = alt; // start is unchanged; end is alt
case '?': case '?':
// (start,end)?: // (start,end)?:
alt := new(_Alt); alt := new(_Alt);
p.re.Add(alt); p.re.add(alt);
nop := new(_Nop); nop := new(_Nop);
p.re.Add(nop); p.re.add(nop);
alt.left = start; // alternate branch is start alt.left = start; // alternate branch is start
alt.next = nop; // follow on to nop alt.setNext(nop); // follow on to nop
end.SetNext(nop); // after end, go to nop end.setNext(nop); // after end, go to nop
start = alt; // start is now alt start = alt; // start is now alt
end = nop; // end is nop pointed to by both branches end = nop; // end is nop pointed to by both branches
default: default:
@ -467,49 +469,49 @@ func (p *parser) Closure() (start, end instr) {
} }
switch p.nextc() { switch p.nextc() {
case '*', '+', '?': case '*', '+', '?':
p.re.Error(ErrBadClosure); p.re.setError(ErrBadClosure);
} }
return return
} }
func (p *parser) Concatenation() (start, end instr) { func (p *parser) concatenation() (start, end instr) {
start, end = iNULL, iNULL; start, end = iNULL, iNULL;
for { for {
nstart, nend := p.Closure(); nstart, nend := p.closure();
switch { switch {
case nstart == iNULL: // end of this concatenation case nstart == iNULL: // end of this concatenation
if start == iNULL { // this is the empty string if start == iNULL { // this is the empty string
nop := p.re.Add(new(_Nop)); nop := p.re.add(new(_Nop));
return nop, nop; return nop, nop;
} }
return; return;
case start == iNULL: // this is first element of concatenation case start == iNULL: // this is first element of concatenation
start, end = nstart, nend; start, end = nstart, nend;
default: default:
end.SetNext(nstart); end.setNext(nstart);
end = nend; end = nend;
} }
} }
panic("unreachable"); panic("unreachable");
} }
func (p *parser) Regexp() (start, end instr) { func (p *parser) regexp() (start, end instr) {
start, end = p.Concatenation(); start, end = p.concatenation();
for { for {
switch p.c() { switch p.c() {
default: default:
return; return;
case '|': case '|':
p.nextc(); p.nextc();
nstart, nend := p.Concatenation(); nstart, nend := p.concatenation();
alt := new(_Alt); alt := new(_Alt);
p.re.Add(alt); p.re.add(alt);
alt.left = start; alt.left = start;
alt.next = nstart; alt.setNext(nstart);
nop := new(_Nop); nop := new(_Nop);
p.re.Add(nop); p.re.add(nop);
end.SetNext(nop); end.setNext(nop);
nend.SetNext(nop); nend.setNext(nop);
start, end = alt, nop; start, end = alt, nop;
} }
} }
@ -517,80 +519,75 @@ func (p *parser) Regexp() (start, end instr) {
} }
func unNop(i instr) instr { func unNop(i instr) instr {
for i.Type() == _NOP { for i.kind() == _NOP {
i = i.Next() i = i.next()
} }
return i return i
} }
func (re *_RE) EliminateNops() { func (re *Regexp) eliminateNops() {
for i := 0; i < re.inst.Len(); i++ { for i := 0; i < re.inst.Len(); i++ {
inst := re.inst.At(i).(instr); inst := re.inst.At(i).(instr);
if inst.Type() == _END { if inst.kind() == _END {
continue continue
} }
inst.SetNext(unNop(inst.Next())); inst.setNext(unNop(inst.next()));
if inst.Type() == _ALT { if inst.kind() == _ALT {
alt := inst.(*_Alt); alt := inst.(*_Alt);
alt.left = unNop(alt.left); alt.left = unNop(alt.left);
} }
} }
} }
func (re *_RE) Dump() { func (re *Regexp) dump() {
for i := 0; i < re.inst.Len(); i++ { for i := 0; i < re.inst.Len(); i++ {
inst := re.inst.At(i).(instr); inst := re.inst.At(i).(instr);
print(inst.Index(), ": "); print(inst.index(), ": ");
inst.Print(); inst.print();
if inst.Type() != _END { if inst.kind() != _END {
print(" -> ", inst.Next().Index()) print(" -> ", inst.next().index())
} }
print("\n"); print("\n");
} }
} }
func (re *_RE) DoParse() { func (re *Regexp) doParse() {
p := newParser(re); p := newParser(re);
start := new(_Start); start := new(_Start);
re.Add(start); re.add(start);
s, e := p.Regexp(); s, e := p.regexp();
start.next = s; start.setNext(s);
re.start = start; re.start = start;
e.SetNext(re.Add(new(_End))); e.setNext(re.add(new(_End)));
if debug { if debug {
re.Dump(); re.dump();
println(); println();
} }
re.EliminateNops(); re.eliminateNops();
if debug { if debug {
re.Dump(); re.dump();
println(); println();
} }
} }
func compiler(str string, ch chan *_RE) { func compiler(str string, ch chan *Regexp) {
re := new(_RE); re := new(Regexp);
re.expr = str; re.expr = str;
re.inst = array.New(0); re.inst = array.New(0);
re.ch = ch; re.ch = ch;
re.DoParse(); re.doParse();
ch <- re; ch <- re;
} }
// Public interface has only execute functionality // Compile parses a regular expression and returns, if successful, a Regexp
type Regexp interface { // object that can be used to match against text.
Execute(s string) []int; func Compile(str string) (regexp *Regexp, error *os.Error) {
Match(s string) bool; // Compile in a separate goroutine and wait for the result.
MatchStrings(s string) []string; ch := make(chan *Regexp);
}
// Compile in separate goroutine; wait for result
func Compile(str string) (regexp Regexp, error *os.Error) {
ch := make(chan *_RE);
go compiler(str, ch); go compiler(str, ch);
re := <-ch; re := <-ch;
return re, re.error return re, re.error
@ -604,13 +601,13 @@ type state struct {
// Append new state to to-do list. Leftmost-longest wins so avoid // Append new state to to-do list. Leftmost-longest wins so avoid
// adding a state that's already active. // adding a state that's already active.
func addState(s []state, inst instr, match []int) []state { func addState(s []state, inst instr, match []int) []state {
index := inst.Index(); index := inst.index();
l := len(s); l := len(s);
pos := match[0]; pos := match[0];
// TODO: Once the state is a vector and we can do insert, have inputs always // 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, // go in order correctly and this "earlier" test is never necessary,
for i := 0; i < l; i++ { for i := 0; i < l; i++ {
if s[i].inst.Index() == index && // same instruction if s[i].inst.index() == index && // same instruction
s[i].match[0] < pos { // earlier match already going; lefmost wins s[i].match[0] < pos { // earlier match already going; lefmost wins
return s return s
} }
@ -628,7 +625,7 @@ func addState(s []state, inst instr, match []int) []state {
return s; return s;
} }
func (re *_RE) DoExecute(str string, pos int) []int { func (re *Regexp) doExecute(str string, pos int) []int {
var s [2][]state; // TODO: use a vector when state values (not ptrs) can be vector elements 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[0] = make([]state, 10)[0:0];
s[1] = make([]state, 10)[0:0]; s[1] = make([]state, 10)[0:0];
@ -643,7 +640,7 @@ func (re *_RE) DoExecute(str string, pos int) []int {
match[i] = -1; // no match seen; catches cases like "a(b)?c" on "ac" match[i] = -1; // no match seen; catches cases like "a(b)?c" on "ac"
} }
match[0] = pos; match[0] = pos;
s[out] = addState(s[out], re.start.Next(), match); s[out] = addState(s[out], re.start.next(), match);
} }
in, out = out, in; // old out state is new in state in, out = out, in; // old out state is new in state
s[out] = s[out][0:0]; // clear out state s[out] = s[out][0:0]; // clear out state
@ -658,35 +655,35 @@ func (re *_RE) DoExecute(str string, pos int) []int {
} }
for i := 0; i < len(s[in]); i++ { for i := 0; i < len(s[in]); i++ {
st := s[in][i]; st := s[in][i];
switch s[in][i].inst.Type() { switch s[in][i].inst.kind() {
case _BOT: case _BOT:
if pos == 0 { if pos == 0 {
s[in] = addState(s[in], st.inst.Next(), st.match) s[in] = addState(s[in], st.inst.next(), st.match)
} }
case _EOT: case _EOT:
if pos == len(str) { if pos == len(str) {
s[in] = addState(s[in], st.inst.Next(), st.match) s[in] = addState(s[in], st.inst.next(), st.match)
} }
case _CHAR: case _CHAR:
if c == st.inst.(*_Char).char { if c == st.inst.(*_Char).char {
s[out] = addState(s[out], st.inst.Next(), st.match) s[out] = addState(s[out], st.inst.next(), st.match)
} }
case _CHARCLASS: case _CHARCLASS:
if st.inst.(*_CharClass).Matches(c) { if st.inst.(*_CharClass).matches(c) {
s[out] = addState(s[out], st.inst.Next(), st.match) s[out] = addState(s[out], st.inst.next(), st.match)
} }
case _ANY: case _ANY:
if c != endOfFile { if c != endOfFile {
s[out] = addState(s[out], st.inst.Next(), st.match) s[out] = addState(s[out], st.inst.next(), st.match)
} }
case _BRA: case _BRA:
n := st.inst.(*_Bra).n; n := st.inst.(*_Bra).n;
st.match[2*n] = pos; st.match[2*n] = pos;
s[in] = addState(s[in], st.inst.Next(), st.match); s[in] = addState(s[in], st.inst.next(), st.match);
case _EBRA: case _EBRA:
n := st.inst.(*_Ebra).n; n := st.inst.(*_Ebra).n;
st.match[2*n+1] = pos; st.match[2*n+1] = pos;
s[in] = addState(s[in], st.inst.Next(), st.match); s[in] = addState(s[in], st.inst.next(), st.match);
case _ALT: case _ALT:
s[in] = addState(s[in], st.inst.(*_Alt).left, st.match); s[in] = addState(s[in], st.inst.(*_Alt).left, st.match);
// give other branch a copy of this match vector // give other branch a copy of this match vector
@ -694,7 +691,7 @@ func (re *_RE) DoExecute(str string, pos int) []int {
for i := 0; i < len(s1); i++ { for i := 0; i < len(s1); i++ {
s1[i] = st.match[i] s1[i] = st.match[i]
} }
s[in] = addState(s[in], st.inst.Next(), s1); s[in] = addState(s[in], st.inst.next(), s1);
case _END: case _END:
// choose leftmost longest // choose leftmost longest
if !found || // first if !found || // first
@ -705,7 +702,7 @@ func (re *_RE) DoExecute(str string, pos int) []int {
} }
found = true; found = true;
default: default:
st.inst.Print(); st.inst.print();
panic("unknown instruction in execute"); panic("unknown instruction in execute");
} }
} }
@ -715,30 +712,50 @@ func (re *_RE) DoExecute(str string, pos int) []int {
} }
func (re *_RE) Execute(s string) []int { // Execute matches the Regexp against the string s.
return re.DoExecute(s, 0) // The return value is an array of integers, in pairs, identifying the positions of
// substrings matched by the expression.
// s[a[0]:a[1]] is the substring matched by the entire expression.
// s[a[2*i]:a[2*i+1]] for i > 0 is the substring matched by the ith parenthesized subexpression.
// A value of -1 means the subexpression did not match any element of the string.
// An empty array means "no match".
// (Regexp is an internal type that implements the Regexp interface.)
func (re *Regexp) Execute(s string) (a []int) {
return re.doExecute(s, 0)
} }
func (re *_RE) Match(s string) bool { // Match returns whether the Regexp matches the string s.
return len(re.DoExecute(s, 0)) > 0 // The return value is a boolean: true for match, false for no match.
// (Regexp is an internal type that implements the Regexp interface.)
func (re *Regexp) Match(s string) bool {
return len(re.doExecute(s, 0)) > 0
} }
func (re *_RE) MatchStrings(s string) []string { // MatchStrings matches the Regexp against the string s.
r := re.DoExecute(s, 0); // The return value is an array of strings matched by the expression.
// a[0] is the substring matched by the entire expression.
// a[i] for i > 0 is the substring matched by the ith parenthesized subexpression.
// An empty array means ``no match''.
// (Regexp is an internal type that implements the Regexp interface.)
func (re *Regexp) MatchStrings(s string) (a []string) {
r := re.doExecute(s, 0);
if r == nil { if r == nil {
return nil return nil
} }
a := make([]string, len(r)/2); a = make([]string, len(r)/2);
for i := 0; i < len(r); i += 2 { for i := 0; i < len(r); i += 2 {
a[i/2] = s[r[i] : r[i+1]] if r[i] != -1 { // -1 means no match for this subexpression
a[i/2] = s[r[i] : r[i+1]]
}
} }
return a return
} }
// Exported function for simple boolean check. Anything more fancy // Match checks whether a textual regular expression
// needs a call to Compile. // matches a substring. More complicated queries need
// to use Compile and the full Regexp interface.
func Match(pattern string, s string) (matched bool, error *os.Error) { func Match(pattern string, s string) (matched bool, error *os.Error) {
re, err := Compile(pattern); re, err := Compile(pattern);
if err != nil { if err != nil {