index/suffixarray: revert change from int -> int32
CL 5040041 (https://golang.org/cl/5040041)
changed the use of []int to []int32 internally so
that encoding/binary could be used. This is no
longer needed (gobs can encode ints), and using
[]int is more in sync w/ the semantics of the data
structure (the index elements are indices which are
ints). Changing it back.
R=r
CC=golang-dev
https://golang.org/cl/5141049
diff --git a/src/pkg/index/suffixarray/qsufsort.go b/src/pkg/index/suffixarray/qsufsort.go
index f4ec3a1..c69be43 100644
--- a/src/pkg/index/suffixarray/qsufsort.go
+++ b/src/pkg/index/suffixarray/qsufsort.go
@@ -26,7 +26,7 @@
import "sort"
-func qsufsort(data []byte) []int32 {
+func qsufsort(data []byte) []int {
// initial sorting by first byte of suffix
sa := sortedByFirstByte(data)
if len(sa) < 2 {
@@ -39,20 +39,20 @@
// the index starts 1-ordered
sufSortable := &suffixSortable{sa: sa, inv: inv, h: 1}
- for int(sa[0]) > -len(sa) { // until all suffixes are one big sorted group
+ for sa[0] > -len(sa) { // until all suffixes are one big sorted group
// The suffixes are h-ordered, make them 2*h-ordered
pi := 0 // pi is first position of first group
sl := 0 // sl is negated length of sorted groups
for pi < len(sa) {
- if s := int(sa[pi]); s < 0 { // if pi starts sorted group
+ if s := sa[pi]; s < 0 { // if pi starts sorted group
pi -= s // skip over sorted group
sl += s // add negated length to sl
} else { // if pi starts unsorted group
if sl != 0 {
- sa[pi+sl] = int32(sl) // combine sorted groups before pi
+ sa[pi+sl] = sl // combine sorted groups before pi
sl = 0
}
- pk := int(inv[s]) + 1 // pk-1 is last position of unsorted group
+ pk := inv[s] + 1 // pk-1 is last position of unsorted group
sufSortable.sa = sa[pi:pk]
sort.Sort(sufSortable)
sufSortable.updateGroups(pi)
@@ -60,19 +60,19 @@
}
}
if sl != 0 { // if the array ends with a sorted group
- sa[pi+sl] = int32(sl) // combine sorted groups at end of sa
+ sa[pi+sl] = sl // combine sorted groups at end of sa
}
sufSortable.h *= 2 // double sorted depth
}
for i := range sa { // reconstruct suffix array from inverse
- sa[inv[i]] = int32(i)
+ sa[inv[i]] = i
}
return sa
}
-func sortedByFirstByte(data []byte) []int32 {
+func sortedByFirstByte(data []byte) []int {
// total byte counts
var count [256]int
for _, b := range data {
@@ -84,17 +84,17 @@
count[b], sum = sum, count[b]+sum
}
// iterate through bytes, placing index into the correct spot in sa
- sa := make([]int32, len(data))
+ sa := make([]int, len(data))
for i, b := range data {
- sa[count[b]] = int32(i)
+ sa[count[b]] = i
count[b]++
}
return sa
}
-func initGroups(sa []int32, data []byte) []int32 {
+func initGroups(sa []int, data []byte) []int {
// label contiguous same-letter groups with the same group number
- inv := make([]int32, len(data))
+ inv := make([]int, len(data))
prevGroup := len(sa) - 1
groupByte := data[sa[prevGroup]]
for i := len(sa) - 1; i >= 0; i-- {
@@ -105,7 +105,7 @@
groupByte = b
prevGroup = i
}
- inv[sa[i]] = int32(prevGroup)
+ inv[sa[i]] = prevGroup
if prevGroup == 0 {
sa[0] = -1
}
@@ -120,9 +120,9 @@
if data[sa[i]] == lastByte && s == -1 {
s = i
}
- if int(sa[i]) == len(sa)-1 {
+ if sa[i] == len(sa)-1 {
sa[i], sa[s] = sa[s], sa[i]
- inv[sa[s]] = int32(s)
+ inv[sa[s]] = s
sa[s] = -1 // mark it as an isolated sorted group
break
}
@@ -132,9 +132,9 @@
}
type suffixSortable struct {
- sa []int32
- inv []int32
- h int32
+ sa []int
+ inv []int
+ h int
buf []int // common scratch space
}
@@ -158,7 +158,7 @@
prev := 0
for _, b := range bounds {
for i := prev; i < b; i++ {
- x.inv[x.sa[i]] = int32(offset + b - 1)
+ x.inv[x.sa[i]] = offset + b - 1
}
if b-prev == 1 {
x.sa[prev] = -1
diff --git a/src/pkg/index/suffixarray/suffixarray.go b/src/pkg/index/suffixarray/suffixarray.go
index 4747ba9..05b06c6 100644
--- a/src/pkg/index/suffixarray/suffixarray.go
+++ b/src/pkg/index/suffixarray/suffixarray.go
@@ -28,7 +28,7 @@
// Index implements a suffix array for fast substring search.
type Index struct {
data []byte
- sa []int32 // suffix array for data; len(sa) == len(data)
+ sa []int // suffix array for data; len(sa) == len(data)
}
// New creates a new Index for data.
@@ -52,7 +52,7 @@
// new data is significantly smaller or larger then
// existing buffers - allocate new ones
x.data = make([]byte, n)
- x.sa = make([]int32, n)
+ x.sa = make([]int, n)
} else {
// re-use existing buffers
x.data = x.data[0:n]
@@ -120,7 +120,7 @@
// lookupAll returns a slice into the matching region of the index.
// The runtime is O(log(N)*len(s)).
-func (x *Index) lookupAll(s []byte) []int32 {
+func (x *Index) lookupAll(s []byte) []int {
// find matching suffix index range [i:j]
// find the first index where s would be the prefix
i := sort.Search(len(x.sa), func(i int) bool { return bytes.Compare(x.at(i), s) >= 0 })
@@ -144,9 +144,7 @@
// 0 <= n <= len(matches)
if n > 0 {
result = make([]int, n)
- for i, x := range matches[0:n] {
- result[i] = int(x)
- }
+ copy(result, matches)
}
}
return
