special case: recognize '[^\n]' and make it as fast as '.'

R=rsc
DELTA=25  (23 added, 1 deleted, 1 changed)
OCL=32793
CL=32799

diff --git a/src/pkg/regexp/all_test.go b/src/pkg/regexp/all_test.go
index 23c22003eef27c096707cd83788e3d8fa7ad336d..0d16b24e3a0b7018dfb9cdfabd468ff037ad3018 100644 (file)
--- a/src/pkg/regexp/all_test.go
+++ b/src/pkg/regexp/all_test.go
@@ -27,6 +27,7 @@ var good_re = []string{
        `[]`,
        `[abc]`,
        `[^1234]`,
+       `[^\n]`,
 }
 
 // TODO: nice to do this with a map
@@ -72,6 +73,7 @@ var matches = []tester {
        tester{ `[a-z]+`,       "abcd", vec{0,4} },
        tester{ `[^a-z]+`,      "ab1234cd",     vec{2,6} },
        tester{ `[a\-\]z]+`,    "az]-bcz",      vec{0,4} },
+       tester{ `[^\n]+`,       "abcd\n",       vec{0,4} },
        tester{ `[日本語]+`, "日本語日本語",   vec{0,18} },
        tester{ `()`,   "",     vec{0,0, 0,0} },
        tester{ `(a)`,  "a",    vec{0,1, 0,1} },

diff --git a/src/pkg/regexp/regexp.go b/src/pkg/regexp/regexp.go
index 1ab9246f686ebb6263c0ce9e8fdbcfef833f97a4..745a3ae724f2b41e6c2e722dd261c72045001cbc 100644 (file)
--- a/src/pkg/regexp/regexp.go
+++ b/src/pkg/regexp/regexp.go
@@ -87,7 +87,8 @@ const (
        _EOT;           // '$' end of text
        _CHAR;  // 'a' regular character
        _CHARCLASS;     // [a-z] character class
-       _ANY;           // '.' any character
+       _ANY;           // '.' any character including newline
+       _NOTNL;         // [^\n] special case: any character but newline
        _BRA;           // '(' parenthesized expression
        _EBRA;  // ')'; end of '(' parenthesized expression
        _ALT;           // '|' alternation
@@ -200,6 +201,14 @@ type _Any struct {
 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;
@@ -305,7 +314,6 @@ func specialcclass(c int) bool {
 
 func (p *parser) charClass() instr {
        cc := newCharClass();
-       p.re.add(cc);
        if p.c() == '^' {
                cc.negate = true;
                p.nextc();
@@ -317,6 +325,14 @@ func (p *parser) charClass() instr {
                        if left >= 0 {
                                p.re.setError(ErrBadRange);
                        }
+                       // Is it [^\n]?
+                       if cc.negate && cc.ranges.Len() == 2 &&
+                               cc.ranges.At(0) == '\n' && cc.ranges.At(1) == '\n' {
+                               nl := new(_NotNl);
+                               p.re.add(nl);
+                               return nl;
+                       }
+                       p.re.add(cc);
                        return cc;
                case '-':       // do this before backslash processing
                        p.re.setError(ErrBadRange);
@@ -680,6 +696,10 @@ func (re *Regexp) doExecute(str string, pos int) []int {
                                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;