utf8.String: Slice(i,j)

R=rsc
CC=golang-dev
https://golang.org/cl/2225048

diff --git a/src/pkg/utf8/string.go b/src/pkg/utf8/string.go
index ce74f720dbf584e501f9b2f37b8e09b7dc6aad65..59107ac19db26e10517874fab4dae7e292af5559 100644 (file)
--- a/src/pkg/utf8/string.go
+++ b/src/pkg/utf8/string.go
@@ -11,6 +11,8 @@ package utf8
 // O(N) in the length of the string, but the overhead is less than always
 // scanning from the beginning.
 // If the string is ASCII, random access is O(1).
+// Unlike the built-in string type, String has internal mutable state and
+// is not thread-safe.
 type String struct {
        str      string
        numRunes int
@@ -55,6 +57,39 @@ func (s *String) IsASCII() bool {
        return s.width == 0
 }
 
+// Slice returns the string sliced at rune positions [i:j].
+func (s *String) Slice(i, j int) string {
+       // ASCII is easy.  Let the compiler catch the indexing error if there is one.
+       if j < s.nonASCII {
+               return s.str[i:j]
+       }
+       if i < 0 || j > s.numRunes || i > j {
+               panic(sliceOutOfRange)
+       }
+       if i == j {
+               return ""
+       }
+       // For non-ASCII, after At(i), bytePos is always the position of the indexed character.
+       var low, high int
+       switch {
+       case i < s.nonASCII:
+               low = i
+       case i == s.numRunes:
+               low = len(s.str)
+       default:
+               s.At(i)
+               low = s.bytePos
+       }
+       switch {
+       case j == s.numRunes:
+               high = len(s.str)
+       default:
+               s.At(j)
+               high = s.bytePos
+       }
+       return s.str[low:high]
+}
+
 // At returns the rune with index i in the String.  The sequence of runes is the same
 // as iterating over the contents with a "for range" clause.
 func (s *String) At(i int) int {
@@ -163,4 +198,5 @@ func (err error) String() string {
 func (err error) RunTimeError() {
 }
 
-var outOfRange = error("utf8.String: index out of Range")
+var outOfRange = error("utf8.String: index out of range")
+var sliceOutOfRange = error("utf8.String: slice index out of range")

diff --git a/src/pkg/utf8/string_test.go b/src/pkg/utf8/string_test.go
index 7ce8f8a7ea1a558a703805d19f95a23004b33307..484d46fbffde68855c1a8e68662502aa7f033055 100644 (file)
--- a/src/pkg/utf8/string_test.go
+++ b/src/pkg/utf8/string_test.go
@@ -68,3 +68,42 @@ func TestRandomAccess(t *testing.T) {
                }
        }
 }
+
+func TestRandomSliceAccess(t *testing.T) {
+       for _, s := range testStrings {
+               if len(s) == 0 || s[0] == '\x80' { // the bad-UTF-8 string fools this simple test
+                       continue
+               }
+               runes := []int(s)
+               str := NewString(s)
+               if str.RuneCount() != len(runes) {
+                       t.Error("%s: expected %d runes; got %d", s, len(runes), str.RuneCount())
+                       break
+               }
+               for k := 0; k < randCount; k++ {
+                       i := rand.Intn(len(runes))
+                       j := rand.Intn(len(runes) + 1)
+                       if i > j { // include empty strings
+                               continue
+                       }
+                       expect := string(runes[i:j])
+                       got := str.Slice(i, j)
+                       if got != expect {
+                               t.Errorf("%s[%d:%d]: expected %q got %q", s, i, j, expect, got)
+                       }
+               }
+       }
+}
+
+func TestLimitSliceAccess(t *testing.T) {
+       for _, s := range testStrings {
+               str := NewString(s)
+               if str.Slice(0, 0) != "" {
+                       t.Error("failure with empty slice at beginning")
+               }
+               nr := RuneCountInString(s)
+               if str.Slice(nr, nr) != "" {
+                       t.Error("failure with empty slice at end")
+               }
+       }
+}