qbit/go
1
0
Fork 0
mirror of https://github.com/golang/go synced 2024-11-21 20:54:45 -07:00
Code Releases Activity

The prefix optimization applies only to the first iteration.

Browse Source 
Fixes #596.

R=rsc
CC=golang-dev
https://golang.org/cl/206101
This commit is contained in:
Rob Pike 2010-02-17 08:49:00 +11:00
parent ca075494a6
commit 7db472fd34
2 changed files with 17 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
src/pkg/regexp/all_test.go
View File

@ -30,7 +30,6 @@ var good_re = []string{
`[^\n]`,
}
// TODO: nice to do this with a map
type stringError struct {
re string
err os.Error
@ -97,6 +96,10 @@ var matches = []tester{
tester{`[.]`, ".", vec{0, 1}},
tester{`/$`, "/abc/", vec{4, 5}},
tester{`/$`, "/abc", vec{}},
// fixed bugs
tester{`ab$`, "cab", vec{1, 3}},
tester{`axxb$`, "axxcb", vec{}},
}
func compileTest(t *testing.T, expr string, error os.Error) *Regexp {

18
src/pkg/regexp/regexp.go
View File

@ -75,8 +75,9 @@ type Regexp struct {
prefix string // initial plain text string
prefixBytes []byte // initial plain text bytes
inst *vector.Vector
start instr
nbra int // number of brackets in expression, for subexpressions
start instr // first instruction of machine
prefixStart instr // where to start if there is a prefix
nbra int // number of brackets in expression, for subexpressions
}
const (
@ -650,8 +651,8 @@ Loop:
b = bytes.Add(b, utf[0:n])
i = inst.next().index()
}
// point start instruction to first non-CHAR
re.inst.At(0).(instr).setNext(re.inst.At(i).(instr))
// point prefixStart instruction to first non-CHAR after prefix
re.prefixStart = re.inst.At(i).(instr)
re.prefixBytes = b
re.prefix = string(b)
}
@ -807,6 +808,7 @@ func (re *Regexp) doExecute(str string, bytestr []byte, pos int) []int {
end = len(bytestr)
}
// fast check for initial plain substring
prefixed := false // has this iteration begun by skipping a prefix?
if re.prefix != "" {
var advance int
if bytestr == nil {
@ -818,6 +820,7 @@ func (re *Regexp) doExecute(str string, bytestr []byte, pos int) []int {
return []int{}
}
pos += advance + len(re.prefix)
prefixed = true
}
arena := &matchArena{nil, 2 * (re.nbra + 1)}
for pos <= end {
@ -825,7 +828,12 @@ func (re *Regexp) doExecute(str string, bytestr []byte, pos int) []int {
// prime the pump if we haven't seen a match yet
match := arena.noMatch()
match.m[0] = pos
s[out] = arena.addState(s[out], re.start.next(), match, pos, end)
if prefixed {
s[out] = arena.addState(s[out], re.prefixStart, match, pos, end)
prefixed = false // next iteration should start at beginning of machine.
} else {
s[out] = arena.addState(s[out], re.start.next(), match, pos, end)
}
arena.free(match) // if addState saved it, ref was incremented
}
in, out = out, in // old out state is new in state