2009-06-04 14:41:31 -07:00
|
|
|
|
// Copyright 2009 The Go Authors. All rights reserved.
|
|
|
|
|
|
// Use of this source code is governed by a BSD-style
|
|
|
|
|
|
// license that can be found in the LICENSE file.
|
|
|
|
|
|
|
2009-08-12 13:18:37 -07:00
|
|
|
|
package bytes_test
|
2009-06-04 14:41:31 -07:00
|
|
|
|
|
|
|
|
|
|
import (
|
2009-12-15 15:33:31 -08:00
|
|
|
|
. "bytes"
|
2016-05-26 10:54:25 +02:00
|
|
|
|
"fmt"
|
2016-09-14 18:03:26 +00:00
|
|
|
|
"internal/testenv"
|
2012-09-18 15:02:08 -04:00
|
|
|
|
"math/rand"
|
2011-06-28 09:43:14 +10:00
|
|
|
|
"reflect"
|
2016-09-14 18:03:26 +00:00
|
|
|
|
"strings"
|
2009-12-15 15:33:31 -08:00
|
|
|
|
"testing"
|
|
|
|
|
|
"unicode"
|
2011-11-08 15:40:58 -08:00
|
|
|
|
"unicode/utf8"
|
2009-06-04 14:41:31 -07:00
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
|
|
func eq(a, b []string) bool {
|
|
|
|
|
|
if len(a) != len(b) {
|
2009-11-09 12:07:39 -08:00
|
|
|
|
return false
|
2009-06-04 14:41:31 -07:00
|
|
|
|
}
|
|
|
|
|
|
for i := 0; i < len(a); i++ {
|
|
|
|
|
|
if a[i] != b[i] {
|
2009-11-09 12:07:39 -08:00
|
|
|
|
return false
|
2009-06-04 14:41:31 -07:00
|
|
|
|
}
|
|
|
|
|
|
}
|
2009-12-15 15:33:31 -08:00
|
|
|
|
return true
|
2009-06-04 14:41:31 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
2013-01-07 10:48:06 +11:00
|
|
|
|
func sliceOfString(s [][]byte) []string {
|
|
|
|
|
|
result := make([]string, len(s))
|
|
|
|
|
|
for i, v := range s {
|
|
|
|
|
|
result[i] = string(v)
|
2009-06-04 14:41:31 -07:00
|
|
|
|
}
|
2009-12-15 15:33:31 -08:00
|
|
|
|
return result
|
2009-06-04 14:41:31 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
// For ease of reading, the test cases use strings that are converted to byte
|
2013-01-07 10:48:06 +11:00
|
|
|
|
// slices before invoking the functions.
|
2009-06-04 14:41:31 -07:00
|
|
|
|
|
|
|
|
|
|
var abcd = "abcd"
|
|
|
|
|
|
var faces = "☺☻☹"
|
|
|
|
|
|
var commas = "1,2,3,4"
|
|
|
|
|
|
var dots = "1....2....3....4"
|
|
|
|
|
|
|
2009-11-18 19:23:08 -08:00
|
|
|
|
type BinOpTest struct {
|
2009-12-15 15:33:31 -08:00
|
|
|
|
a string
|
|
|
|
|
|
b string
|
|
|
|
|
|
i int
|
2009-06-04 14:41:31 -07:00
|
|
|
|
}
|
2009-10-08 15:14:54 -07:00
|
|
|
|
|
2013-05-14 16:05:51 -07:00
|
|
|
|
func TestEqual(t *testing.T) {
|
2012-02-18 17:39:40 +11:00
|
|
|
|
for _, tt := range compareTests {
|
|
|
|
|
|
eql := Equal(tt.a, tt.b)
|
2009-11-18 19:23:08 -08:00
|
|
|
|
if eql != (tt.i == 0) {
|
2009-11-09 12:07:39 -08:00
|
|
|
|
t.Errorf(`Equal(%q, %q) = %v`, tt.a, tt.b, eql)
|
2009-06-04 14:41:31 -07:00
|
|
|
|
}
|
2012-02-18 17:39:40 +11:00
|
|
|
|
eql = EqualPortable(tt.a, tt.b)
|
2011-12-07 15:09:56 -05:00
|
|
|
|
if eql != (tt.i == 0) {
|
|
|
|
|
|
t.Errorf(`EqualPortable(%q, %q) = %v`, tt.a, tt.b, eql)
|
|
|
|
|
|
}
|
2009-06-04 14:41:31 -07:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2013-05-14 16:05:51 -07:00
|
|
|
|
func TestEqualExhaustive(t *testing.T) {
|
2013-04-02 16:26:15 -07:00
|
|
|
|
var size = 128
|
|
|
|
|
|
if testing.Short() {
|
|
|
|
|
|
size = 32
|
|
|
|
|
|
}
|
|
|
|
|
|
a := make([]byte, size)
|
|
|
|
|
|
b := make([]byte, size)
|
|
|
|
|
|
b_init := make([]byte, size)
|
|
|
|
|
|
// randomish but deterministic data
|
|
|
|
|
|
for i := 0; i < size; i++ {
|
|
|
|
|
|
a[i] = byte(17 * i)
|
|
|
|
|
|
b_init[i] = byte(23*i + 100)
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
for len := 0; len <= size; len++ {
|
|
|
|
|
|
for x := 0; x <= size-len; x++ {
|
|
|
|
|
|
for y := 0; y <= size-len; y++ {
|
|
|
|
|
|
copy(b, b_init)
|
|
|
|
|
|
copy(b[y:y+len], a[x:x+len])
|
|
|
|
|
|
if !Equal(a[x:x+len], b[y:y+len]) || !Equal(b[y:y+len], a[x:x+len]) {
|
|
|
|
|
|
t.Errorf("Equal(%d, %d, %d) = false", len, x, y)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2016-03-01 23:21:55 +00:00
|
|
|
|
// make sure Equal returns false for minimally different strings. The data
|
2013-04-02 16:26:15 -07:00
|
|
|
|
// is all zeros except for a single one in one location.
|
|
|
|
|
|
func TestNotEqual(t *testing.T) {
|
|
|
|
|
|
var size = 128
|
|
|
|
|
|
if testing.Short() {
|
|
|
|
|
|
size = 32
|
|
|
|
|
|
}
|
|
|
|
|
|
a := make([]byte, size)
|
|
|
|
|
|
b := make([]byte, size)
|
|
|
|
|
|
|
|
|
|
|
|
for len := 0; len <= size; len++ {
|
|
|
|
|
|
for x := 0; x <= size-len; x++ {
|
|
|
|
|
|
for y := 0; y <= size-len; y++ {
|
|
|
|
|
|
for diffpos := x; diffpos < x+len; diffpos++ {
|
|
|
|
|
|
a[diffpos] = 1
|
|
|
|
|
|
if Equal(a[x:x+len], b[y:y+len]) || Equal(b[y:y+len], a[x:x+len]) {
|
|
|
|
|
|
t.Errorf("NotEqual(%d, %d, %d, %d) = true", len, x, y, diffpos)
|
|
|
|
|
|
}
|
|
|
|
|
|
a[diffpos] = 0
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2010-03-26 13:05:04 -07:00
|
|
|
|
var indexTests = []BinOpTest{
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{"", "", 0},
|
|
|
|
|
|
{"", "a", -1},
|
|
|
|
|
|
{"", "foo", -1},
|
|
|
|
|
|
{"fo", "foo", -1},
|
2013-09-11 21:20:15 +10:00
|
|
|
|
{"foo", "baz", -1},
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{"foo", "foo", 0},
|
|
|
|
|
|
{"oofofoofooo", "f", 2},
|
|
|
|
|
|
{"oofofoofooo", "foo", 4},
|
|
|
|
|
|
{"barfoobarfoo", "foo", 3},
|
|
|
|
|
|
{"foo", "", 0},
|
|
|
|
|
|
{"foo", "o", 1},
|
|
|
|
|
|
{"abcABCabc", "A", 3},
|
2010-03-26 13:05:04 -07:00
|
|
|
|
// cases with one byte strings - test IndexByte and special case in Index()
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{"", "a", -1},
|
|
|
|
|
|
{"x", "a", -1},
|
|
|
|
|
|
{"x", "x", 0},
|
|
|
|
|
|
{"abc", "a", 0},
|
|
|
|
|
|
{"abc", "b", 1},
|
|
|
|
|
|
{"abc", "c", 2},
|
|
|
|
|
|
{"abc", "x", -1},
|
2010-11-08 17:33:53 -08:00
|
|
|
|
{"barfoobarfooyyyzzzyyyzzzyyyzzzyyyxxxzzzyyy", "x", 33},
|
|
|
|
|
|
{"foofyfoobarfoobar", "y", 4},
|
|
|
|
|
|
{"oooooooooooooooooooooo", "r", -1},
|
2017-11-04 10:19:53 -07:00
|
|
|
|
// test fallback to Rabin-Karp.
|
|
|
|
|
|
{"oxoxoxoxoxoxoxoxoxoxoxoy", "oy", 22},
|
|
|
|
|
|
{"oxoxoxoxoxoxoxoxoxoxoxox", "oy", -1},
|
2009-11-18 19:23:08 -08:00
|
|
|
|
}
|
|
|
|
|
|
|
2010-03-26 13:05:04 -07:00
|
|
|
|
var lastIndexTests = []BinOpTest{
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{"", "", 0},
|
|
|
|
|
|
{"", "a", -1},
|
|
|
|
|
|
{"", "foo", -1},
|
|
|
|
|
|
{"fo", "foo", -1},
|
|
|
|
|
|
{"foo", "foo", 0},
|
|
|
|
|
|
{"foo", "f", 0},
|
|
|
|
|
|
{"oofofoofooo", "f", 7},
|
|
|
|
|
|
{"oofofoofooo", "foo", 7},
|
|
|
|
|
|
{"barfoobarfoo", "foo", 9},
|
|
|
|
|
|
{"foo", "", 3},
|
|
|
|
|
|
{"foo", "o", 2},
|
|
|
|
|
|
{"abcABCabc", "A", 3},
|
|
|
|
|
|
{"abcABCabc", "a", 6},
|
2010-03-26 13:05:04 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
var indexAnyTests = []BinOpTest{
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{"", "", -1},
|
|
|
|
|
|
{"", "a", -1},
|
|
|
|
|
|
{"", "abc", -1},
|
|
|
|
|
|
{"a", "", -1},
|
|
|
|
|
|
{"a", "a", 0},
|
|
|
|
|
|
{"aaa", "a", 0},
|
|
|
|
|
|
{"abc", "xyz", -1},
|
|
|
|
|
|
{"abc", "xcz", 2},
|
|
|
|
|
|
{"ab☺c", "x☺yz", 2},
|
bytes, strings: optimize for ASCII sets
In a large codebase within Google, there are thousands of uses of:
ContainsAny|IndexAny|LastIndexAny|Trim|TrimLeft|TrimRight
An analysis of their usage shows that over 97% of them only use character
sets consisting of only ASCII symbols.
Uses of ContainsAny|IndexAny|LastIndexAny:
6% are 1 character (e.g., "\n" or " ")
58% are 2-4 characters (e.g., "<>" or "\r\n\t ")
24% are 5-9 characters (e.g., "()[]*^$")
10% are 10+ characters (e.g., "+-=&|><!(){}[]^\"~*?:\\/ ")
We optimize for ASCII sets, which are commonly used to search for
"control" characters in some string. We don't optimize for the
single character scenario since IndexRune or IndexByte could be used.
Uses of Trim|TrimLeft|TrimRight:
71% are 1 character (e.g., "\n" or " ")
14% are 2 characters (e.g., "\r\n")
10% are 3-4 characters (e.g., " \t\r\n")
5% are 10+ characters (e.g., "0123456789abcdefABCDEF")
We optimize for the single character case with a simple closured function
that only checks for that character's value. We optimize for the medium
and larger sets using a 16-byte bit-map representing a set of ASCII characters.
The benchmarks below have the following suffix name "%d:%d" where the first
number is the length of the input and the second number is the length
of the charset.
== bytes package ==
benchmark old ns/op new ns/op delta
BenchmarkIndexAnyASCII/1:1-4 5.09 5.23 +2.75%
BenchmarkIndexAnyASCII/1:2-4 5.81 5.85 +0.69%
BenchmarkIndexAnyASCII/1:4-4 7.22 7.50 +3.88%
BenchmarkIndexAnyASCII/1:8-4 11.0 11.1 +0.91%
BenchmarkIndexAnyASCII/1:16-4 17.5 17.8 +1.71%
BenchmarkIndexAnyASCII/16:1-4 36.0 34.0 -5.56%
BenchmarkIndexAnyASCII/16:2-4 46.6 36.5 -21.67%
BenchmarkIndexAnyASCII/16:4-4 78.0 40.4 -48.21%
BenchmarkIndexAnyASCII/16:8-4 136 47.4 -65.15%
BenchmarkIndexAnyASCII/16:16-4 254 61.5 -75.79%
BenchmarkIndexAnyASCII/256:1-4 542 388 -28.41%
BenchmarkIndexAnyASCII/256:2-4 705 382 -45.82%
BenchmarkIndexAnyASCII/256:4-4 1089 386 -64.55%
BenchmarkIndexAnyASCII/256:8-4 1994 394 -80.24%
BenchmarkIndexAnyASCII/256:16-4 3843 411 -89.31%
BenchmarkIndexAnyASCII/4096:1-4 8522 5873 -31.08%
BenchmarkIndexAnyASCII/4096:2-4 11253 5861 -47.92%
BenchmarkIndexAnyASCII/4096:4-4 17824 5883 -66.99%
BenchmarkIndexAnyASCII/4096:8-4 32053 5871 -81.68%
BenchmarkIndexAnyASCII/4096:16-4 60512 5888 -90.27%
BenchmarkTrimASCII/1:1-4 79.5 70.8 -10.94%
BenchmarkTrimASCII/1:2-4 79.0 105 +32.91%
BenchmarkTrimASCII/1:4-4 79.6 109 +36.93%
BenchmarkTrimASCII/1:8-4 78.8 118 +49.75%
BenchmarkTrimASCII/1:16-4 80.2 132 +64.59%
BenchmarkTrimASCII/16:1-4 243 116 -52.26%
BenchmarkTrimASCII/16:2-4 243 171 -29.63%
BenchmarkTrimASCII/16:4-4 243 176 -27.57%
BenchmarkTrimASCII/16:8-4 241 184 -23.65%
BenchmarkTrimASCII/16:16-4 238 199 -16.39%
BenchmarkTrimASCII/256:1-4 2580 840 -67.44%
BenchmarkTrimASCII/256:2-4 2603 1175 -54.86%
BenchmarkTrimASCII/256:4-4 2572 1188 -53.81%
BenchmarkTrimASCII/256:8-4 2550 1191 -53.29%
BenchmarkTrimASCII/256:16-4 2585 1208 -53.27%
BenchmarkTrimASCII/4096:1-4 39773 12181 -69.37%
BenchmarkTrimASCII/4096:2-4 39946 17231 -56.86%
BenchmarkTrimASCII/4096:4-4 39641 17179 -56.66%
BenchmarkTrimASCII/4096:8-4 39835 17175 -56.88%
BenchmarkTrimASCII/4096:16-4 40229 17215 -57.21%
== strings package ==
benchmark old ns/op new ns/op delta
BenchmarkIndexAnyASCII/1:1-4 5.94 4.97 -16.33%
BenchmarkIndexAnyASCII/1:2-4 5.94 5.55 -6.57%
BenchmarkIndexAnyASCII/1:4-4 7.45 7.21 -3.22%
BenchmarkIndexAnyASCII/1:8-4 10.8 10.6 -1.85%
BenchmarkIndexAnyASCII/1:16-4 17.4 17.2 -1.15%
BenchmarkIndexAnyASCII/16:1-4 36.4 32.2 -11.54%
BenchmarkIndexAnyASCII/16:2-4 49.6 34.6 -30.24%
BenchmarkIndexAnyASCII/16:4-4 77.5 37.9 -51.10%
BenchmarkIndexAnyASCII/16:8-4 138 45.5 -67.03%
BenchmarkIndexAnyASCII/16:16-4 241 59.1 -75.48%
BenchmarkIndexAnyASCII/256:1-4 509 378 -25.74%
BenchmarkIndexAnyASCII/256:2-4 720 381 -47.08%
BenchmarkIndexAnyASCII/256:4-4 1142 384 -66.37%
BenchmarkIndexAnyASCII/256:8-4 1999 391 -80.44%
BenchmarkIndexAnyASCII/256:16-4 3735 403 -89.21%
BenchmarkIndexAnyASCII/4096:1-4 7973 5824 -26.95%
BenchmarkIndexAnyASCII/4096:2-4 11432 5809 -49.19%
BenchmarkIndexAnyASCII/4096:4-4 18327 5819 -68.25%
BenchmarkIndexAnyASCII/4096:8-4 33059 5828 -82.37%
BenchmarkIndexAnyASCII/4096:16-4 59703 5817 -90.26%
BenchmarkTrimASCII/1:1-4 71.9 71.8 -0.14%
BenchmarkTrimASCII/1:2-4 73.3 103 +40.52%
BenchmarkTrimASCII/1:4-4 71.8 106 +47.63%
BenchmarkTrimASCII/1:8-4 71.2 113 +58.71%
BenchmarkTrimASCII/1:16-4 71.6 128 +78.77%
BenchmarkTrimASCII/16:1-4 152 116 -23.68%
BenchmarkTrimASCII/16:2-4 160 168 +5.00%
BenchmarkTrimASCII/16:4-4 172 170 -1.16%
BenchmarkTrimASCII/16:8-4 200 177 -11.50%
BenchmarkTrimASCII/16:16-4 254 193 -24.02%
BenchmarkTrimASCII/256:1-4 1438 864 -39.92%
BenchmarkTrimASCII/256:2-4 1551 1195 -22.95%
BenchmarkTrimASCII/256:4-4 1770 1200 -32.20%
BenchmarkTrimASCII/256:8-4 2195 1216 -44.60%
BenchmarkTrimASCII/256:16-4 3054 1224 -59.92%
BenchmarkTrimASCII/4096:1-4 21726 12557 -42.20%
BenchmarkTrimASCII/4096:2-4 23586 17508 -25.77%
BenchmarkTrimASCII/4096:4-4 26898 17510 -34.90%
BenchmarkTrimASCII/4096:8-4 33714 17595 -47.81%
BenchmarkTrimASCII/4096:16-4 47429 17700 -62.68%
The benchmarks added test the worst case. For IndexAny, that is when the
charset matches none of the input. For Trim, it is when the charset matches
all of the input.
Change-Id: I970874d101a96b33528fc99b165379abe58cf6ea
Reviewed-on: https://go-review.googlesource.com/31593
Run-TryBot: Joe Tsai <thebrokentoaster@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Martin Möhrmann <martisch@uos.de>
2016-10-20 03:16:22 -07:00
|
|
|
|
{"a☺b☻c☹d", "cx", len("a☺b☻")},
|
|
|
|
|
|
{"a☺b☻c☹d", "uvw☻xyz", len("a☺b")},
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{"aRegExp*", ".(|)*+?^$[]", 7},
|
|
|
|
|
|
{dots + dots + dots, " ", -1},
|
bytes, strings: optimize for ASCII sets
In a large codebase within Google, there are thousands of uses of:
ContainsAny|IndexAny|LastIndexAny|Trim|TrimLeft|TrimRight
An analysis of their usage shows that over 97% of them only use character
sets consisting of only ASCII symbols.
Uses of ContainsAny|IndexAny|LastIndexAny:
6% are 1 character (e.g., "\n" or " ")
58% are 2-4 characters (e.g., "<>" or "\r\n\t ")
24% are 5-9 characters (e.g., "()[]*^$")
10% are 10+ characters (e.g., "+-=&|><!(){}[]^\"~*?:\\/ ")
We optimize for ASCII sets, which are commonly used to search for
"control" characters in some string. We don't optimize for the
single character scenario since IndexRune or IndexByte could be used.
Uses of Trim|TrimLeft|TrimRight:
71% are 1 character (e.g., "\n" or " ")
14% are 2 characters (e.g., "\r\n")
10% are 3-4 characters (e.g., " \t\r\n")
5% are 10+ characters (e.g., "0123456789abcdefABCDEF")
We optimize for the single character case with a simple closured function
that only checks for that character's value. We optimize for the medium
and larger sets using a 16-byte bit-map representing a set of ASCII characters.
The benchmarks below have the following suffix name "%d:%d" where the first
number is the length of the input and the second number is the length
of the charset.
== bytes package ==
benchmark old ns/op new ns/op delta
BenchmarkIndexAnyASCII/1:1-4 5.09 5.23 +2.75%
BenchmarkIndexAnyASCII/1:2-4 5.81 5.85 +0.69%
BenchmarkIndexAnyASCII/1:4-4 7.22 7.50 +3.88%
BenchmarkIndexAnyASCII/1:8-4 11.0 11.1 +0.91%
BenchmarkIndexAnyASCII/1:16-4 17.5 17.8 +1.71%
BenchmarkIndexAnyASCII/16:1-4 36.0 34.0 -5.56%
BenchmarkIndexAnyASCII/16:2-4 46.6 36.5 -21.67%
BenchmarkIndexAnyASCII/16:4-4 78.0 40.4 -48.21%
BenchmarkIndexAnyASCII/16:8-4 136 47.4 -65.15%
BenchmarkIndexAnyASCII/16:16-4 254 61.5 -75.79%
BenchmarkIndexAnyASCII/256:1-4 542 388 -28.41%
BenchmarkIndexAnyASCII/256:2-4 705 382 -45.82%
BenchmarkIndexAnyASCII/256:4-4 1089 386 -64.55%
BenchmarkIndexAnyASCII/256:8-4 1994 394 -80.24%
BenchmarkIndexAnyASCII/256:16-4 3843 411 -89.31%
BenchmarkIndexAnyASCII/4096:1-4 8522 5873 -31.08%
BenchmarkIndexAnyASCII/4096:2-4 11253 5861 -47.92%
BenchmarkIndexAnyASCII/4096:4-4 17824 5883 -66.99%
BenchmarkIndexAnyASCII/4096:8-4 32053 5871 -81.68%
BenchmarkIndexAnyASCII/4096:16-4 60512 5888 -90.27%
BenchmarkTrimASCII/1:1-4 79.5 70.8 -10.94%
BenchmarkTrimASCII/1:2-4 79.0 105 +32.91%
BenchmarkTrimASCII/1:4-4 79.6 109 +36.93%
BenchmarkTrimASCII/1:8-4 78.8 118 +49.75%
BenchmarkTrimASCII/1:16-4 80.2 132 +64.59%
BenchmarkTrimASCII/16:1-4 243 116 -52.26%
BenchmarkTrimASCII/16:2-4 243 171 -29.63%
BenchmarkTrimASCII/16:4-4 243 176 -27.57%
BenchmarkTrimASCII/16:8-4 241 184 -23.65%
BenchmarkTrimASCII/16:16-4 238 199 -16.39%
BenchmarkTrimASCII/256:1-4 2580 840 -67.44%
BenchmarkTrimASCII/256:2-4 2603 1175 -54.86%
BenchmarkTrimASCII/256:4-4 2572 1188 -53.81%
BenchmarkTrimASCII/256:8-4 2550 1191 -53.29%
BenchmarkTrimASCII/256:16-4 2585 1208 -53.27%
BenchmarkTrimASCII/4096:1-4 39773 12181 -69.37%
BenchmarkTrimASCII/4096:2-4 39946 17231 -56.86%
BenchmarkTrimASCII/4096:4-4 39641 17179 -56.66%
BenchmarkTrimASCII/4096:8-4 39835 17175 -56.88%
BenchmarkTrimASCII/4096:16-4 40229 17215 -57.21%
== strings package ==
benchmark old ns/op new ns/op delta
BenchmarkIndexAnyASCII/1:1-4 5.94 4.97 -16.33%
BenchmarkIndexAnyASCII/1:2-4 5.94 5.55 -6.57%
BenchmarkIndexAnyASCII/1:4-4 7.45 7.21 -3.22%
BenchmarkIndexAnyASCII/1:8-4 10.8 10.6 -1.85%
BenchmarkIndexAnyASCII/1:16-4 17.4 17.2 -1.15%
BenchmarkIndexAnyASCII/16:1-4 36.4 32.2 -11.54%
BenchmarkIndexAnyASCII/16:2-4 49.6 34.6 -30.24%
BenchmarkIndexAnyASCII/16:4-4 77.5 37.9 -51.10%
BenchmarkIndexAnyASCII/16:8-4 138 45.5 -67.03%
BenchmarkIndexAnyASCII/16:16-4 241 59.1 -75.48%
BenchmarkIndexAnyASCII/256:1-4 509 378 -25.74%
BenchmarkIndexAnyASCII/256:2-4 720 381 -47.08%
BenchmarkIndexAnyASCII/256:4-4 1142 384 -66.37%
BenchmarkIndexAnyASCII/256:8-4 1999 391 -80.44%
BenchmarkIndexAnyASCII/256:16-4 3735 403 -89.21%
BenchmarkIndexAnyASCII/4096:1-4 7973 5824 -26.95%
BenchmarkIndexAnyASCII/4096:2-4 11432 5809 -49.19%
BenchmarkIndexAnyASCII/4096:4-4 18327 5819 -68.25%
BenchmarkIndexAnyASCII/4096:8-4 33059 5828 -82.37%
BenchmarkIndexAnyASCII/4096:16-4 59703 5817 -90.26%
BenchmarkTrimASCII/1:1-4 71.9 71.8 -0.14%
BenchmarkTrimASCII/1:2-4 73.3 103 +40.52%
BenchmarkTrimASCII/1:4-4 71.8 106 +47.63%
BenchmarkTrimASCII/1:8-4 71.2 113 +58.71%
BenchmarkTrimASCII/1:16-4 71.6 128 +78.77%
BenchmarkTrimASCII/16:1-4 152 116 -23.68%
BenchmarkTrimASCII/16:2-4 160 168 +5.00%
BenchmarkTrimASCII/16:4-4 172 170 -1.16%
BenchmarkTrimASCII/16:8-4 200 177 -11.50%
BenchmarkTrimASCII/16:16-4 254 193 -24.02%
BenchmarkTrimASCII/256:1-4 1438 864 -39.92%
BenchmarkTrimASCII/256:2-4 1551 1195 -22.95%
BenchmarkTrimASCII/256:4-4 1770 1200 -32.20%
BenchmarkTrimASCII/256:8-4 2195 1216 -44.60%
BenchmarkTrimASCII/256:16-4 3054 1224 -59.92%
BenchmarkTrimASCII/4096:1-4 21726 12557 -42.20%
BenchmarkTrimASCII/4096:2-4 23586 17508 -25.77%
BenchmarkTrimASCII/4096:4-4 26898 17510 -34.90%
BenchmarkTrimASCII/4096:8-4 33714 17595 -47.81%
BenchmarkTrimASCII/4096:16-4 47429 17700 -62.68%
The benchmarks added test the worst case. For IndexAny, that is when the
charset matches none of the input. For Trim, it is when the charset matches
all of the input.
Change-Id: I970874d101a96b33528fc99b165379abe58cf6ea
Reviewed-on: https://go-review.googlesource.com/31593
Run-TryBot: Joe Tsai <thebrokentoaster@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Martin Möhrmann <martisch@uos.de>
2016-10-20 03:16:22 -07:00
|
|
|
|
{"012abcba210", "\xffb", 4},
|
|
|
|
|
|
{"012\x80bcb\x80210", "\xffb", 3},
|
2010-03-26 13:05:04 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
2010-11-12 12:47:50 -08:00
|
|
|
|
var lastIndexAnyTests = []BinOpTest{
|
|
|
|
|
|
{"", "", -1},
|
|
|
|
|
|
{"", "a", -1},
|
|
|
|
|
|
{"", "abc", -1},
|
|
|
|
|
|
{"a", "", -1},
|
|
|
|
|
|
{"a", "a", 0},
|
|
|
|
|
|
{"aaa", "a", 2},
|
|
|
|
|
|
{"abc", "xyz", -1},
|
|
|
|
|
|
{"abc", "ab", 1},
|
bytes, strings: optimize for ASCII sets
In a large codebase within Google, there are thousands of uses of:
ContainsAny|IndexAny|LastIndexAny|Trim|TrimLeft|TrimRight
An analysis of their usage shows that over 97% of them only use character
sets consisting of only ASCII symbols.
Uses of ContainsAny|IndexAny|LastIndexAny:
6% are 1 character (e.g., "\n" or " ")
58% are 2-4 characters (e.g., "<>" or "\r\n\t ")
24% are 5-9 characters (e.g., "()[]*^$")
10% are 10+ characters (e.g., "+-=&|><!(){}[]^\"~*?:\\/ ")
We optimize for ASCII sets, which are commonly used to search for
"control" characters in some string. We don't optimize for the
single character scenario since IndexRune or IndexByte could be used.
Uses of Trim|TrimLeft|TrimRight:
71% are 1 character (e.g., "\n" or " ")
14% are 2 characters (e.g., "\r\n")
10% are 3-4 characters (e.g., " \t\r\n")
5% are 10+ characters (e.g., "0123456789abcdefABCDEF")
We optimize for the single character case with a simple closured function
that only checks for that character's value. We optimize for the medium
and larger sets using a 16-byte bit-map representing a set of ASCII characters.
The benchmarks below have the following suffix name "%d:%d" where the first
number is the length of the input and the second number is the length
of the charset.
== bytes package ==
benchmark old ns/op new ns/op delta
BenchmarkIndexAnyASCII/1:1-4 5.09 5.23 +2.75%
BenchmarkIndexAnyASCII/1:2-4 5.81 5.85 +0.69%
BenchmarkIndexAnyASCII/1:4-4 7.22 7.50 +3.88%
BenchmarkIndexAnyASCII/1:8-4 11.0 11.1 +0.91%
BenchmarkIndexAnyASCII/1:16-4 17.5 17.8 +1.71%
BenchmarkIndexAnyASCII/16:1-4 36.0 34.0 -5.56%
BenchmarkIndexAnyASCII/16:2-4 46.6 36.5 -21.67%
BenchmarkIndexAnyASCII/16:4-4 78.0 40.4 -48.21%
BenchmarkIndexAnyASCII/16:8-4 136 47.4 -65.15%
BenchmarkIndexAnyASCII/16:16-4 254 61.5 -75.79%
BenchmarkIndexAnyASCII/256:1-4 542 388 -28.41%
BenchmarkIndexAnyASCII/256:2-4 705 382 -45.82%
BenchmarkIndexAnyASCII/256:4-4 1089 386 -64.55%
BenchmarkIndexAnyASCII/256:8-4 1994 394 -80.24%
BenchmarkIndexAnyASCII/256:16-4 3843 411 -89.31%
BenchmarkIndexAnyASCII/4096:1-4 8522 5873 -31.08%
BenchmarkIndexAnyASCII/4096:2-4 11253 5861 -47.92%
BenchmarkIndexAnyASCII/4096:4-4 17824 5883 -66.99%
BenchmarkIndexAnyASCII/4096:8-4 32053 5871 -81.68%
BenchmarkIndexAnyASCII/4096:16-4 60512 5888 -90.27%
BenchmarkTrimASCII/1:1-4 79.5 70.8 -10.94%
BenchmarkTrimASCII/1:2-4 79.0 105 +32.91%
BenchmarkTrimASCII/1:4-4 79.6 109 +36.93%
BenchmarkTrimASCII/1:8-4 78.8 118 +49.75%
BenchmarkTrimASCII/1:16-4 80.2 132 +64.59%
BenchmarkTrimASCII/16:1-4 243 116 -52.26%
BenchmarkTrimASCII/16:2-4 243 171 -29.63%
BenchmarkTrimASCII/16:4-4 243 176 -27.57%
BenchmarkTrimASCII/16:8-4 241 184 -23.65%
BenchmarkTrimASCII/16:16-4 238 199 -16.39%
BenchmarkTrimASCII/256:1-4 2580 840 -67.44%
BenchmarkTrimASCII/256:2-4 2603 1175 -54.86%
BenchmarkTrimASCII/256:4-4 2572 1188 -53.81%
BenchmarkTrimASCII/256:8-4 2550 1191 -53.29%
BenchmarkTrimASCII/256:16-4 2585 1208 -53.27%
BenchmarkTrimASCII/4096:1-4 39773 12181 -69.37%
BenchmarkTrimASCII/4096:2-4 39946 17231 -56.86%
BenchmarkTrimASCII/4096:4-4 39641 17179 -56.66%
BenchmarkTrimASCII/4096:8-4 39835 17175 -56.88%
BenchmarkTrimASCII/4096:16-4 40229 17215 -57.21%
== strings package ==
benchmark old ns/op new ns/op delta
BenchmarkIndexAnyASCII/1:1-4 5.94 4.97 -16.33%
BenchmarkIndexAnyASCII/1:2-4 5.94 5.55 -6.57%
BenchmarkIndexAnyASCII/1:4-4 7.45 7.21 -3.22%
BenchmarkIndexAnyASCII/1:8-4 10.8 10.6 -1.85%
BenchmarkIndexAnyASCII/1:16-4 17.4 17.2 -1.15%
BenchmarkIndexAnyASCII/16:1-4 36.4 32.2 -11.54%
BenchmarkIndexAnyASCII/16:2-4 49.6 34.6 -30.24%
BenchmarkIndexAnyASCII/16:4-4 77.5 37.9 -51.10%
BenchmarkIndexAnyASCII/16:8-4 138 45.5 -67.03%
BenchmarkIndexAnyASCII/16:16-4 241 59.1 -75.48%
BenchmarkIndexAnyASCII/256:1-4 509 378 -25.74%
BenchmarkIndexAnyASCII/256:2-4 720 381 -47.08%
BenchmarkIndexAnyASCII/256:4-4 1142 384 -66.37%
BenchmarkIndexAnyASCII/256:8-4 1999 391 -80.44%
BenchmarkIndexAnyASCII/256:16-4 3735 403 -89.21%
BenchmarkIndexAnyASCII/4096:1-4 7973 5824 -26.95%
BenchmarkIndexAnyASCII/4096:2-4 11432 5809 -49.19%
BenchmarkIndexAnyASCII/4096:4-4 18327 5819 -68.25%
BenchmarkIndexAnyASCII/4096:8-4 33059 5828 -82.37%
BenchmarkIndexAnyASCII/4096:16-4 59703 5817 -90.26%
BenchmarkTrimASCII/1:1-4 71.9 71.8 -0.14%
BenchmarkTrimASCII/1:2-4 73.3 103 +40.52%
BenchmarkTrimASCII/1:4-4 71.8 106 +47.63%
BenchmarkTrimASCII/1:8-4 71.2 113 +58.71%
BenchmarkTrimASCII/1:16-4 71.6 128 +78.77%
BenchmarkTrimASCII/16:1-4 152 116 -23.68%
BenchmarkTrimASCII/16:2-4 160 168 +5.00%
BenchmarkTrimASCII/16:4-4 172 170 -1.16%
BenchmarkTrimASCII/16:8-4 200 177 -11.50%
BenchmarkTrimASCII/16:16-4 254 193 -24.02%
BenchmarkTrimASCII/256:1-4 1438 864 -39.92%
BenchmarkTrimASCII/256:2-4 1551 1195 -22.95%
BenchmarkTrimASCII/256:4-4 1770 1200 -32.20%
BenchmarkTrimASCII/256:8-4 2195 1216 -44.60%
BenchmarkTrimASCII/256:16-4 3054 1224 -59.92%
BenchmarkTrimASCII/4096:1-4 21726 12557 -42.20%
BenchmarkTrimASCII/4096:2-4 23586 17508 -25.77%
BenchmarkTrimASCII/4096:4-4 26898 17510 -34.90%
BenchmarkTrimASCII/4096:8-4 33714 17595 -47.81%
BenchmarkTrimASCII/4096:16-4 47429 17700 -62.68%
The benchmarks added test the worst case. For IndexAny, that is when the
charset matches none of the input. For Trim, it is when the charset matches
all of the input.
Change-Id: I970874d101a96b33528fc99b165379abe58cf6ea
Reviewed-on: https://go-review.googlesource.com/31593
Run-TryBot: Joe Tsai <thebrokentoaster@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Martin Möhrmann <martisch@uos.de>
2016-10-20 03:16:22 -07:00
|
|
|
|
{"ab☺c", "x☺yz", 2},
|
|
|
|
|
|
{"a☺b☻c☹d", "cx", len("a☺b☻")},
|
|
|
|
|
|
{"a☺b☻c☹d", "uvw☻xyz", len("a☺b")},
|
2010-11-12 12:47:50 -08:00
|
|
|
|
{"a.RegExp*", ".(|)*+?^$[]", 8},
|
|
|
|
|
|
{dots + dots + dots, " ", -1},
|
bytes, strings: optimize for ASCII sets
In a large codebase within Google, there are thousands of uses of:
ContainsAny|IndexAny|LastIndexAny|Trim|TrimLeft|TrimRight
An analysis of their usage shows that over 97% of them only use character
sets consisting of only ASCII symbols.
Uses of ContainsAny|IndexAny|LastIndexAny:
6% are 1 character (e.g., "\n" or " ")
58% are 2-4 characters (e.g., "<>" or "\r\n\t ")
24% are 5-9 characters (e.g., "()[]*^$")
10% are 10+ characters (e.g., "+-=&|><!(){}[]^\"~*?:\\/ ")
We optimize for ASCII sets, which are commonly used to search for
"control" characters in some string. We don't optimize for the
single character scenario since IndexRune or IndexByte could be used.
Uses of Trim|TrimLeft|TrimRight:
71% are 1 character (e.g., "\n" or " ")
14% are 2 characters (e.g., "\r\n")
10% are 3-4 characters (e.g., " \t\r\n")
5% are 10+ characters (e.g., "0123456789abcdefABCDEF")
We optimize for the single character case with a simple closured function
that only checks for that character's value. We optimize for the medium
and larger sets using a 16-byte bit-map representing a set of ASCII characters.
The benchmarks below have the following suffix name "%d:%d" where the first
number is the length of the input and the second number is the length
of the charset.
== bytes package ==
benchmark old ns/op new ns/op delta
BenchmarkIndexAnyASCII/1:1-4 5.09 5.23 +2.75%
BenchmarkIndexAnyASCII/1:2-4 5.81 5.85 +0.69%
BenchmarkIndexAnyASCII/1:4-4 7.22 7.50 +3.88%
BenchmarkIndexAnyASCII/1:8-4 11.0 11.1 +0.91%
BenchmarkIndexAnyASCII/1:16-4 17.5 17.8 +1.71%
BenchmarkIndexAnyASCII/16:1-4 36.0 34.0 -5.56%
BenchmarkIndexAnyASCII/16:2-4 46.6 36.5 -21.67%
BenchmarkIndexAnyASCII/16:4-4 78.0 40.4 -48.21%
BenchmarkIndexAnyASCII/16:8-4 136 47.4 -65.15%
BenchmarkIndexAnyASCII/16:16-4 254 61.5 -75.79%
BenchmarkIndexAnyASCII/256:1-4 542 388 -28.41%
BenchmarkIndexAnyASCII/256:2-4 705 382 -45.82%
BenchmarkIndexAnyASCII/256:4-4 1089 386 -64.55%
BenchmarkIndexAnyASCII/256:8-4 1994 394 -80.24%
BenchmarkIndexAnyASCII/256:16-4 3843 411 -89.31%
BenchmarkIndexAnyASCII/4096:1-4 8522 5873 -31.08%
BenchmarkIndexAnyASCII/4096:2-4 11253 5861 -47.92%
BenchmarkIndexAnyASCII/4096:4-4 17824 5883 -66.99%
BenchmarkIndexAnyASCII/4096:8-4 32053 5871 -81.68%
BenchmarkIndexAnyASCII/4096:16-4 60512 5888 -90.27%
BenchmarkTrimASCII/1:1-4 79.5 70.8 -10.94%
BenchmarkTrimASCII/1:2-4 79.0 105 +32.91%
BenchmarkTrimASCII/1:4-4 79.6 109 +36.93%
BenchmarkTrimASCII/1:8-4 78.8 118 +49.75%
BenchmarkTrimASCII/1:16-4 80.2 132 +64.59%
BenchmarkTrimASCII/16:1-4 243 116 -52.26%
BenchmarkTrimASCII/16:2-4 243 171 -29.63%
BenchmarkTrimASCII/16:4-4 243 176 -27.57%
BenchmarkTrimASCII/16:8-4 241 184 -23.65%
BenchmarkTrimASCII/16:16-4 238 199 -16.39%
BenchmarkTrimASCII/256:1-4 2580 840 -67.44%
BenchmarkTrimASCII/256:2-4 2603 1175 -54.86%
BenchmarkTrimASCII/256:4-4 2572 1188 -53.81%
BenchmarkTrimASCII/256:8-4 2550 1191 -53.29%
BenchmarkTrimASCII/256:16-4 2585 1208 -53.27%
BenchmarkTrimASCII/4096:1-4 39773 12181 -69.37%
BenchmarkTrimASCII/4096:2-4 39946 17231 -56.86%
BenchmarkTrimASCII/4096:4-4 39641 17179 -56.66%
BenchmarkTrimASCII/4096:8-4 39835 17175 -56.88%
BenchmarkTrimASCII/4096:16-4 40229 17215 -57.21%
== strings package ==
benchmark old ns/op new ns/op delta
BenchmarkIndexAnyASCII/1:1-4 5.94 4.97 -16.33%
BenchmarkIndexAnyASCII/1:2-4 5.94 5.55 -6.57%
BenchmarkIndexAnyASCII/1:4-4 7.45 7.21 -3.22%
BenchmarkIndexAnyASCII/1:8-4 10.8 10.6 -1.85%
BenchmarkIndexAnyASCII/1:16-4 17.4 17.2 -1.15%
BenchmarkIndexAnyASCII/16:1-4 36.4 32.2 -11.54%
BenchmarkIndexAnyASCII/16:2-4 49.6 34.6 -30.24%
BenchmarkIndexAnyASCII/16:4-4 77.5 37.9 -51.10%
BenchmarkIndexAnyASCII/16:8-4 138 45.5 -67.03%
BenchmarkIndexAnyASCII/16:16-4 241 59.1 -75.48%
BenchmarkIndexAnyASCII/256:1-4 509 378 -25.74%
BenchmarkIndexAnyASCII/256:2-4 720 381 -47.08%
BenchmarkIndexAnyASCII/256:4-4 1142 384 -66.37%
BenchmarkIndexAnyASCII/256:8-4 1999 391 -80.44%
BenchmarkIndexAnyASCII/256:16-4 3735 403 -89.21%
BenchmarkIndexAnyASCII/4096:1-4 7973 5824 -26.95%
BenchmarkIndexAnyASCII/4096:2-4 11432 5809 -49.19%
BenchmarkIndexAnyASCII/4096:4-4 18327 5819 -68.25%
BenchmarkIndexAnyASCII/4096:8-4 33059 5828 -82.37%
BenchmarkIndexAnyASCII/4096:16-4 59703 5817 -90.26%
BenchmarkTrimASCII/1:1-4 71.9 71.8 -0.14%
BenchmarkTrimASCII/1:2-4 73.3 103 +40.52%
BenchmarkTrimASCII/1:4-4 71.8 106 +47.63%
BenchmarkTrimASCII/1:8-4 71.2 113 +58.71%
BenchmarkTrimASCII/1:16-4 71.6 128 +78.77%
BenchmarkTrimASCII/16:1-4 152 116 -23.68%
BenchmarkTrimASCII/16:2-4 160 168 +5.00%
BenchmarkTrimASCII/16:4-4 172 170 -1.16%
BenchmarkTrimASCII/16:8-4 200 177 -11.50%
BenchmarkTrimASCII/16:16-4 254 193 -24.02%
BenchmarkTrimASCII/256:1-4 1438 864 -39.92%
BenchmarkTrimASCII/256:2-4 1551 1195 -22.95%
BenchmarkTrimASCII/256:4-4 1770 1200 -32.20%
BenchmarkTrimASCII/256:8-4 2195 1216 -44.60%
BenchmarkTrimASCII/256:16-4 3054 1224 -59.92%
BenchmarkTrimASCII/4096:1-4 21726 12557 -42.20%
BenchmarkTrimASCII/4096:2-4 23586 17508 -25.77%
BenchmarkTrimASCII/4096:4-4 26898 17510 -34.90%
BenchmarkTrimASCII/4096:8-4 33714 17595 -47.81%
BenchmarkTrimASCII/4096:16-4 47429 17700 -62.68%
The benchmarks added test the worst case. For IndexAny, that is when the
charset matches none of the input. For Trim, it is when the charset matches
all of the input.
Change-Id: I970874d101a96b33528fc99b165379abe58cf6ea
Reviewed-on: https://go-review.googlesource.com/31593
Run-TryBot: Joe Tsai <thebrokentoaster@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Martin Möhrmann <martisch@uos.de>
2016-10-20 03:16:22 -07:00
|
|
|
|
{"012abcba210", "\xffb", 6},
|
|
|
|
|
|
{"012\x80bcb\x80210", "\xffb", 7},
|
2010-11-12 12:47:50 -08:00
|
|
|
|
}
|
|
|
|
|
|
|
2010-03-26 13:05:04 -07:00
|
|
|
|
// Execute f on each test case. funcName should be the name of f; it's used
|
|
|
|
|
|
// in failure reports.
|
|
|
|
|
|
func runIndexTests(t *testing.T, f func(s, sep []byte) int, funcName string, testCases []BinOpTest) {
|
|
|
|
|
|
for _, test := range testCases {
|
|
|
|
|
|
a := []byte(test.a)
|
|
|
|
|
|
b := []byte(test.b)
|
|
|
|
|
|
actual := f(a, b)
|
|
|
|
|
|
if actual != test.i {
|
|
|
|
|
|
t.Errorf("%s(%q,%q) = %v; want %v", funcName, a, b, actual, test.i)
|
2009-11-18 19:23:08 -08:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2010-11-12 12:47:50 -08:00
|
|
|
|
func runIndexAnyTests(t *testing.T, f func(s []byte, chars string) int, funcName string, testCases []BinOpTest) {
|
|
|
|
|
|
for _, test := range testCases {
|
2010-05-03 10:59:00 -07:00
|
|
|
|
a := []byte(test.a)
|
2010-11-12 12:47:50 -08:00
|
|
|
|
actual := f(a, test.b)
|
2010-05-03 10:59:00 -07:00
|
|
|
|
if actual != test.i {
|
2010-11-12 12:47:50 -08:00
|
|
|
|
t.Errorf("%s(%q,%q) = %v; want %v", funcName, a, test.b, actual, test.i)
|
2010-05-03 10:59:00 -07:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2010-03-26 13:05:04 -07:00
|
|
|
|
|
2010-11-12 12:47:50 -08:00
|
|
|
|
func TestIndex(t *testing.T) { runIndexTests(t, Index, "Index", indexTests) }
|
|
|
|
|
|
func TestLastIndex(t *testing.T) { runIndexTests(t, LastIndex, "LastIndex", lastIndexTests) }
|
|
|
|
|
|
func TestIndexAny(t *testing.T) { runIndexAnyTests(t, IndexAny, "IndexAny", indexAnyTests) }
|
|
|
|
|
|
func TestLastIndexAny(t *testing.T) {
|
|
|
|
|
|
runIndexAnyTests(t, LastIndexAny, "LastIndexAny", lastIndexAnyTests)
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2009-11-18 19:23:08 -08:00
|
|
|
|
func TestIndexByte(t *testing.T) {
|
2010-03-26 13:05:04 -07:00
|
|
|
|
for _, tt := range indexTests {
|
2009-11-18 19:23:08 -08:00
|
|
|
|
if len(tt.b) != 1 {
|
|
|
|
|
|
continue
|
|
|
|
|
|
}
|
2010-02-25 16:01:29 -08:00
|
|
|
|
a := []byte(tt.a)
|
2009-12-15 15:33:31 -08:00
|
|
|
|
b := tt.b[0]
|
|
|
|
|
|
pos := IndexByte(a, b)
|
2009-11-18 19:23:08 -08:00
|
|
|
|
if pos != tt.i {
|
|
|
|
|
|
t.Errorf(`IndexByte(%q, '%c') = %v`, tt.a, b, pos)
|
|
|
|
|
|
}
|
2009-12-15 15:33:31 -08:00
|
|
|
|
posp := IndexBytePortable(a, b)
|
2009-12-04 10:23:43 -08:00
|
|
|
|
if posp != tt.i {
|
|
|
|
|
|
t.Errorf(`indexBytePortable(%q, '%c') = %v`, tt.a, b, posp)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2015-04-29 20:45:55 +03:00
|
|
|
|
func TestLastIndexByte(t *testing.T) {
|
|
|
|
|
|
testCases := []BinOpTest{
|
|
|
|
|
|
{"", "q", -1},
|
|
|
|
|
|
{"abcdef", "q", -1},
|
|
|
|
|
|
{"abcdefabcdef", "a", len("abcdef")}, // something in the middle
|
|
|
|
|
|
{"abcdefabcdef", "f", len("abcdefabcde")}, // last byte
|
|
|
|
|
|
{"zabcdefabcdef", "z", 0}, // first byte
|
|
|
|
|
|
{"a☺b☻c☹d", "b", len("a☺")}, // non-ascii
|
|
|
|
|
|
}
|
|
|
|
|
|
for _, test := range testCases {
|
|
|
|
|
|
actual := LastIndexByte([]byte(test.a), test.b[0])
|
|
|
|
|
|
if actual != test.i {
|
|
|
|
|
|
t.Errorf("LastIndexByte(%q,%c) = %v; want %v", test.a, test.b[0], actual, test.i)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2010-11-08 17:33:53 -08:00
|
|
|
|
// test a larger buffer with different sizes and alignments
|
|
|
|
|
|
func TestIndexByteBig(t *testing.T) {
|
2011-03-25 16:31:10 -07:00
|
|
|
|
var n = 1024
|
|
|
|
|
|
if testing.Short() {
|
|
|
|
|
|
n = 128
|
|
|
|
|
|
}
|
2010-11-08 17:33:53 -08:00
|
|
|
|
b := make([]byte, n)
|
|
|
|
|
|
for i := 0; i < n; i++ {
|
|
|
|
|
|
// different start alignments
|
|
|
|
|
|
b1 := b[i:]
|
|
|
|
|
|
for j := 0; j < len(b1); j++ {
|
|
|
|
|
|
b1[j] = 'x'
|
|
|
|
|
|
pos := IndexByte(b1, 'x')
|
|
|
|
|
|
if pos != j {
|
|
|
|
|
|
t.Errorf("IndexByte(%q, 'x') = %v", b1, pos)
|
|
|
|
|
|
}
|
|
|
|
|
|
b1[j] = 0
|
|
|
|
|
|
pos = IndexByte(b1, 'x')
|
|
|
|
|
|
if pos != -1 {
|
|
|
|
|
|
t.Errorf("IndexByte(%q, 'x') = %v", b1, pos)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
// different end alignments
|
|
|
|
|
|
b1 = b[:i]
|
|
|
|
|
|
for j := 0; j < len(b1); j++ {
|
|
|
|
|
|
b1[j] = 'x'
|
|
|
|
|
|
pos := IndexByte(b1, 'x')
|
|
|
|
|
|
if pos != j {
|
|
|
|
|
|
t.Errorf("IndexByte(%q, 'x') = %v", b1, pos)
|
|
|
|
|
|
}
|
|
|
|
|
|
b1[j] = 0
|
|
|
|
|
|
pos = IndexByte(b1, 'x')
|
|
|
|
|
|
if pos != -1 {
|
|
|
|
|
|
t.Errorf("IndexByte(%q, 'x') = %v", b1, pos)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
// different start and end alignments
|
|
|
|
|
|
b1 = b[i/2 : n-(i+1)/2]
|
|
|
|
|
|
for j := 0; j < len(b1); j++ {
|
|
|
|
|
|
b1[j] = 'x'
|
|
|
|
|
|
pos := IndexByte(b1, 'x')
|
|
|
|
|
|
if pos != j {
|
|
|
|
|
|
t.Errorf("IndexByte(%q, 'x') = %v", b1, pos)
|
|
|
|
|
|
}
|
|
|
|
|
|
b1[j] = 0
|
|
|
|
|
|
pos = IndexByte(b1, 'x')
|
|
|
|
|
|
if pos != -1 {
|
|
|
|
|
|
t.Errorf("IndexByte(%q, 'x') = %v", b1, pos)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2016-01-15 18:17:09 -08:00
|
|
|
|
// test a small index across all page offsets
|
|
|
|
|
|
func TestIndexByteSmall(t *testing.T) {
|
|
|
|
|
|
b := make([]byte, 5015) // bigger than a page
|
|
|
|
|
|
// Make sure we find the correct byte even when straddling a page.
|
|
|
|
|
|
for i := 0; i <= len(b)-15; i++ {
|
|
|
|
|
|
for j := 0; j < 15; j++ {
|
|
|
|
|
|
b[i+j] = byte(100 + j)
|
|
|
|
|
|
}
|
|
|
|
|
|
for j := 0; j < 15; j++ {
|
|
|
|
|
|
p := IndexByte(b[i:i+15], byte(100+j))
|
|
|
|
|
|
if p != j {
|
|
|
|
|
|
t.Errorf("IndexByte(%q, %d) = %d", b[i:i+15], 100+j, p)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
for j := 0; j < 15; j++ {
|
|
|
|
|
|
b[i+j] = 0
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
// Make sure matches outside the slice never trigger.
|
|
|
|
|
|
for i := 0; i <= len(b)-15; i++ {
|
|
|
|
|
|
for j := 0; j < 15; j++ {
|
|
|
|
|
|
b[i+j] = 1
|
|
|
|
|
|
}
|
|
|
|
|
|
for j := 0; j < 15; j++ {
|
|
|
|
|
|
p := IndexByte(b[i:i+15], byte(0))
|
|
|
|
|
|
if p != -1 {
|
|
|
|
|
|
t.Errorf("IndexByte(%q, %d) = %d", b[i:i+15], 0, p)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
for j := 0; j < 15; j++ {
|
|
|
|
|
|
b[i+j] = 0
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2010-08-05 23:11:06 +10:00
|
|
|
|
func TestIndexRune(t *testing.T) {
|
2016-10-26 14:18:37 -07:00
|
|
|
|
tests := []struct {
|
|
|
|
|
|
in string
|
|
|
|
|
|
rune rune
|
|
|
|
|
|
want int
|
|
|
|
|
|
}{
|
|
|
|
|
|
{"", 'a', -1},
|
|
|
|
|
|
{"", '☺', -1},
|
|
|
|
|
|
{"foo", '☹', -1},
|
|
|
|
|
|
{"foo", 'o', 1},
|
|
|
|
|
|
{"foo☺bar", '☺', 3},
|
|
|
|
|
|
{"foo☺☻☹bar", '☹', 9},
|
|
|
|
|
|
{"a A x", 'A', 2},
|
|
|
|
|
|
{"some_text=some_value", '=', 9},
|
|
|
|
|
|
{"☺a", 'a', 3},
|
|
|
|
|
|
{"a☻☺b", '☺', 4},
|
|
|
|
|
|
|
|
|
|
|
|
// RuneError should match any invalid UTF-8 byte sequence.
|
|
|
|
|
|
{"<22>", '<27>', 0},
|
|
|
|
|
|
{"\xff", '<27>', 0},
|
|
|
|
|
|
{"☻x<E298BB>", '<27>', len("☻x")},
|
|
|
|
|
|
{"☻x\xe2\x98", '<27>', len("☻x")},
|
|
|
|
|
|
{"☻x\xe2\x98<39>", '<27>', len("☻x")},
|
|
|
|
|
|
{"☻x\xe2\x98x", '<27>', len("☻x")},
|
|
|
|
|
|
|
|
|
|
|
|
// Invalid rune values should never match.
|
|
|
|
|
|
{"a☺b☻c☹d\xe2\x98<39>\xff<66>\xed\xa0\x80", -1, -1},
|
|
|
|
|
|
{"a☺b☻c☹d\xe2\x98<39>\xff<66>\xed\xa0\x80", 0xD800, -1}, // Surrogate pair
|
|
|
|
|
|
{"a☺b☻c☹d\xe2\x98<39>\xff<66>\xed\xa0\x80", utf8.MaxRune + 1, -1},
|
|
|
|
|
|
}
|
|
|
|
|
|
for _, tt := range tests {
|
|
|
|
|
|
if got := IndexRune([]byte(tt.in), tt.rune); got != tt.want {
|
|
|
|
|
|
t.Errorf("IndexRune(%q, %d) = %v; want %v", tt.in, tt.rune, got, tt.want)
|
2010-08-05 23:11:06 +10:00
|
|
|
|
}
|
|
|
|
|
|
}
|
2016-09-06 08:09:27 +09:00
|
|
|
|
|
|
|
|
|
|
haystack := []byte("test世界")
|
|
|
|
|
|
allocs := testing.AllocsPerRun(1000, func() {
|
|
|
|
|
|
if i := IndexRune(haystack, 's'); i != 2 {
|
|
|
|
|
|
t.Fatalf("'s' at %d; want 2", i)
|
|
|
|
|
|
}
|
|
|
|
|
|
if i := IndexRune(haystack, '世'); i != 4 {
|
|
|
|
|
|
t.Fatalf("'世' at %d; want 4", i)
|
|
|
|
|
|
}
|
|
|
|
|
|
})
|
|
|
|
|
|
if allocs != 0 {
|
2016-10-26 14:18:37 -07:00
|
|
|
|
t.Errorf("expected no allocations, got %f", allocs)
|
2016-09-06 08:09:27 +09:00
|
|
|
|
}
|
2010-08-05 23:11:06 +10:00
|
|
|
|
}
|
|
|
|
|
|
|
2017-03-19 12:18:08 +01:00
|
|
|
|
// test count of a single byte across page offsets
|
|
|
|
|
|
func TestCountByte(t *testing.T) {
|
|
|
|
|
|
b := make([]byte, 5015) // bigger than a page
|
|
|
|
|
|
windows := []int{1, 2, 3, 4, 15, 16, 17, 31, 32, 33, 63, 64, 65, 128}
|
|
|
|
|
|
testCountWindow := func(i, window int) {
|
|
|
|
|
|
for j := 0; j < window; j++ {
|
|
|
|
|
|
b[i+j] = byte(100)
|
|
|
|
|
|
p := Count(b[i:i+window], []byte{100})
|
|
|
|
|
|
if p != j+1 {
|
|
|
|
|
|
t.Errorf("TestCountByte.Count(%q, 100) = %d", b[i:i+window], p)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
maxWnd := windows[len(windows)-1]
|
|
|
|
|
|
|
|
|
|
|
|
for i := 0; i <= 2*maxWnd; i++ {
|
|
|
|
|
|
for _, window := range windows {
|
|
|
|
|
|
if window > len(b[i:]) {
|
|
|
|
|
|
window = len(b[i:])
|
|
|
|
|
|
}
|
|
|
|
|
|
testCountWindow(i, window)
|
|
|
|
|
|
for j := 0; j < window; j++ {
|
|
|
|
|
|
b[i+j] = byte(0)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
for i := 4096 - (maxWnd + 1); i < len(b); i++ {
|
|
|
|
|
|
for _, window := range windows {
|
|
|
|
|
|
if window > len(b[i:]) {
|
|
|
|
|
|
window = len(b[i:])
|
|
|
|
|
|
}
|
|
|
|
|
|
testCountWindow(i, window)
|
|
|
|
|
|
for j := 0; j < window; j++ {
|
|
|
|
|
|
b[i+j] = byte(0)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
// Make sure we don't count bytes outside our window
|
|
|
|
|
|
func TestCountByteNoMatch(t *testing.T) {
|
|
|
|
|
|
b := make([]byte, 5015)
|
|
|
|
|
|
windows := []int{1, 2, 3, 4, 15, 16, 17, 31, 32, 33, 63, 64, 65, 128}
|
|
|
|
|
|
for i := 0; i <= len(b); i++ {
|
|
|
|
|
|
for _, window := range windows {
|
|
|
|
|
|
if window > len(b[i:]) {
|
|
|
|
|
|
window = len(b[i:])
|
|
|
|
|
|
}
|
|
|
|
|
|
// Fill the window with non-match
|
|
|
|
|
|
for j := 0; j < window; j++ {
|
|
|
|
|
|
b[i+j] = byte(100)
|
|
|
|
|
|
}
|
|
|
|
|
|
// Try to find something that doesn't exist
|
|
|
|
|
|
p := Count(b[i:i+window], []byte{0})
|
|
|
|
|
|
if p != 0 {
|
|
|
|
|
|
t.Errorf("TestCountByteNoMatch(%q, 0) = %d", b[i:i+window], p)
|
|
|
|
|
|
}
|
|
|
|
|
|
for j := 0; j < window; j++ {
|
|
|
|
|
|
b[i+j] = byte(0)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2011-12-07 15:09:56 -05:00
|
|
|
|
var bmbuf []byte
|
2009-12-04 10:23:43 -08:00
|
|
|
|
|
2016-05-26 10:54:25 +02:00
|
|
|
|
func valName(x int) string {
|
|
|
|
|
|
if s := x >> 20; s<<20 == x {
|
|
|
|
|
|
return fmt.Sprintf("%dM", s)
|
|
|
|
|
|
}
|
|
|
|
|
|
if s := x >> 10; s<<10 == x {
|
|
|
|
|
|
return fmt.Sprintf("%dK", s)
|
2011-12-07 15:09:56 -05:00
|
|
|
|
}
|
2016-05-26 10:54:25 +02:00
|
|
|
|
return fmt.Sprint(x)
|
2009-12-04 10:23:43 -08:00
|
|
|
|
}
|
|
|
|
|
|
|
2016-05-26 10:54:25 +02:00
|
|
|
|
func benchBytes(b *testing.B, sizes []int, f func(b *testing.B, n int)) {
|
|
|
|
|
|
for _, n := range sizes {
|
2016-09-14 18:03:26 +00:00
|
|
|
|
if isRaceBuilder && n > 4<<10 {
|
|
|
|
|
|
continue
|
|
|
|
|
|
}
|
2016-05-26 10:54:25 +02:00
|
|
|
|
b.Run(valName(n), func(b *testing.B) {
|
|
|
|
|
|
if len(bmbuf) < n {
|
|
|
|
|
|
bmbuf = make([]byte, n)
|
|
|
|
|
|
}
|
|
|
|
|
|
b.SetBytes(int64(n))
|
|
|
|
|
|
f(b, n)
|
|
|
|
|
|
})
|
2011-12-07 15:09:56 -05:00
|
|
|
|
}
|
2009-12-04 10:23:43 -08:00
|
|
|
|
}
|
|
|
|
|
|
|
2016-05-26 10:54:25 +02:00
|
|
|
|
var indexSizes = []int{10, 32, 4 << 10, 4 << 20, 64 << 20}
|
2011-12-07 15:09:56 -05:00
|
|
|
|
|
2016-09-14 18:03:26 +00:00
|
|
|
|
var isRaceBuilder = strings.HasSuffix(testenv.Builder(), "-race")
|
|
|
|
|
|
|
2016-05-26 10:54:25 +02:00
|
|
|
|
func BenchmarkIndexByte(b *testing.B) {
|
|
|
|
|
|
benchBytes(b, indexSizes, bmIndexByte(IndexByte))
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func BenchmarkIndexBytePortable(b *testing.B) {
|
|
|
|
|
|
benchBytes(b, indexSizes, bmIndexByte(IndexBytePortable))
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func bmIndexByte(index func([]byte, byte) int) func(b *testing.B, n int) {
|
|
|
|
|
|
return func(b *testing.B, n int) {
|
|
|
|
|
|
buf := bmbuf[0:n]
|
|
|
|
|
|
buf[n-1] = 'x'
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
j := index(buf, 'x')
|
|
|
|
|
|
if j != n-1 {
|
|
|
|
|
|
b.Fatal("bad index", j)
|
|
|
|
|
|
}
|
2011-12-07 15:09:56 -05:00
|
|
|
|
}
|
2016-05-26 10:54:25 +02:00
|
|
|
|
buf[n-1] = '\x00'
|
2011-12-07 15:09:56 -05:00
|
|
|
|
}
|
2009-12-04 10:23:43 -08:00
|
|
|
|
}
|
|
|
|
|
|
|
2016-09-06 08:09:27 +09:00
|
|
|
|
func BenchmarkIndexRune(b *testing.B) {
|
|
|
|
|
|
benchBytes(b, indexSizes, bmIndexRune(IndexRune))
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func BenchmarkIndexRuneASCII(b *testing.B) {
|
|
|
|
|
|
benchBytes(b, indexSizes, bmIndexRuneASCII(IndexRune))
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func bmIndexRuneASCII(index func([]byte, rune) int) func(b *testing.B, n int) {
|
|
|
|
|
|
return func(b *testing.B, n int) {
|
|
|
|
|
|
buf := bmbuf[0:n]
|
|
|
|
|
|
buf[n-1] = 'x'
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
j := index(buf, 'x')
|
|
|
|
|
|
if j != n-1 {
|
|
|
|
|
|
b.Fatal("bad index", j)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
buf[n-1] = '\x00'
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func bmIndexRune(index func([]byte, rune) int) func(b *testing.B, n int) {
|
|
|
|
|
|
return func(b *testing.B, n int) {
|
|
|
|
|
|
buf := bmbuf[0:n]
|
|
|
|
|
|
utf8.EncodeRune(buf[n-3:], '世')
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
j := index(buf, '世')
|
|
|
|
|
|
if j != n-3 {
|
|
|
|
|
|
b.Fatal("bad index", j)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
buf[n-3] = '\x00'
|
|
|
|
|
|
buf[n-2] = '\x00'
|
|
|
|
|
|
buf[n-1] = '\x00'
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2016-05-26 10:54:25 +02:00
|
|
|
|
func BenchmarkEqual(b *testing.B) {
|
|
|
|
|
|
b.Run("0", func(b *testing.B) {
|
|
|
|
|
|
var buf [4]byte
|
|
|
|
|
|
buf1 := buf[0:0]
|
|
|
|
|
|
buf2 := buf[1:1]
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
eq := Equal(buf1, buf2)
|
|
|
|
|
|
if !eq {
|
|
|
|
|
|
b.Fatal("bad equal")
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
})
|
2009-12-04 10:23:43 -08:00
|
|
|
|
|
2016-05-26 10:54:25 +02:00
|
|
|
|
sizes := []int{1, 6, 9, 15, 16, 20, 32, 4 << 10, 4 << 20, 64 << 20}
|
|
|
|
|
|
benchBytes(b, sizes, bmEqual(Equal))
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func BenchmarkEqualPort(b *testing.B) {
|
|
|
|
|
|
sizes := []int{1, 6, 32, 4 << 10, 4 << 20, 64 << 20}
|
|
|
|
|
|
benchBytes(b, sizes, bmEqual(EqualPortable))
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func bmEqual(equal func([]byte, []byte) bool) func(b *testing.B, n int) {
|
|
|
|
|
|
return func(b *testing.B, n int) {
|
|
|
|
|
|
if len(bmbuf) < 2*n {
|
|
|
|
|
|
bmbuf = make([]byte, 2*n)
|
|
|
|
|
|
}
|
|
|
|
|
|
buf1 := bmbuf[0:n]
|
|
|
|
|
|
buf2 := bmbuf[n : 2*n]
|
|
|
|
|
|
buf1[n-1] = 'x'
|
|
|
|
|
|
buf2[n-1] = 'x'
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
eq := equal(buf1, buf2)
|
|
|
|
|
|
if !eq {
|
|
|
|
|
|
b.Fatal("bad equal")
|
|
|
|
|
|
}
|
2009-12-04 10:23:43 -08:00
|
|
|
|
}
|
2016-05-26 10:54:25 +02:00
|
|
|
|
buf1[n-1] = '\x00'
|
|
|
|
|
|
buf2[n-1] = '\x00'
|
2009-11-18 19:23:08 -08:00
|
|
|
|
}
|
2011-12-07 15:09:56 -05:00
|
|
|
|
}
|
|
|
|
|
|
|
2016-05-26 10:54:25 +02:00
|
|
|
|
func BenchmarkIndex(b *testing.B) {
|
|
|
|
|
|
benchBytes(b, indexSizes, func(b *testing.B, n int) {
|
|
|
|
|
|
buf := bmbuf[0:n]
|
|
|
|
|
|
buf[n-1] = 'x'
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
j := Index(buf, buf[n-7:])
|
|
|
|
|
|
if j != n-7 {
|
|
|
|
|
|
b.Fatal("bad index", j)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
buf[n-1] = '\x00'
|
|
|
|
|
|
})
|
|
|
|
|
|
}
|
2011-12-07 15:09:56 -05:00
|
|
|
|
|
2016-05-26 10:54:25 +02:00
|
|
|
|
func BenchmarkIndexEasy(b *testing.B) {
|
|
|
|
|
|
benchBytes(b, indexSizes, func(b *testing.B, n int) {
|
|
|
|
|
|
buf := bmbuf[0:n]
|
|
|
|
|
|
buf[n-1] = 'x'
|
|
|
|
|
|
buf[n-7] = 'x'
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
j := Index(buf, buf[n-7:])
|
|
|
|
|
|
if j != n-7 {
|
|
|
|
|
|
b.Fatal("bad index", j)
|
|
|
|
|
|
}
|
2011-12-07 15:09:56 -05:00
|
|
|
|
}
|
2016-05-26 10:54:25 +02:00
|
|
|
|
buf[n-1] = '\x00'
|
|
|
|
|
|
buf[n-7] = '\x00'
|
|
|
|
|
|
})
|
2011-12-07 15:09:56 -05:00
|
|
|
|
}
|
|
|
|
|
|
|
2016-05-26 10:54:25 +02:00
|
|
|
|
func BenchmarkCount(b *testing.B) {
|
|
|
|
|
|
benchBytes(b, indexSizes, func(b *testing.B, n int) {
|
|
|
|
|
|
buf := bmbuf[0:n]
|
|
|
|
|
|
buf[n-1] = 'x'
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
j := Count(buf, buf[n-7:])
|
|
|
|
|
|
if j != 1 {
|
|
|
|
|
|
b.Fatal("bad count", j)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
buf[n-1] = '\x00'
|
|
|
|
|
|
})
|
|
|
|
|
|
}
|
2011-12-07 15:09:56 -05:00
|
|
|
|
|
2016-05-26 10:54:25 +02:00
|
|
|
|
func BenchmarkCountEasy(b *testing.B) {
|
|
|
|
|
|
benchBytes(b, indexSizes, func(b *testing.B, n int) {
|
|
|
|
|
|
buf := bmbuf[0:n]
|
|
|
|
|
|
buf[n-1] = 'x'
|
|
|
|
|
|
buf[n-7] = 'x'
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
j := Count(buf, buf[n-7:])
|
|
|
|
|
|
if j != 1 {
|
|
|
|
|
|
b.Fatal("bad count", j)
|
|
|
|
|
|
}
|
2011-12-07 15:09:56 -05:00
|
|
|
|
}
|
2016-05-26 10:54:25 +02:00
|
|
|
|
buf[n-1] = '\x00'
|
|
|
|
|
|
buf[n-7] = '\x00'
|
|
|
|
|
|
})
|
2009-11-18 19:23:08 -08:00
|
|
|
|
}
|
2009-06-04 14:41:31 -07:00
|
|
|
|
|
2017-03-19 12:18:08 +01:00
|
|
|
|
func BenchmarkCountSingle(b *testing.B) {
|
|
|
|
|
|
benchBytes(b, indexSizes, func(b *testing.B, n int) {
|
|
|
|
|
|
buf := bmbuf[0:n]
|
|
|
|
|
|
step := 8
|
|
|
|
|
|
for i := 0; i < len(buf); i += step {
|
|
|
|
|
|
buf[i] = 1
|
|
|
|
|
|
}
|
|
|
|
|
|
expect := (len(buf) + (step - 1)) / step
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
j := Count(buf, []byte{1})
|
|
|
|
|
|
if j != expect {
|
|
|
|
|
|
b.Fatal("bad count", j, expect)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
for i := 0; i < len(buf); i++ {
|
|
|
|
|
|
buf[i] = 0
|
|
|
|
|
|
}
|
|
|
|
|
|
})
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2009-06-04 14:41:31 -07:00
|
|
|
|
type SplitTest struct {
|
2009-12-15 15:33:31 -08:00
|
|
|
|
s string
|
|
|
|
|
|
sep string
|
|
|
|
|
|
n int
|
|
|
|
|
|
a []string
|
2009-06-04 14:41:31 -07:00
|
|
|
|
}
|
2009-10-08 15:14:54 -07:00
|
|
|
|
|
|
|
|
|
|
var splittests = []SplitTest{
|
2019-04-02 22:38:39 +02:00
|
|
|
|
{"", "", -1, []string{}},
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{abcd, "a", 0, nil},
|
2019-04-02 22:38:39 +02:00
|
|
|
|
{abcd, "", 2, []string{"a", "bcd"}},
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{abcd, "a", -1, []string{"", "bcd"}},
|
|
|
|
|
|
{abcd, "z", -1, []string{"abcd"}},
|
|
|
|
|
|
{abcd, "", -1, []string{"a", "b", "c", "d"}},
|
|
|
|
|
|
{commas, ",", -1, []string{"1", "2", "3", "4"}},
|
|
|
|
|
|
{dots, "...", -1, []string{"1", ".2", ".3", ".4"}},
|
|
|
|
|
|
{faces, "☹", -1, []string{"☺☻", ""}},
|
|
|
|
|
|
{faces, "~", -1, []string{faces}},
|
|
|
|
|
|
{faces, "", -1, []string{"☺", "☻", "☹"}},
|
|
|
|
|
|
{"1 2 3 4", " ", 3, []string{"1", "2", "3 4"}},
|
|
|
|
|
|
{"1 2", " ", 3, []string{"1", "2"}},
|
|
|
|
|
|
{"123", "", 2, []string{"1", "23"}},
|
|
|
|
|
|
{"123", "", 17, []string{"1", "2", "3"}},
|
2009-06-04 14:41:31 -07:00
|
|
|
|
}
|
2009-09-13 21:35:18 -07:00
|
|
|
|
|
2009-06-04 14:41:31 -07:00
|
|
|
|
func TestSplit(t *testing.T) {
|
2009-06-24 19:02:29 -07:00
|
|
|
|
for _, tt := range splittests {
|
2011-06-28 09:43:14 +10:00
|
|
|
|
a := SplitN([]byte(tt.s), []byte(tt.sep), tt.n)
|
2017-10-30 16:05:18 -07:00
|
|
|
|
|
|
|
|
|
|
// Appending to the results should not change future results.
|
|
|
|
|
|
var x []byte
|
|
|
|
|
|
for _, v := range a {
|
|
|
|
|
|
x = append(v, 'z')
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2013-01-07 10:48:06 +11:00
|
|
|
|
result := sliceOfString(a)
|
2009-06-04 14:41:31 -07:00
|
|
|
|
if !eq(result, tt.a) {
|
2009-12-15 15:33:31 -08:00
|
|
|
|
t.Errorf(`Split(%q, %q, %d) = %v; want %v`, tt.s, tt.sep, tt.n, result, tt.a)
|
|
|
|
|
|
continue
|
2009-06-04 14:41:31 -07:00
|
|
|
|
}
|
2019-04-02 22:38:39 +02:00
|
|
|
|
if tt.n == 0 || len(a) == 0 {
|
2010-07-01 14:08:14 -07:00
|
|
|
|
continue
|
|
|
|
|
|
}
|
2017-10-30 16:05:18 -07:00
|
|
|
|
|
|
|
|
|
|
if want := tt.a[len(tt.a)-1] + "z"; string(x) != want {
|
|
|
|
|
|
t.Errorf("last appended result was %s; want %s", x, want)
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2010-02-25 16:01:29 -08:00
|
|
|
|
s := Join(a, []byte(tt.sep))
|
2009-06-04 14:41:31 -07:00
|
|
|
|
if string(s) != tt.s {
|
2009-11-09 12:07:39 -08:00
|
|
|
|
t.Errorf(`Join(Split(%q, %q, %d), %q) = %q`, tt.s, tt.sep, tt.n, tt.sep, s)
|
2009-06-04 14:41:31 -07:00
|
|
|
|
}
|
2011-06-28 09:43:14 +10:00
|
|
|
|
if tt.n < 0 {
|
|
|
|
|
|
b := Split([]byte(tt.s), []byte(tt.sep))
|
|
|
|
|
|
if !reflect.DeepEqual(a, b) {
|
|
|
|
|
|
t.Errorf("Split disagrees withSplitN(%q, %q, %d) = %v; want %v", tt.s, tt.sep, tt.n, b, a)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2012-07-20 16:04:22 -03:00
|
|
|
|
if len(a) > 0 {
|
|
|
|
|
|
in, out := a[0], s
|
|
|
|
|
|
if cap(in) == cap(out) && &in[:1][0] == &out[:1][0] {
|
|
|
|
|
|
t.Errorf("Join(%#v, %q) didn't copy", a, tt.sep)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2009-06-04 14:41:31 -07:00
|
|
|
|
}
|
|
|
|
|
|
}
|
2009-06-04 15:00:15 -07:00
|
|
|
|
|
2009-11-04 15:19:30 -08:00
|
|
|
|
var splitaftertests = []SplitTest{
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{abcd, "a", -1, []string{"a", "bcd"}},
|
|
|
|
|
|
{abcd, "z", -1, []string{"abcd"}},
|
|
|
|
|
|
{abcd, "", -1, []string{"a", "b", "c", "d"}},
|
|
|
|
|
|
{commas, ",", -1, []string{"1,", "2,", "3,", "4"}},
|
|
|
|
|
|
{dots, "...", -1, []string{"1...", ".2...", ".3...", ".4"}},
|
|
|
|
|
|
{faces, "☹", -1, []string{"☺☻☹", ""}},
|
|
|
|
|
|
{faces, "~", -1, []string{faces}},
|
|
|
|
|
|
{faces, "", -1, []string{"☺", "☻", "☹"}},
|
|
|
|
|
|
{"1 2 3 4", " ", 3, []string{"1 ", "2 ", "3 4"}},
|
|
|
|
|
|
{"1 2 3", " ", 3, []string{"1 ", "2 ", "3"}},
|
|
|
|
|
|
{"1 2", " ", 3, []string{"1 ", "2"}},
|
|
|
|
|
|
{"123", "", 2, []string{"1", "23"}},
|
|
|
|
|
|
{"123", "", 17, []string{"1", "2", "3"}},
|
2009-11-04 15:19:30 -08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestSplitAfter(t *testing.T) {
|
|
|
|
|
|
for _, tt := range splitaftertests {
|
2011-06-28 09:43:14 +10:00
|
|
|
|
a := SplitAfterN([]byte(tt.s), []byte(tt.sep), tt.n)
|
2017-10-30 16:05:18 -07:00
|
|
|
|
|
|
|
|
|
|
// Appending to the results should not change future results.
|
|
|
|
|
|
var x []byte
|
|
|
|
|
|
for _, v := range a {
|
|
|
|
|
|
x = append(v, 'z')
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2013-01-07 10:48:06 +11:00
|
|
|
|
result := sliceOfString(a)
|
2009-11-04 15:19:30 -08:00
|
|
|
|
if !eq(result, tt.a) {
|
2009-12-15 15:33:31 -08:00
|
|
|
|
t.Errorf(`Split(%q, %q, %d) = %v; want %v`, tt.s, tt.sep, tt.n, result, tt.a)
|
|
|
|
|
|
continue
|
2009-11-04 15:19:30 -08:00
|
|
|
|
}
|
2017-10-30 16:05:18 -07:00
|
|
|
|
|
|
|
|
|
|
if want := tt.a[len(tt.a)-1] + "z"; string(x) != want {
|
|
|
|
|
|
t.Errorf("last appended result was %s; want %s", x, want)
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2009-12-15 15:33:31 -08:00
|
|
|
|
s := Join(a, nil)
|
2009-11-04 15:19:30 -08:00
|
|
|
|
if string(s) != tt.s {
|
2009-11-09 12:07:39 -08:00
|
|
|
|
t.Errorf(`Join(Split(%q, %q, %d), %q) = %q`, tt.s, tt.sep, tt.n, tt.sep, s)
|
2009-11-04 15:19:30 -08:00
|
|
|
|
}
|
2011-06-28 09:43:14 +10:00
|
|
|
|
if tt.n < 0 {
|
|
|
|
|
|
b := SplitAfter([]byte(tt.s), []byte(tt.sep))
|
|
|
|
|
|
if !reflect.DeepEqual(a, b) {
|
|
|
|
|
|
t.Errorf("SplitAfter disagrees withSplitAfterN(%q, %q, %d) = %v; want %v", tt.s, tt.sep, tt.n, b, a)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2009-11-04 15:19:30 -08:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2009-12-15 21:09:55 -08:00
|
|
|
|
type FieldsTest struct {
|
|
|
|
|
|
s string
|
|
|
|
|
|
a []string
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
var fieldstests = []FieldsTest{
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{"", []string{}},
|
|
|
|
|
|
{" ", []string{}},
|
|
|
|
|
|
{" \t ", []string{}},
|
|
|
|
|
|
{" abc ", []string{"abc"}},
|
|
|
|
|
|
{"1 2 3 4", []string{"1", "2", "3", "4"}},
|
|
|
|
|
|
{"1 2 3 4", []string{"1", "2", "3", "4"}},
|
|
|
|
|
|
{"1\t\t2\t\t3\t4", []string{"1", "2", "3", "4"}},
|
|
|
|
|
|
{"1\u20002\u20013\u20024", []string{"1", "2", "3", "4"}},
|
|
|
|
|
|
{"\u2000\u2001\u2002", []string{}},
|
|
|
|
|
|
{"\n™\t™\n", []string{"™", "™"}},
|
|
|
|
|
|
{faces, []string{faces}},
|
2009-12-15 21:09:55 -08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestFields(t *testing.T) {
|
|
|
|
|
|
for _, tt := range fieldstests {
|
2017-10-30 16:05:18 -07:00
|
|
|
|
b := []byte(tt.s)
|
|
|
|
|
|
a := Fields(b)
|
|
|
|
|
|
|
|
|
|
|
|
// Appending to the results should not change future results.
|
|
|
|
|
|
var x []byte
|
|
|
|
|
|
for _, v := range a {
|
|
|
|
|
|
x = append(v, 'z')
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2013-01-07 10:48:06 +11:00
|
|
|
|
result := sliceOfString(a)
|
2009-12-15 21:09:55 -08:00
|
|
|
|
if !eq(result, tt.a) {
|
|
|
|
|
|
t.Errorf("Fields(%q) = %v; want %v", tt.s, a, tt.a)
|
|
|
|
|
|
continue
|
|
|
|
|
|
}
|
2017-10-30 16:05:18 -07:00
|
|
|
|
|
|
|
|
|
|
if string(b) != tt.s {
|
|
|
|
|
|
t.Errorf("slice changed to %s; want %s", string(b), tt.s)
|
|
|
|
|
|
}
|
|
|
|
|
|
if len(tt.a) > 0 {
|
|
|
|
|
|
if want := tt.a[len(tt.a)-1] + "z"; string(x) != want {
|
|
|
|
|
|
t.Errorf("last appended result was %s; want %s", x, want)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2009-12-15 21:09:55 -08:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2010-08-05 23:11:06 +10:00
|
|
|
|
func TestFieldsFunc(t *testing.T) {
|
2012-09-18 15:02:08 -04:00
|
|
|
|
for _, tt := range fieldstests {
|
|
|
|
|
|
a := FieldsFunc([]byte(tt.s), unicode.IsSpace)
|
2013-01-07 10:48:06 +11:00
|
|
|
|
result := sliceOfString(a)
|
2012-09-18 15:02:08 -04:00
|
|
|
|
if !eq(result, tt.a) {
|
|
|
|
|
|
t.Errorf("FieldsFunc(%q, unicode.IsSpace) = %v; want %v", tt.s, a, tt.a)
|
|
|
|
|
|
continue
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2011-10-25 22:22:09 -07:00
|
|
|
|
pred := func(c rune) bool { return c == 'X' }
|
2010-08-05 23:11:06 +10:00
|
|
|
|
var fieldsFuncTests = []FieldsTest{
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{"", []string{}},
|
|
|
|
|
|
{"XX", []string{}},
|
|
|
|
|
|
{"XXhiXXX", []string{"hi"}},
|
|
|
|
|
|
{"aXXbXXXcX", []string{"a", "b", "c"}},
|
2010-08-05 23:11:06 +10:00
|
|
|
|
}
|
|
|
|
|
|
for _, tt := range fieldsFuncTests {
|
2017-10-30 16:05:18 -07:00
|
|
|
|
b := []byte(tt.s)
|
|
|
|
|
|
a := FieldsFunc(b, pred)
|
|
|
|
|
|
|
|
|
|
|
|
// Appending to the results should not change future results.
|
|
|
|
|
|
var x []byte
|
|
|
|
|
|
for _, v := range a {
|
|
|
|
|
|
x = append(v, 'z')
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2013-01-07 10:48:06 +11:00
|
|
|
|
result := sliceOfString(a)
|
2010-08-05 23:11:06 +10:00
|
|
|
|
if !eq(result, tt.a) {
|
|
|
|
|
|
t.Errorf("FieldsFunc(%q) = %v, want %v", tt.s, a, tt.a)
|
|
|
|
|
|
}
|
2017-10-30 16:05:18 -07:00
|
|
|
|
|
|
|
|
|
|
if string(b) != tt.s {
|
|
|
|
|
|
t.Errorf("slice changed to %s; want %s", b, tt.s)
|
|
|
|
|
|
}
|
|
|
|
|
|
if len(tt.a) > 0 {
|
|
|
|
|
|
if want := tt.a[len(tt.a)-1] + "z"; string(x) != want {
|
|
|
|
|
|
t.Errorf("last appended result was %s; want %s", x, want)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2010-08-05 23:11:06 +10:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2013-01-07 10:48:06 +11:00
|
|
|
|
// Test case for any function which accepts and returns a byte slice.
|
2019-03-27 07:36:27 -04:00
|
|
|
|
// For ease of creation, we write the input byte slice as a string.
|
2009-09-01 13:46:59 -07:00
|
|
|
|
type StringTest struct {
|
2019-03-27 07:36:27 -04:00
|
|
|
|
in string
|
|
|
|
|
|
out []byte
|
2009-09-01 13:46:59 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
2009-10-08 15:14:54 -07:00
|
|
|
|
var upperTests = []StringTest{
|
2019-03-27 07:36:27 -04:00
|
|
|
|
{"", []byte("")},
|
2019-04-08 15:53:35 +02:00
|
|
|
|
{"ONLYUPPER", []byte("ONLYUPPER")},
|
2019-03-27 07:36:27 -04:00
|
|
|
|
{"abc", []byte("ABC")},
|
|
|
|
|
|
{"AbC123", []byte("ABC123")},
|
|
|
|
|
|
{"azAZ09_", []byte("AZAZ09_")},
|
2019-04-08 15:53:35 +02:00
|
|
|
|
{"longStrinGwitHmixofsmaLLandcAps", []byte("LONGSTRINGWITHMIXOFSMALLANDCAPS")},
|
|
|
|
|
|
{"long\u0250string\u0250with\u0250nonascii\u2C6Fchars", []byte("LONG\u2C6FSTRING\u2C6FWITH\u2C6FNONASCII\u2C6FCHARS")},
|
2019-03-27 07:36:27 -04:00
|
|
|
|
{"\u0250\u0250\u0250\u0250\u0250", []byte("\u2C6F\u2C6F\u2C6F\u2C6F\u2C6F")}, // grows one byte per char
|
2019-04-08 15:53:35 +02:00
|
|
|
|
{"a\u0080\U0010FFFF", []byte("A\u0080\U0010FFFF")}, // test utf8.RuneSelf and utf8.MaxRune
|
2009-09-01 13:46:59 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
2009-10-08 15:14:54 -07:00
|
|
|
|
var lowerTests = []StringTest{
|
2019-03-27 07:36:27 -04:00
|
|
|
|
{"", []byte("")},
|
|
|
|
|
|
{"abc", []byte("abc")},
|
|
|
|
|
|
{"AbC123", []byte("abc123")},
|
|
|
|
|
|
{"azAZ09_", []byte("azaz09_")},
|
2019-04-08 15:53:35 +02:00
|
|
|
|
{"longStrinGwitHmixofsmaLLandcAps", []byte("longstringwithmixofsmallandcaps")},
|
|
|
|
|
|
{"LONG\u2C6FSTRING\u2C6FWITH\u2C6FNONASCII\u2C6FCHARS", []byte("long\u0250string\u0250with\u0250nonascii\u0250chars")},
|
2019-03-27 07:36:27 -04:00
|
|
|
|
{"\u2C6D\u2C6D\u2C6D\u2C6D\u2C6D", []byte("\u0251\u0251\u0251\u0251\u0251")}, // shrinks one byte per char
|
2019-04-08 15:53:35 +02:00
|
|
|
|
{"A\u0080\U0010FFFF", []byte("a\u0080\U0010FFFF")}, // test utf8.RuneSelf and utf8.MaxRune
|
2009-09-01 13:46:59 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
const space = "\t\v\r\f\n\u0085\u00a0\u2000\u3000"
|
|
|
|
|
|
|
2009-10-08 15:14:54 -07:00
|
|
|
|
var trimSpaceTests = []StringTest{
|
2019-03-27 07:36:27 -04:00
|
|
|
|
{"", nil},
|
|
|
|
|
|
{" a", []byte("a")},
|
|
|
|
|
|
{"b ", []byte("b")},
|
|
|
|
|
|
{"abc", []byte("abc")},
|
|
|
|
|
|
{space + "abc" + space, []byte("abc")},
|
|
|
|
|
|
{" ", nil},
|
|
|
|
|
|
{"\u3000 ", nil},
|
|
|
|
|
|
{" \u3000", nil},
|
|
|
|
|
|
{" \t\r\n \t\t\r\r\n\n ", nil},
|
|
|
|
|
|
{" \t\r\n x\t\t\r\r\n\n ", []byte("x")},
|
|
|
|
|
|
{" \u2000\t\r\n x\t\t\r\r\ny\n \u3000", []byte("x\t\t\r\r\ny")},
|
|
|
|
|
|
{"1 \t\r\n2", []byte("1 \t\r\n2")},
|
|
|
|
|
|
{" x\x80", []byte("x\x80")},
|
|
|
|
|
|
{" x\xc0", []byte("x\xc0")},
|
|
|
|
|
|
{"x \xc0\xc0 ", []byte("x \xc0\xc0")},
|
|
|
|
|
|
{"x \xc0", []byte("x \xc0")},
|
|
|
|
|
|
{"x \xc0 ", []byte("x \xc0")},
|
|
|
|
|
|
{"x \xc0\xc0 ", []byte("x \xc0\xc0")},
|
|
|
|
|
|
{"x ☺\xc0\xc0 ", []byte("x ☺\xc0\xc0")},
|
|
|
|
|
|
{"x ☺ ", []byte("x ☺")},
|
2009-09-01 13:46:59 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
// Execute f on each test case. funcName should be the name of f; it's used
|
|
|
|
|
|
// in failure reports.
|
|
|
|
|
|
func runStringTests(t *testing.T, f func([]byte) []byte, funcName string, testCases []StringTest) {
|
2009-09-15 09:41:59 -07:00
|
|
|
|
for _, tc := range testCases {
|
2019-03-27 07:36:27 -04:00
|
|
|
|
actual := f([]byte(tc.in))
|
|
|
|
|
|
if actual == nil && tc.out != nil {
|
|
|
|
|
|
t.Errorf("%s(%q) = nil; want %q", funcName, tc.in, tc.out)
|
|
|
|
|
|
}
|
|
|
|
|
|
if actual != nil && tc.out == nil {
|
|
|
|
|
|
t.Errorf("%s(%q) = %q; want nil", funcName, tc.in, actual)
|
|
|
|
|
|
}
|
|
|
|
|
|
if !Equal(actual, tc.out) {
|
2009-11-09 12:07:39 -08:00
|
|
|
|
t.Errorf("%s(%q) = %q; want %q", funcName, tc.in, actual, tc.out)
|
2009-09-01 13:46:59 -07:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2011-10-25 22:22:09 -07:00
|
|
|
|
func tenRunes(r rune) string {
|
|
|
|
|
|
runes := make([]rune, 10)
|
|
|
|
|
|
for i := range runes {
|
|
|
|
|
|
runes[i] = r
|
2009-09-01 13:46:59 -07:00
|
|
|
|
}
|
2011-10-25 22:22:09 -07:00
|
|
|
|
return string(runes)
|
2009-09-01 13:46:59 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
2009-11-15 12:07:27 -08:00
|
|
|
|
// User-defined self-inverse mapping function
|
2011-10-25 22:22:09 -07:00
|
|
|
|
func rot13(r rune) rune {
|
|
|
|
|
|
const step = 13
|
|
|
|
|
|
if r >= 'a' && r <= 'z' {
|
|
|
|
|
|
return ((r - 'a' + step) % 26) + 'a'
|
2009-11-15 12:07:27 -08:00
|
|
|
|
}
|
2011-10-25 22:22:09 -07:00
|
|
|
|
if r >= 'A' && r <= 'Z' {
|
|
|
|
|
|
return ((r - 'A' + step) % 26) + 'A'
|
2009-11-15 12:07:27 -08:00
|
|
|
|
}
|
2011-10-25 22:22:09 -07:00
|
|
|
|
return r
|
2009-11-15 12:07:27 -08:00
|
|
|
|
}
|
|
|
|
|
|
|
2009-09-01 13:46:59 -07:00
|
|
|
|
func TestMap(t *testing.T) {
|
|
|
|
|
|
// Run a couple of awful growth/shrinkage tests
|
2009-12-15 15:33:31 -08:00
|
|
|
|
a := tenRunes('a')
|
2009-11-15 12:07:27 -08:00
|
|
|
|
|
2016-03-01 23:21:55 +00:00
|
|
|
|
// 1. Grow. This triggers two reallocations in Map.
|
2011-10-25 22:22:09 -07:00
|
|
|
|
maxRune := func(r rune) rune { return unicode.MaxRune }
|
2010-10-26 21:52:54 -07:00
|
|
|
|
m := Map(maxRune, []byte(a))
|
2009-12-15 15:33:31 -08:00
|
|
|
|
expect := tenRunes(unicode.MaxRune)
|
2009-09-01 13:46:59 -07:00
|
|
|
|
if string(m) != expect {
|
2009-11-09 12:07:39 -08:00
|
|
|
|
t.Errorf("growing: expected %q got %q", expect, m)
|
2009-09-01 13:46:59 -07:00
|
|
|
|
}
|
2009-11-15 12:07:27 -08:00
|
|
|
|
|
2009-09-01 13:46:59 -07:00
|
|
|
|
// 2. Shrink
|
2011-10-25 22:22:09 -07:00
|
|
|
|
minRune := func(r rune) rune { return 'a' }
|
2010-10-26 21:52:54 -07:00
|
|
|
|
m = Map(minRune, []byte(tenRunes(unicode.MaxRune)))
|
2009-12-15 15:33:31 -08:00
|
|
|
|
expect = a
|
2009-09-01 13:46:59 -07:00
|
|
|
|
if string(m) != expect {
|
2009-11-09 12:07:39 -08:00
|
|
|
|
t.Errorf("shrinking: expected %q got %q", expect, m)
|
2009-09-01 13:46:59 -07:00
|
|
|
|
}
|
2009-11-15 12:07:27 -08:00
|
|
|
|
|
|
|
|
|
|
// 3. Rot13
|
2010-10-26 21:52:54 -07:00
|
|
|
|
m = Map(rot13, []byte("a to zed"))
|
2009-12-15 15:33:31 -08:00
|
|
|
|
expect = "n gb mrq"
|
2009-11-15 12:07:27 -08:00
|
|
|
|
if string(m) != expect {
|
|
|
|
|
|
t.Errorf("rot13: expected %q got %q", expect, m)
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
// 4. Rot13^2
|
2010-10-26 21:52:54 -07:00
|
|
|
|
m = Map(rot13, Map(rot13, []byte("a to zed")))
|
2009-12-15 15:33:31 -08:00
|
|
|
|
expect = "a to zed"
|
2009-11-15 12:07:27 -08:00
|
|
|
|
if string(m) != expect {
|
|
|
|
|
|
t.Errorf("rot13: expected %q got %q", expect, m)
|
|
|
|
|
|
}
|
2009-12-11 10:37:48 -08:00
|
|
|
|
|
|
|
|
|
|
// 5. Drop
|
2011-10-25 22:22:09 -07:00
|
|
|
|
dropNotLatin := func(r rune) rune {
|
|
|
|
|
|
if unicode.Is(unicode.Latin, r) {
|
|
|
|
|
|
return r
|
2009-12-11 10:37:48 -08:00
|
|
|
|
}
|
2009-12-15 15:33:31 -08:00
|
|
|
|
return -1
|
|
|
|
|
|
}
|
2010-10-26 21:52:54 -07:00
|
|
|
|
m = Map(dropNotLatin, []byte("Hello, 세계"))
|
2009-12-15 15:33:31 -08:00
|
|
|
|
expect = "Hello"
|
2009-12-11 10:37:48 -08:00
|
|
|
|
if string(m) != expect {
|
|
|
|
|
|
t.Errorf("drop: expected %q got %q", expect, m)
|
|
|
|
|
|
}
|
2014-03-18 20:52:58 -07:00
|
|
|
|
|
|
|
|
|
|
// 6. Invalid rune
|
|
|
|
|
|
invalidRune := func(r rune) rune {
|
|
|
|
|
|
return utf8.MaxRune + 1
|
|
|
|
|
|
}
|
|
|
|
|
|
m = Map(invalidRune, []byte("x"))
|
|
|
|
|
|
expect = "\uFFFD"
|
|
|
|
|
|
if string(m) != expect {
|
|
|
|
|
|
t.Errorf("invalidRune: expected %q got %q", expect, m)
|
|
|
|
|
|
}
|
2009-09-01 13:46:59 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
2009-12-15 15:33:31 -08:00
|
|
|
|
func TestToUpper(t *testing.T) { runStringTests(t, ToUpper, "ToUpper", upperTests) }
|
2009-09-01 13:46:59 -07:00
|
|
|
|
|
2009-12-15 15:33:31 -08:00
|
|
|
|
func TestToLower(t *testing.T) { runStringTests(t, ToLower, "ToLower", lowerTests) }
|
2009-09-01 13:46:59 -07:00
|
|
|
|
|
2019-04-08 15:53:35 +02:00
|
|
|
|
func BenchmarkToUpper(b *testing.B) {
|
|
|
|
|
|
for _, tc := range upperTests {
|
|
|
|
|
|
tin := []byte(tc.in)
|
|
|
|
|
|
b.Run(tc.in, func(b *testing.B) {
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
actual := ToUpper(tin)
|
|
|
|
|
|
if !Equal(actual, tc.out) {
|
|
|
|
|
|
b.Errorf("ToUpper(%q) = %q; want %q", tc.in, actual, tc.out)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
})
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func BenchmarkToLower(b *testing.B) {
|
|
|
|
|
|
for _, tc := range lowerTests {
|
|
|
|
|
|
tin := []byte(tc.in)
|
|
|
|
|
|
b.Run(tc.in, func(b *testing.B) {
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
actual := ToLower(tin)
|
|
|
|
|
|
if !Equal(actual, tc.out) {
|
|
|
|
|
|
b.Errorf("ToLower(%q) = %q; want %q", tc.in, actual, tc.out)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
})
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2009-12-15 15:33:31 -08:00
|
|
|
|
func TestTrimSpace(t *testing.T) { runStringTests(t, TrimSpace, "TrimSpace", trimSpaceTests) }
|
2009-09-13 21:35:18 -07:00
|
|
|
|
|
2009-11-16 12:40:01 -08:00
|
|
|
|
type RepeatTest struct {
|
2009-12-15 15:33:31 -08:00
|
|
|
|
in, out string
|
|
|
|
|
|
count int
|
2009-11-16 12:40:01 -08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
var RepeatTests = []RepeatTest{
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{"", "", 0},
|
|
|
|
|
|
{"", "", 1},
|
|
|
|
|
|
{"", "", 2},
|
|
|
|
|
|
{"-", "", 0},
|
|
|
|
|
|
{"-", "-", 1},
|
|
|
|
|
|
{"-", "----------", 10},
|
|
|
|
|
|
{"abc ", "abc abc abc ", 3},
|
2009-11-16 12:40:01 -08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestRepeat(t *testing.T) {
|
|
|
|
|
|
for _, tt := range RepeatTests {
|
2010-02-25 16:01:29 -08:00
|
|
|
|
tin := []byte(tt.in)
|
|
|
|
|
|
tout := []byte(tt.out)
|
2009-12-15 15:33:31 -08:00
|
|
|
|
a := Repeat(tin, tt.count)
|
2009-11-16 12:40:01 -08:00
|
|
|
|
if !Equal(a, tout) {
|
2009-12-15 15:33:31 -08:00
|
|
|
|
t.Errorf("Repeat(%q, %d) = %q; want %q", tin, tt.count, a, tout)
|
|
|
|
|
|
continue
|
2009-11-16 12:40:01 -08:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2009-12-02 20:47:38 -08:00
|
|
|
|
|
2016-09-28 01:54:38 -07:00
|
|
|
|
func repeat(b []byte, count int) (err error) {
|
|
|
|
|
|
defer func() {
|
|
|
|
|
|
if r := recover(); r != nil {
|
|
|
|
|
|
switch v := r.(type) {
|
|
|
|
|
|
case error:
|
|
|
|
|
|
err = v
|
|
|
|
|
|
default:
|
|
|
|
|
|
err = fmt.Errorf("%s", v)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}()
|
|
|
|
|
|
|
|
|
|
|
|
Repeat(b, count)
|
|
|
|
|
|
|
|
|
|
|
|
return
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
// See Issue golang.org/issue/16237
|
|
|
|
|
|
func TestRepeatCatchesOverflow(t *testing.T) {
|
|
|
|
|
|
tests := [...]struct {
|
|
|
|
|
|
s string
|
|
|
|
|
|
count int
|
|
|
|
|
|
errStr string
|
|
|
|
|
|
}{
|
|
|
|
|
|
0: {"--", -2147483647, "negative"},
|
|
|
|
|
|
1: {"", int(^uint(0) >> 1), ""},
|
|
|
|
|
|
2: {"-", 10, ""},
|
|
|
|
|
|
3: {"gopher", 0, ""},
|
|
|
|
|
|
4: {"-", -1, "negative"},
|
|
|
|
|
|
5: {"--", -102, "negative"},
|
|
|
|
|
|
6: {string(make([]byte, 255)), int((^uint(0))/255 + 1), "overflow"},
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
for i, tt := range tests {
|
|
|
|
|
|
err := repeat([]byte(tt.s), tt.count)
|
|
|
|
|
|
if tt.errStr == "" {
|
|
|
|
|
|
if err != nil {
|
|
|
|
|
|
t.Errorf("#%d panicked %v", i, err)
|
|
|
|
|
|
}
|
|
|
|
|
|
continue
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
if err == nil || !strings.Contains(err.Error(), tt.errStr) {
|
|
|
|
|
|
t.Errorf("#%d expected %q got %q", i, tt.errStr, err)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2011-10-25 22:22:09 -07:00
|
|
|
|
func runesEqual(a, b []rune) bool {
|
2009-12-02 20:47:38 -08:00
|
|
|
|
if len(a) != len(b) {
|
|
|
|
|
|
return false
|
|
|
|
|
|
}
|
|
|
|
|
|
for i, r := range a {
|
|
|
|
|
|
if r != b[i] {
|
|
|
|
|
|
return false
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2009-12-15 15:33:31 -08:00
|
|
|
|
return true
|
2009-12-02 20:47:38 -08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
type RunesTest struct {
|
2009-12-15 15:33:31 -08:00
|
|
|
|
in string
|
2011-10-25 22:22:09 -07:00
|
|
|
|
out []rune
|
2009-12-15 15:33:31 -08:00
|
|
|
|
lossy bool
|
2009-12-02 20:47:38 -08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
var RunesTests = []RunesTest{
|
2011-10-25 22:22:09 -07:00
|
|
|
|
{"", []rune{}, false},
|
|
|
|
|
|
{" ", []rune{32}, false},
|
|
|
|
|
|
{"ABC", []rune{65, 66, 67}, false},
|
|
|
|
|
|
{"abc", []rune{97, 98, 99}, false},
|
|
|
|
|
|
{"\u65e5\u672c\u8a9e", []rune{26085, 26412, 35486}, false},
|
|
|
|
|
|
{"ab\x80c", []rune{97, 98, 0xFFFD, 99}, true},
|
|
|
|
|
|
{"ab\xc0c", []rune{97, 98, 0xFFFD, 99}, true},
|
2009-12-02 20:47:38 -08:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestRunes(t *testing.T) {
|
|
|
|
|
|
for _, tt := range RunesTests {
|
2010-02-25 16:01:29 -08:00
|
|
|
|
tin := []byte(tt.in)
|
2009-12-15 15:33:31 -08:00
|
|
|
|
a := Runes(tin)
|
2009-12-02 20:47:38 -08:00
|
|
|
|
if !runesEqual(a, tt.out) {
|
2009-12-15 15:33:31 -08:00
|
|
|
|
t.Errorf("Runes(%q) = %v; want %v", tin, a, tt.out)
|
|
|
|
|
|
continue
|
2009-12-02 20:47:38 -08:00
|
|
|
|
}
|
|
|
|
|
|
if !tt.lossy {
|
|
|
|
|
|
// can only test reassembly if we didn't lose information
|
2009-12-15 15:33:31 -08:00
|
|
|
|
s := string(a)
|
2009-12-02 20:47:38 -08:00
|
|
|
|
if s != tt.in {
|
|
|
|
|
|
t.Errorf("string(Runes(%q)) = %x; want %x", tin, s, tin)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2010-05-18 23:01:05 -07:00
|
|
|
|
|
|
|
|
|
|
type TrimTest struct {
|
2013-02-01 08:41:25 -08:00
|
|
|
|
f string
|
|
|
|
|
|
in, arg, out string
|
2010-05-18 23:01:05 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
var trimTests = []TrimTest{
|
2011-11-13 22:57:19 -05:00
|
|
|
|
{"Trim", "abba", "a", "bb"},
|
|
|
|
|
|
{"Trim", "abba", "ab", ""},
|
|
|
|
|
|
{"TrimLeft", "abba", "ab", ""},
|
|
|
|
|
|
{"TrimRight", "abba", "ab", ""},
|
|
|
|
|
|
{"TrimLeft", "abba", "a", "bba"},
|
|
|
|
|
|
{"TrimRight", "abba", "a", "abb"},
|
|
|
|
|
|
{"Trim", "<tag>", "<>", "tag"},
|
|
|
|
|
|
{"Trim", "* listitem", " *", "listitem"},
|
|
|
|
|
|
{"Trim", `"quote"`, `"`, "quote"},
|
|
|
|
|
|
{"Trim", "\u2C6F\u2C6F\u0250\u0250\u2C6F\u2C6F", "\u2C6F", "\u0250\u0250"},
|
bytes, strings: optimize for ASCII sets
In a large codebase within Google, there are thousands of uses of:
ContainsAny|IndexAny|LastIndexAny|Trim|TrimLeft|TrimRight
An analysis of their usage shows that over 97% of them only use character
sets consisting of only ASCII symbols.
Uses of ContainsAny|IndexAny|LastIndexAny:
6% are 1 character (e.g., "\n" or " ")
58% are 2-4 characters (e.g., "<>" or "\r\n\t ")
24% are 5-9 characters (e.g., "()[]*^$")
10% are 10+ characters (e.g., "+-=&|><!(){}[]^\"~*?:\\/ ")
We optimize for ASCII sets, which are commonly used to search for
"control" characters in some string. We don't optimize for the
single character scenario since IndexRune or IndexByte could be used.
Uses of Trim|TrimLeft|TrimRight:
71% are 1 character (e.g., "\n" or " ")
14% are 2 characters (e.g., "\r\n")
10% are 3-4 characters (e.g., " \t\r\n")
5% are 10+ characters (e.g., "0123456789abcdefABCDEF")
We optimize for the single character case with a simple closured function
that only checks for that character's value. We optimize for the medium
and larger sets using a 16-byte bit-map representing a set of ASCII characters.
The benchmarks below have the following suffix name "%d:%d" where the first
number is the length of the input and the second number is the length
of the charset.
== bytes package ==
benchmark old ns/op new ns/op delta
BenchmarkIndexAnyASCII/1:1-4 5.09 5.23 +2.75%
BenchmarkIndexAnyASCII/1:2-4 5.81 5.85 +0.69%
BenchmarkIndexAnyASCII/1:4-4 7.22 7.50 +3.88%
BenchmarkIndexAnyASCII/1:8-4 11.0 11.1 +0.91%
BenchmarkIndexAnyASCII/1:16-4 17.5 17.8 +1.71%
BenchmarkIndexAnyASCII/16:1-4 36.0 34.0 -5.56%
BenchmarkIndexAnyASCII/16:2-4 46.6 36.5 -21.67%
BenchmarkIndexAnyASCII/16:4-4 78.0 40.4 -48.21%
BenchmarkIndexAnyASCII/16:8-4 136 47.4 -65.15%
BenchmarkIndexAnyASCII/16:16-4 254 61.5 -75.79%
BenchmarkIndexAnyASCII/256:1-4 542 388 -28.41%
BenchmarkIndexAnyASCII/256:2-4 705 382 -45.82%
BenchmarkIndexAnyASCII/256:4-4 1089 386 -64.55%
BenchmarkIndexAnyASCII/256:8-4 1994 394 -80.24%
BenchmarkIndexAnyASCII/256:16-4 3843 411 -89.31%
BenchmarkIndexAnyASCII/4096:1-4 8522 5873 -31.08%
BenchmarkIndexAnyASCII/4096:2-4 11253 5861 -47.92%
BenchmarkIndexAnyASCII/4096:4-4 17824 5883 -66.99%
BenchmarkIndexAnyASCII/4096:8-4 32053 5871 -81.68%
BenchmarkIndexAnyASCII/4096:16-4 60512 5888 -90.27%
BenchmarkTrimASCII/1:1-4 79.5 70.8 -10.94%
BenchmarkTrimASCII/1:2-4 79.0 105 +32.91%
BenchmarkTrimASCII/1:4-4 79.6 109 +36.93%
BenchmarkTrimASCII/1:8-4 78.8 118 +49.75%
BenchmarkTrimASCII/1:16-4 80.2 132 +64.59%
BenchmarkTrimASCII/16:1-4 243 116 -52.26%
BenchmarkTrimASCII/16:2-4 243 171 -29.63%
BenchmarkTrimASCII/16:4-4 243 176 -27.57%
BenchmarkTrimASCII/16:8-4 241 184 -23.65%
BenchmarkTrimASCII/16:16-4 238 199 -16.39%
BenchmarkTrimASCII/256:1-4 2580 840 -67.44%
BenchmarkTrimASCII/256:2-4 2603 1175 -54.86%
BenchmarkTrimASCII/256:4-4 2572 1188 -53.81%
BenchmarkTrimASCII/256:8-4 2550 1191 -53.29%
BenchmarkTrimASCII/256:16-4 2585 1208 -53.27%
BenchmarkTrimASCII/4096:1-4 39773 12181 -69.37%
BenchmarkTrimASCII/4096:2-4 39946 17231 -56.86%
BenchmarkTrimASCII/4096:4-4 39641 17179 -56.66%
BenchmarkTrimASCII/4096:8-4 39835 17175 -56.88%
BenchmarkTrimASCII/4096:16-4 40229 17215 -57.21%
== strings package ==
benchmark old ns/op new ns/op delta
BenchmarkIndexAnyASCII/1:1-4 5.94 4.97 -16.33%
BenchmarkIndexAnyASCII/1:2-4 5.94 5.55 -6.57%
BenchmarkIndexAnyASCII/1:4-4 7.45 7.21 -3.22%
BenchmarkIndexAnyASCII/1:8-4 10.8 10.6 -1.85%
BenchmarkIndexAnyASCII/1:16-4 17.4 17.2 -1.15%
BenchmarkIndexAnyASCII/16:1-4 36.4 32.2 -11.54%
BenchmarkIndexAnyASCII/16:2-4 49.6 34.6 -30.24%
BenchmarkIndexAnyASCII/16:4-4 77.5 37.9 -51.10%
BenchmarkIndexAnyASCII/16:8-4 138 45.5 -67.03%
BenchmarkIndexAnyASCII/16:16-4 241 59.1 -75.48%
BenchmarkIndexAnyASCII/256:1-4 509 378 -25.74%
BenchmarkIndexAnyASCII/256:2-4 720 381 -47.08%
BenchmarkIndexAnyASCII/256:4-4 1142 384 -66.37%
BenchmarkIndexAnyASCII/256:8-4 1999 391 -80.44%
BenchmarkIndexAnyASCII/256:16-4 3735 403 -89.21%
BenchmarkIndexAnyASCII/4096:1-4 7973 5824 -26.95%
BenchmarkIndexAnyASCII/4096:2-4 11432 5809 -49.19%
BenchmarkIndexAnyASCII/4096:4-4 18327 5819 -68.25%
BenchmarkIndexAnyASCII/4096:8-4 33059 5828 -82.37%
BenchmarkIndexAnyASCII/4096:16-4 59703 5817 -90.26%
BenchmarkTrimASCII/1:1-4 71.9 71.8 -0.14%
BenchmarkTrimASCII/1:2-4 73.3 103 +40.52%
BenchmarkTrimASCII/1:4-4 71.8 106 +47.63%
BenchmarkTrimASCII/1:8-4 71.2 113 +58.71%
BenchmarkTrimASCII/1:16-4 71.6 128 +78.77%
BenchmarkTrimASCII/16:1-4 152 116 -23.68%
BenchmarkTrimASCII/16:2-4 160 168 +5.00%
BenchmarkTrimASCII/16:4-4 172 170 -1.16%
BenchmarkTrimASCII/16:8-4 200 177 -11.50%
BenchmarkTrimASCII/16:16-4 254 193 -24.02%
BenchmarkTrimASCII/256:1-4 1438 864 -39.92%
BenchmarkTrimASCII/256:2-4 1551 1195 -22.95%
BenchmarkTrimASCII/256:4-4 1770 1200 -32.20%
BenchmarkTrimASCII/256:8-4 2195 1216 -44.60%
BenchmarkTrimASCII/256:16-4 3054 1224 -59.92%
BenchmarkTrimASCII/4096:1-4 21726 12557 -42.20%
BenchmarkTrimASCII/4096:2-4 23586 17508 -25.77%
BenchmarkTrimASCII/4096:4-4 26898 17510 -34.90%
BenchmarkTrimASCII/4096:8-4 33714 17595 -47.81%
BenchmarkTrimASCII/4096:16-4 47429 17700 -62.68%
The benchmarks added test the worst case. For IndexAny, that is when the
charset matches none of the input. For Trim, it is when the charset matches
all of the input.
Change-Id: I970874d101a96b33528fc99b165379abe58cf6ea
Reviewed-on: https://go-review.googlesource.com/31593
Run-TryBot: Joe Tsai <thebrokentoaster@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Martin Möhrmann <martisch@uos.de>
2016-10-20 03:16:22 -07:00
|
|
|
|
{"Trim", "\x80test\xff", "\xff", "test"},
|
|
|
|
|
|
{"Trim", " Ġ ", " ", "Ġ"},
|
|
|
|
|
|
{"Trim", " Ġİ0", "0 ", "Ġİ"},
|
2010-05-18 23:01:05 -07:00
|
|
|
|
//empty string tests
|
2011-11-13 22:57:19 -05:00
|
|
|
|
{"Trim", "abba", "", "abba"},
|
|
|
|
|
|
{"Trim", "", "123", ""},
|
|
|
|
|
|
{"Trim", "", "", ""},
|
|
|
|
|
|
{"TrimLeft", "abba", "", "abba"},
|
|
|
|
|
|
{"TrimLeft", "", "123", ""},
|
|
|
|
|
|
{"TrimLeft", "", "", ""},
|
|
|
|
|
|
{"TrimRight", "abba", "", "abba"},
|
|
|
|
|
|
{"TrimRight", "", "123", ""},
|
|
|
|
|
|
{"TrimRight", "", "", ""},
|
|
|
|
|
|
{"TrimRight", "☺\xc0", "☺", "☺\xc0"},
|
2013-02-01 08:41:25 -08:00
|
|
|
|
{"TrimPrefix", "aabb", "a", "abb"},
|
|
|
|
|
|
{"TrimPrefix", "aabb", "b", "aabb"},
|
|
|
|
|
|
{"TrimSuffix", "aabb", "a", "aabb"},
|
|
|
|
|
|
{"TrimSuffix", "aabb", "b", "aab"},
|
2010-05-18 23:01:05 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestTrim(t *testing.T) {
|
|
|
|
|
|
for _, tc := range trimTests {
|
2011-11-13 22:57:19 -05:00
|
|
|
|
name := tc.f
|
|
|
|
|
|
var f func([]byte, string) []byte
|
2013-02-01 08:41:25 -08:00
|
|
|
|
var fb func([]byte, []byte) []byte
|
2011-11-13 22:57:19 -05:00
|
|
|
|
switch name {
|
|
|
|
|
|
case "Trim":
|
|
|
|
|
|
f = Trim
|
|
|
|
|
|
case "TrimLeft":
|
|
|
|
|
|
f = TrimLeft
|
|
|
|
|
|
case "TrimRight":
|
|
|
|
|
|
f = TrimRight
|
2013-02-01 08:41:25 -08:00
|
|
|
|
case "TrimPrefix":
|
|
|
|
|
|
fb = TrimPrefix
|
|
|
|
|
|
case "TrimSuffix":
|
|
|
|
|
|
fb = TrimSuffix
|
2010-05-18 23:01:05 -07:00
|
|
|
|
default:
|
2011-11-28 09:50:51 -08:00
|
|
|
|
t.Errorf("Undefined trim function %s", name)
|
2010-05-18 23:01:05 -07:00
|
|
|
|
}
|
2013-02-01 08:41:25 -08:00
|
|
|
|
var actual string
|
|
|
|
|
|
if f != nil {
|
|
|
|
|
|
actual = string(f([]byte(tc.in), tc.arg))
|
|
|
|
|
|
} else {
|
|
|
|
|
|
actual = string(fb([]byte(tc.in), []byte(tc.arg)))
|
|
|
|
|
|
}
|
2010-05-18 23:01:05 -07:00
|
|
|
|
if actual != tc.out {
|
2013-02-01 08:41:25 -08:00
|
|
|
|
t.Errorf("%s(%q, %q) = %q; want %q", name, tc.in, tc.arg, actual, tc.out)
|
2010-05-18 23:01:05 -07:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2010-07-23 12:34:35 -07:00
|
|
|
|
type predicate struct {
|
2011-10-25 22:22:09 -07:00
|
|
|
|
f func(r rune) bool
|
2010-07-23 12:34:35 -07:00
|
|
|
|
name string
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
var isSpace = predicate{unicode.IsSpace, "IsSpace"}
|
|
|
|
|
|
var isDigit = predicate{unicode.IsDigit, "IsDigit"}
|
|
|
|
|
|
var isUpper = predicate{unicode.IsUpper, "IsUpper"}
|
|
|
|
|
|
var isValidRune = predicate{
|
2011-10-25 22:22:09 -07:00
|
|
|
|
func(r rune) bool {
|
2010-07-23 12:34:35 -07:00
|
|
|
|
return r != utf8.RuneError
|
|
|
|
|
|
},
|
|
|
|
|
|
"IsValidRune",
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2010-05-18 23:01:05 -07:00
|
|
|
|
type TrimFuncTest struct {
|
2019-03-28 17:49:43 -04:00
|
|
|
|
f predicate
|
|
|
|
|
|
in string
|
|
|
|
|
|
trimOut []byte
|
|
|
|
|
|
leftOut []byte
|
|
|
|
|
|
rightOut []byte
|
2010-07-23 12:34:35 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func not(p predicate) predicate {
|
|
|
|
|
|
return predicate{
|
2011-10-25 22:22:09 -07:00
|
|
|
|
func(r rune) bool {
|
2010-07-23 12:34:35 -07:00
|
|
|
|
return !p.f(r)
|
|
|
|
|
|
},
|
|
|
|
|
|
"not " + p.name,
|
|
|
|
|
|
}
|
2010-05-18 23:01:05 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
var trimFuncTests = []TrimFuncTest{
|
2019-03-28 17:49:43 -04:00
|
|
|
|
{isSpace, space + " hello " + space,
|
|
|
|
|
|
[]byte("hello"),
|
|
|
|
|
|
[]byte("hello " + space),
|
|
|
|
|
|
[]byte(space + " hello")},
|
|
|
|
|
|
{isDigit, "\u0e50\u0e5212hello34\u0e50\u0e51",
|
|
|
|
|
|
[]byte("hello"),
|
|
|
|
|
|
[]byte("hello34\u0e50\u0e51"),
|
|
|
|
|
|
[]byte("\u0e50\u0e5212hello")},
|
|
|
|
|
|
{isUpper, "\u2C6F\u2C6F\u2C6F\u2C6FABCDhelloEF\u2C6F\u2C6FGH\u2C6F\u2C6F",
|
|
|
|
|
|
[]byte("hello"),
|
|
|
|
|
|
[]byte("helloEF\u2C6F\u2C6FGH\u2C6F\u2C6F"),
|
|
|
|
|
|
[]byte("\u2C6F\u2C6F\u2C6F\u2C6FABCDhello")},
|
|
|
|
|
|
{not(isSpace), "hello" + space + "hello",
|
|
|
|
|
|
[]byte(space),
|
|
|
|
|
|
[]byte(space + "hello"),
|
|
|
|
|
|
[]byte("hello" + space)},
|
|
|
|
|
|
{not(isDigit), "hello\u0e50\u0e521234\u0e50\u0e51helo",
|
|
|
|
|
|
[]byte("\u0e50\u0e521234\u0e50\u0e51"),
|
|
|
|
|
|
[]byte("\u0e50\u0e521234\u0e50\u0e51helo"),
|
|
|
|
|
|
[]byte("hello\u0e50\u0e521234\u0e50\u0e51")},
|
|
|
|
|
|
{isValidRune, "ab\xc0a\xc0cd",
|
|
|
|
|
|
[]byte("\xc0a\xc0"),
|
|
|
|
|
|
[]byte("\xc0a\xc0cd"),
|
|
|
|
|
|
[]byte("ab\xc0a\xc0")},
|
|
|
|
|
|
{not(isValidRune), "\xc0a\xc0",
|
|
|
|
|
|
[]byte("a"),
|
|
|
|
|
|
[]byte("a\xc0"),
|
|
|
|
|
|
[]byte("\xc0a")},
|
|
|
|
|
|
// The nils returned by TrimLeftFunc are odd behavior, but we need
|
|
|
|
|
|
// to preserve backwards compatibility.
|
|
|
|
|
|
{isSpace, "",
|
|
|
|
|
|
nil,
|
|
|
|
|
|
nil,
|
|
|
|
|
|
[]byte("")},
|
|
|
|
|
|
{isSpace, " ",
|
|
|
|
|
|
nil,
|
|
|
|
|
|
nil,
|
|
|
|
|
|
[]byte("")},
|
2010-05-18 23:01:05 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestTrimFunc(t *testing.T) {
|
|
|
|
|
|
for _, tc := range trimFuncTests {
|
2019-03-28 17:49:43 -04:00
|
|
|
|
trimmers := []struct {
|
|
|
|
|
|
name string
|
|
|
|
|
|
trim func(s []byte, f func(r rune) bool) []byte
|
|
|
|
|
|
out []byte
|
|
|
|
|
|
}{
|
|
|
|
|
|
{"TrimFunc", TrimFunc, tc.trimOut},
|
|
|
|
|
|
{"TrimLeftFunc", TrimLeftFunc, tc.leftOut},
|
|
|
|
|
|
{"TrimRightFunc", TrimRightFunc, tc.rightOut},
|
|
|
|
|
|
}
|
|
|
|
|
|
for _, trimmer := range trimmers {
|
|
|
|
|
|
actual := trimmer.trim([]byte(tc.in), tc.f.f)
|
|
|
|
|
|
if actual == nil && trimmer.out != nil {
|
|
|
|
|
|
t.Errorf("%s(%q, %q) = nil; want %q", trimmer.name, tc.in, tc.f.name, trimmer.out)
|
|
|
|
|
|
}
|
|
|
|
|
|
if actual != nil && trimmer.out == nil {
|
|
|
|
|
|
t.Errorf("%s(%q, %q) = %q; want nil", trimmer.name, tc.in, tc.f.name, actual)
|
|
|
|
|
|
}
|
|
|
|
|
|
if !Equal(actual, trimmer.out) {
|
|
|
|
|
|
t.Errorf("%s(%q, %q) = %q; want %q", trimmer.name, tc.in, tc.f.name, actual, trimmer.out)
|
|
|
|
|
|
}
|
2010-07-23 12:34:35 -07:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
type IndexFuncTest struct {
|
|
|
|
|
|
in string
|
|
|
|
|
|
f predicate
|
|
|
|
|
|
first, last int
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
var indexFuncTests = []IndexFuncTest{
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{"", isValidRune, -1, -1},
|
|
|
|
|
|
{"abc", isDigit, -1, -1},
|
|
|
|
|
|
{"0123", isDigit, 0, 3},
|
|
|
|
|
|
{"a1b", isDigit, 1, 1},
|
|
|
|
|
|
{space, isSpace, 0, len(space) - 3}, // last rune in space is 3 bytes
|
|
|
|
|
|
{"\u0e50\u0e5212hello34\u0e50\u0e51", isDigit, 0, 18},
|
|
|
|
|
|
{"\u2C6F\u2C6F\u2C6F\u2C6FABCDhelloEF\u2C6F\u2C6FGH\u2C6F\u2C6F", isUpper, 0, 34},
|
|
|
|
|
|
{"12\u0e50\u0e52hello34\u0e50\u0e51", not(isDigit), 8, 12},
|
2010-07-23 12:34:35 -07:00
|
|
|
|
|
|
|
|
|
|
// tests of invalid UTF-8
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{"\x801", isDigit, 1, 1},
|
|
|
|
|
|
{"\x80abc", isDigit, -1, -1},
|
|
|
|
|
|
{"\xc0a\xc0", isValidRune, 1, 1},
|
|
|
|
|
|
{"\xc0a\xc0", not(isValidRune), 0, 2},
|
|
|
|
|
|
{"\xc0☺\xc0", not(isValidRune), 0, 4},
|
|
|
|
|
|
{"\xc0☺\xc0\xc0", not(isValidRune), 0, 5},
|
|
|
|
|
|
{"ab\xc0a\xc0cd", not(isValidRune), 2, 4},
|
|
|
|
|
|
{"a\xe0\x80cd", not(isValidRune), 1, 2},
|
2010-07-23 12:34:35 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestIndexFunc(t *testing.T) {
|
|
|
|
|
|
for _, tc := range indexFuncTests {
|
|
|
|
|
|
first := IndexFunc([]byte(tc.in), tc.f.f)
|
|
|
|
|
|
if first != tc.first {
|
|
|
|
|
|
t.Errorf("IndexFunc(%q, %s) = %d; want %d", tc.in, tc.f.name, first, tc.first)
|
|
|
|
|
|
}
|
|
|
|
|
|
last := LastIndexFunc([]byte(tc.in), tc.f.f)
|
|
|
|
|
|
if last != tc.last {
|
|
|
|
|
|
t.Errorf("LastIndexFunc(%q, %s) = %d; want %d", tc.in, tc.f.name, last, tc.last)
|
2010-05-18 23:01:05 -07:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2010-06-30 18:03:09 -07:00
|
|
|
|
|
|
|
|
|
|
type ReplaceTest struct {
|
|
|
|
|
|
in string
|
|
|
|
|
|
old, new string
|
|
|
|
|
|
n int
|
|
|
|
|
|
out string
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
var ReplaceTests = []ReplaceTest{
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{"hello", "l", "L", 0, "hello"},
|
|
|
|
|
|
{"hello", "l", "L", -1, "heLLo"},
|
|
|
|
|
|
{"hello", "x", "X", -1, "hello"},
|
|
|
|
|
|
{"", "x", "X", -1, ""},
|
|
|
|
|
|
{"radar", "r", "<r>", -1, "<r>ada<r>"},
|
|
|
|
|
|
{"", "", "<>", -1, "<>"},
|
|
|
|
|
|
{"banana", "a", "<>", -1, "b<>n<>n<>"},
|
|
|
|
|
|
{"banana", "a", "<>", 1, "b<>nana"},
|
|
|
|
|
|
{"banana", "a", "<>", 1000, "b<>n<>n<>"},
|
|
|
|
|
|
{"banana", "an", "<>", -1, "b<><>a"},
|
|
|
|
|
|
{"banana", "ana", "<>", -1, "b<>na"},
|
|
|
|
|
|
{"banana", "", "<>", -1, "<>b<>a<>n<>a<>n<>a<>"},
|
|
|
|
|
|
{"banana", "", "<>", 10, "<>b<>a<>n<>a<>n<>a<>"},
|
|
|
|
|
|
{"banana", "", "<>", 6, "<>b<>a<>n<>a<>n<>a"},
|
|
|
|
|
|
{"banana", "", "<>", 5, "<>b<>a<>n<>a<>na"},
|
|
|
|
|
|
{"banana", "", "<>", 1, "<>banana"},
|
|
|
|
|
|
{"banana", "a", "a", -1, "banana"},
|
|
|
|
|
|
{"banana", "a", "a", 1, "banana"},
|
|
|
|
|
|
{"☺☻☹", "", "<>", -1, "<>☺<>☻<>☹<>"},
|
2010-06-30 18:03:09 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestReplace(t *testing.T) {
|
|
|
|
|
|
for _, tt := range ReplaceTests {
|
2011-10-12 13:42:04 -07:00
|
|
|
|
in := append([]byte(tt.in), "<spare>"...)
|
2011-09-21 12:36:17 -03:00
|
|
|
|
in = in[:len(tt.in)]
|
|
|
|
|
|
out := Replace(in, []byte(tt.old), []byte(tt.new), tt.n)
|
|
|
|
|
|
if s := string(out); s != tt.out {
|
2010-06-30 18:03:09 -07:00
|
|
|
|
t.Errorf("Replace(%q, %q, %q, %d) = %q, want %q", tt.in, tt.old, tt.new, tt.n, s, tt.out)
|
|
|
|
|
|
}
|
2011-09-21 12:36:17 -03:00
|
|
|
|
if cap(in) == cap(out) && &in[:1][0] == &out[:1][0] {
|
|
|
|
|
|
t.Errorf("Replace(%q, %q, %q, %d) didn't copy", tt.in, tt.old, tt.new, tt.n)
|
|
|
|
|
|
}
|
2018-09-26 20:19:11 +00:00
|
|
|
|
if tt.n == -1 {
|
|
|
|
|
|
out := ReplaceAll(in, []byte(tt.old), []byte(tt.new))
|
|
|
|
|
|
if s := string(out); s != tt.out {
|
|
|
|
|
|
t.Errorf("ReplaceAll(%q, %q, %q) = %q, want %q", tt.in, tt.old, tt.new, s, tt.out)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2010-06-30 18:03:09 -07:00
|
|
|
|
}
|
|
|
|
|
|
}
|
2010-07-20 19:53:59 -07:00
|
|
|
|
|
|
|
|
|
|
type TitleTest struct {
|
|
|
|
|
|
in, out string
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
var TitleTests = []TitleTest{
|
2010-10-22 10:06:33 -07:00
|
|
|
|
{"", ""},
|
|
|
|
|
|
{"a", "A"},
|
|
|
|
|
|
{" aaa aaa aaa ", " Aaa Aaa Aaa "},
|
|
|
|
|
|
{" Aaa Aaa Aaa ", " Aaa Aaa Aaa "},
|
|
|
|
|
|
{"123a456", "123a456"},
|
|
|
|
|
|
{"double-blind", "Double-Blind"},
|
|
|
|
|
|
{"ÿøû", "Ÿøû"},
|
2013-12-20 23:19:32 -08:00
|
|
|
|
{"with_underscore", "With_underscore"},
|
|
|
|
|
|
{"unicode \xe2\x80\xa8 line separator", "Unicode \xe2\x80\xa8 Line Separator"},
|
2010-07-20 19:53:59 -07:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestTitle(t *testing.T) {
|
|
|
|
|
|
for _, tt := range TitleTests {
|
|
|
|
|
|
if s := string(Title([]byte(tt.in))); s != tt.out {
|
|
|
|
|
|
t.Errorf("Title(%q) = %q, want %q", tt.in, s, tt.out)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2011-09-26 19:35:32 -04:00
|
|
|
|
|
2013-09-11 21:20:15 +10:00
|
|
|
|
var ToTitleTests = []TitleTest{
|
|
|
|
|
|
{"", ""},
|
|
|
|
|
|
{"a", "A"},
|
|
|
|
|
|
{" aaa aaa aaa ", " AAA AAA AAA "},
|
|
|
|
|
|
{" Aaa Aaa Aaa ", " AAA AAA AAA "},
|
|
|
|
|
|
{"123a456", "123A456"},
|
|
|
|
|
|
{"double-blind", "DOUBLE-BLIND"},
|
|
|
|
|
|
{"ÿøû", "ŸØÛ"},
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestToTitle(t *testing.T) {
|
|
|
|
|
|
for _, tt := range ToTitleTests {
|
|
|
|
|
|
if s := string(ToTitle([]byte(tt.in))); s != tt.out {
|
|
|
|
|
|
t.Errorf("ToTitle(%q) = %q, want %q", tt.in, s, tt.out)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2011-09-26 19:35:32 -04:00
|
|
|
|
var EqualFoldTests = []struct {
|
|
|
|
|
|
s, t string
|
|
|
|
|
|
out bool
|
|
|
|
|
|
}{
|
|
|
|
|
|
{"abc", "abc", true},
|
|
|
|
|
|
{"ABcd", "ABcd", true},
|
|
|
|
|
|
{"123abc", "123ABC", true},
|
|
|
|
|
|
{"αβδ", "ΑΒΔ", true},
|
|
|
|
|
|
{"abc", "xyz", false},
|
|
|
|
|
|
{"abc", "XYZ", false},
|
|
|
|
|
|
{"abcdefghijk", "abcdefghijX", false},
|
|
|
|
|
|
{"abcdefghijk", "abcdefghij\u212A", true},
|
|
|
|
|
|
{"abcdefghijK", "abcdefghij\u212A", true},
|
|
|
|
|
|
{"abcdefghijkz", "abcdefghij\u212Ay", false},
|
|
|
|
|
|
{"abcdefghijKz", "abcdefghij\u212Ay", false},
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestEqualFold(t *testing.T) {
|
|
|
|
|
|
for _, tt := range EqualFoldTests {
|
|
|
|
|
|
if out := EqualFold([]byte(tt.s), []byte(tt.t)); out != tt.out {
|
|
|
|
|
|
t.Errorf("EqualFold(%#q, %#q) = %v, want %v", tt.s, tt.t, out, tt.out)
|
|
|
|
|
|
}
|
|
|
|
|
|
if out := EqualFold([]byte(tt.t), []byte(tt.s)); out != tt.out {
|
|
|
|
|
|
t.Errorf("EqualFold(%#q, %#q) = %v, want %v", tt.t, tt.s, out, tt.out)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2012-09-18 15:02:08 -04:00
|
|
|
|
|
2013-09-11 21:20:15 +10:00
|
|
|
|
func TestBufferGrowNegative(t *testing.T) {
|
|
|
|
|
|
defer func() {
|
|
|
|
|
|
if err := recover(); err == nil {
|
2014-05-10 07:13:42 +09:00
|
|
|
|
t.Fatal("Grow(-1) should have panicked")
|
2013-09-11 21:20:15 +10:00
|
|
|
|
}
|
|
|
|
|
|
}()
|
|
|
|
|
|
var b Buffer
|
|
|
|
|
|
b.Grow(-1)
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestBufferTruncateNegative(t *testing.T) {
|
|
|
|
|
|
defer func() {
|
|
|
|
|
|
if err := recover(); err == nil {
|
2014-05-10 07:13:42 +09:00
|
|
|
|
t.Fatal("Truncate(-1) should have panicked")
|
2013-09-11 21:20:15 +10:00
|
|
|
|
}
|
|
|
|
|
|
}()
|
|
|
|
|
|
var b Buffer
|
|
|
|
|
|
b.Truncate(-1)
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestBufferTruncateOutOfRange(t *testing.T) {
|
|
|
|
|
|
defer func() {
|
|
|
|
|
|
if err := recover(); err == nil {
|
2014-05-10 07:13:42 +09:00
|
|
|
|
t.Fatal("Truncate(20) should have panicked")
|
2013-09-11 21:20:15 +10:00
|
|
|
|
}
|
|
|
|
|
|
}()
|
|
|
|
|
|
var b Buffer
|
|
|
|
|
|
b.Write(make([]byte, 10))
|
|
|
|
|
|
b.Truncate(20)
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2013-12-28 20:33:05 +11:00
|
|
|
|
var containsTests = []struct {
|
|
|
|
|
|
b, subslice []byte
|
|
|
|
|
|
want bool
|
|
|
|
|
|
}{
|
|
|
|
|
|
{[]byte("hello"), []byte("hel"), true},
|
|
|
|
|
|
{[]byte("日本語"), []byte("日本"), true},
|
|
|
|
|
|
{[]byte("hello"), []byte("Hello, world"), false},
|
|
|
|
|
|
{[]byte("東京"), []byte("京東"), false},
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestContains(t *testing.T) {
|
|
|
|
|
|
for _, tt := range containsTests {
|
|
|
|
|
|
if got := Contains(tt.b, tt.subslice); got != tt.want {
|
|
|
|
|
|
t.Errorf("Contains(%q, %q) = %v, want %v", tt.b, tt.subslice, got, tt.want)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2016-04-05 15:43:07 -07:00
|
|
|
|
var ContainsAnyTests = []struct {
|
|
|
|
|
|
b []byte
|
|
|
|
|
|
substr string
|
|
|
|
|
|
expected bool
|
|
|
|
|
|
}{
|
|
|
|
|
|
{[]byte(""), "", false},
|
|
|
|
|
|
{[]byte(""), "a", false},
|
|
|
|
|
|
{[]byte(""), "abc", false},
|
|
|
|
|
|
{[]byte("a"), "", false},
|
|
|
|
|
|
{[]byte("a"), "a", true},
|
|
|
|
|
|
{[]byte("aaa"), "a", true},
|
|
|
|
|
|
{[]byte("abc"), "xyz", false},
|
|
|
|
|
|
{[]byte("abc"), "xcz", true},
|
|
|
|
|
|
{[]byte("a☺b☻c☹d"), "uvw☻xyz", true},
|
|
|
|
|
|
{[]byte("aRegExp*"), ".(|)*+?^$[]", true},
|
|
|
|
|
|
{[]byte(dots + dots + dots), " ", false},
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestContainsAny(t *testing.T) {
|
|
|
|
|
|
for _, ct := range ContainsAnyTests {
|
|
|
|
|
|
if ContainsAny(ct.b, ct.substr) != ct.expected {
|
|
|
|
|
|
t.Errorf("ContainsAny(%s, %s) = %v, want %v",
|
|
|
|
|
|
ct.b, ct.substr, !ct.expected, ct.expected)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2016-04-07 23:22:30 -07:00
|
|
|
|
var ContainsRuneTests = []struct {
|
|
|
|
|
|
b []byte
|
|
|
|
|
|
r rune
|
|
|
|
|
|
expected bool
|
|
|
|
|
|
}{
|
|
|
|
|
|
{[]byte(""), 'a', false},
|
|
|
|
|
|
{[]byte("a"), 'a', true},
|
|
|
|
|
|
{[]byte("aaa"), 'a', true},
|
|
|
|
|
|
{[]byte("abc"), 'y', false},
|
|
|
|
|
|
{[]byte("abc"), 'c', true},
|
|
|
|
|
|
{[]byte("a☺b☻c☹d"), 'x', false},
|
|
|
|
|
|
{[]byte("a☺b☻c☹d"), '☻', true},
|
|
|
|
|
|
{[]byte("aRegExp*"), '*', true},
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func TestContainsRune(t *testing.T) {
|
|
|
|
|
|
for _, ct := range ContainsRuneTests {
|
|
|
|
|
|
if ContainsRune(ct.b, ct.r) != ct.expected {
|
|
|
|
|
|
t.Errorf("ContainsRune(%q, %q) = %v, want %v",
|
|
|
|
|
|
ct.b, ct.r, !ct.expected, ct.expected)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2012-09-18 15:02:08 -04:00
|
|
|
|
var makeFieldsInput = func() []byte {
|
|
|
|
|
|
x := make([]byte, 1<<20)
|
2012-10-30 13:38:01 -07:00
|
|
|
|
// Input is ~10% space, ~10% 2-byte UTF-8, rest ASCII non-space.
|
2012-09-18 15:02:08 -04:00
|
|
|
|
for i := range x {
|
|
|
|
|
|
switch rand.Intn(10) {
|
|
|
|
|
|
case 0:
|
|
|
|
|
|
x[i] = ' '
|
|
|
|
|
|
case 1:
|
|
|
|
|
|
if i > 0 && x[i-1] == 'x' {
|
|
|
|
|
|
copy(x[i-1:], "χ")
|
|
|
|
|
|
break
|
|
|
|
|
|
}
|
|
|
|
|
|
fallthrough
|
|
|
|
|
|
default:
|
|
|
|
|
|
x[i] = 'x'
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
return x
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2017-08-12 10:20:30 +02:00
|
|
|
|
var makeFieldsInputASCII = func() []byte {
|
|
|
|
|
|
x := make([]byte, 1<<20)
|
|
|
|
|
|
// Input is ~10% space, rest ASCII non-space.
|
|
|
|
|
|
for i := range x {
|
|
|
|
|
|
if rand.Intn(10) == 0 {
|
|
|
|
|
|
x[i] = ' '
|
|
|
|
|
|
} else {
|
|
|
|
|
|
x[i] = 'x'
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
return x
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
var bytesdata = []struct {
|
|
|
|
|
|
name string
|
|
|
|
|
|
data []byte
|
|
|
|
|
|
}{
|
|
|
|
|
|
{"ASCII", makeFieldsInputASCII()},
|
|
|
|
|
|
{"Mixed", makeFieldsInput()},
|
|
|
|
|
|
}
|
2012-09-18 15:02:08 -04:00
|
|
|
|
|
|
|
|
|
|
func BenchmarkFields(b *testing.B) {
|
2017-08-12 10:20:30 +02:00
|
|
|
|
for _, sd := range bytesdata {
|
|
|
|
|
|
b.Run(sd.name, func(b *testing.B) {
|
|
|
|
|
|
for j := 1 << 4; j <= 1<<20; j <<= 4 {
|
|
|
|
|
|
b.Run(fmt.Sprintf("%d", j), func(b *testing.B) {
|
|
|
|
|
|
b.ReportAllocs()
|
|
|
|
|
|
b.SetBytes(int64(j))
|
|
|
|
|
|
data := sd.data[:j]
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
Fields(data)
|
|
|
|
|
|
}
|
|
|
|
|
|
})
|
|
|
|
|
|
}
|
|
|
|
|
|
})
|
2012-09-18 15:02:08 -04:00
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func BenchmarkFieldsFunc(b *testing.B) {
|
2017-08-12 10:20:30 +02:00
|
|
|
|
for _, sd := range bytesdata {
|
|
|
|
|
|
b.Run(sd.name, func(b *testing.B) {
|
|
|
|
|
|
for j := 1 << 4; j <= 1<<20; j <<= 4 {
|
|
|
|
|
|
b.Run(fmt.Sprintf("%d", j), func(b *testing.B) {
|
|
|
|
|
|
b.ReportAllocs()
|
|
|
|
|
|
b.SetBytes(int64(j))
|
|
|
|
|
|
data := sd.data[:j]
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
FieldsFunc(data, unicode.IsSpace)
|
|
|
|
|
|
}
|
|
|
|
|
|
})
|
|
|
|
|
|
}
|
|
|
|
|
|
})
|
2012-09-18 15:02:08 -04:00
|
|
|
|
}
|
|
|
|
|
|
}
|
2013-02-07 16:00:06 -08:00
|
|
|
|
|
|
|
|
|
|
func BenchmarkTrimSpace(b *testing.B) {
|
2018-12-06 08:53:29 -05:00
|
|
|
|
tests := []struct {
|
|
|
|
|
|
name string
|
|
|
|
|
|
input []byte
|
|
|
|
|
|
}{
|
|
|
|
|
|
{"NoTrim", []byte("typical")},
|
|
|
|
|
|
{"ASCII", []byte(" foo bar ")},
|
|
|
|
|
|
{"SomeNonASCII", []byte(" \u2000\t\r\n x\t\t\r\r\ny\n \u3000 ")},
|
|
|
|
|
|
{"JustNonASCII", []byte("\u2000\u2000\u2000☺☺☺☺\u3000\u3000\u3000")},
|
|
|
|
|
|
}
|
|
|
|
|
|
for _, test := range tests {
|
|
|
|
|
|
b.Run(test.name, func(b *testing.B) {
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
TrimSpace(test.input)
|
|
|
|
|
|
}
|
|
|
|
|
|
})
|
2013-02-07 16:00:06 -08:00
|
|
|
|
}
|
|
|
|
|
|
}
|
2014-06-11 19:03:59 -07:00
|
|
|
|
|
2017-02-07 13:38:52 +02:00
|
|
|
|
func makeBenchInputHard() []byte {
|
|
|
|
|
|
tokens := [...]string{
|
|
|
|
|
|
"<a>", "<p>", "<b>", "<strong>",
|
|
|
|
|
|
"</a>", "</p>", "</b>", "</strong>",
|
|
|
|
|
|
"hello", "world",
|
|
|
|
|
|
}
|
|
|
|
|
|
x := make([]byte, 0, 1<<20)
|
|
|
|
|
|
for {
|
|
|
|
|
|
i := rand.Intn(len(tokens))
|
|
|
|
|
|
if len(x)+len(tokens[i]) >= 1<<20 {
|
|
|
|
|
|
break
|
|
|
|
|
|
}
|
|
|
|
|
|
x = append(x, tokens[i]...)
|
|
|
|
|
|
}
|
|
|
|
|
|
return x
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
var benchInputHard = makeBenchInputHard()
|
|
|
|
|
|
|
2019-03-15 11:29:03 +01:00
|
|
|
|
func benchmarkIndexHard(b *testing.B, sep []byte) {
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
Index(benchInputHard, sep)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2019-03-08 11:36:31 +01:00
|
|
|
|
func benchmarkLastIndexHard(b *testing.B, sep []byte) {
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
LastIndex(benchInputHard, sep)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2019-03-15 11:29:03 +01:00
|
|
|
|
func benchmarkCountHard(b *testing.B, sep []byte) {
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
Count(benchInputHard, sep)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func BenchmarkIndexHard1(b *testing.B) { benchmarkIndexHard(b, []byte("<>")) }
|
|
|
|
|
|
func BenchmarkIndexHard2(b *testing.B) { benchmarkIndexHard(b, []byte("</pre>")) }
|
|
|
|
|
|
func BenchmarkIndexHard3(b *testing.B) { benchmarkIndexHard(b, []byte("<b>hello world</b>")) }
|
|
|
|
|
|
func BenchmarkIndexHard4(b *testing.B) {
|
|
|
|
|
|
benchmarkIndexHard(b, []byte("<pre><b>hello</b><strong>world</strong></pre>"))
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2019-03-08 11:36:31 +01:00
|
|
|
|
func BenchmarkLastIndexHard1(b *testing.B) { benchmarkLastIndexHard(b, []byte("<>")) }
|
|
|
|
|
|
func BenchmarkLastIndexHard2(b *testing.B) { benchmarkLastIndexHard(b, []byte("</pre>")) }
|
|
|
|
|
|
func BenchmarkLastIndexHard3(b *testing.B) { benchmarkLastIndexHard(b, []byte("<b>hello world</b>")) }
|
|
|
|
|
|
|
2019-03-15 11:29:03 +01:00
|
|
|
|
func BenchmarkCountHard1(b *testing.B) { benchmarkCountHard(b, []byte("<>")) }
|
|
|
|
|
|
func BenchmarkCountHard2(b *testing.B) { benchmarkCountHard(b, []byte("</pre>")) }
|
|
|
|
|
|
func BenchmarkCountHard3(b *testing.B) { benchmarkCountHard(b, []byte("<b>hello world</b>")) }
|
|
|
|
|
|
|
2017-02-07 13:38:52 +02:00
|
|
|
|
func BenchmarkSplitEmptySeparator(b *testing.B) {
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
Split(benchInputHard, nil)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func BenchmarkSplitSingleByteSeparator(b *testing.B) {
|
|
|
|
|
|
sep := []byte("/")
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
Split(benchInputHard, sep)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func BenchmarkSplitMultiByteSeparator(b *testing.B) {
|
|
|
|
|
|
sep := []byte("hello")
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
Split(benchInputHard, sep)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func BenchmarkSplitNSingleByteSeparator(b *testing.B) {
|
|
|
|
|
|
sep := []byte("/")
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
SplitN(benchInputHard, sep, 10)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func BenchmarkSplitNMultiByteSeparator(b *testing.B) {
|
|
|
|
|
|
sep := []byte("hello")
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
SplitN(benchInputHard, sep, 10)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
2014-06-11 19:03:59 -07:00
|
|
|
|
func BenchmarkRepeat(b *testing.B) {
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
Repeat([]byte("-"), 80)
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2015-07-02 11:43:46 -07:00
|
|
|
|
|
2016-05-26 10:54:25 +02:00
|
|
|
|
func BenchmarkBytesCompare(b *testing.B) {
|
|
|
|
|
|
for n := 1; n <= 2048; n <<= 1 {
|
|
|
|
|
|
b.Run(fmt.Sprint(n), func(b *testing.B) {
|
|
|
|
|
|
var x = make([]byte, n)
|
|
|
|
|
|
var y = make([]byte, n)
|
2015-07-02 11:43:46 -07:00
|
|
|
|
|
2016-05-26 10:54:25 +02:00
|
|
|
|
for i := 0; i < n; i++ {
|
|
|
|
|
|
x[i] = 'a'
|
|
|
|
|
|
}
|
2015-07-02 11:43:46 -07:00
|
|
|
|
|
2016-05-26 10:54:25 +02:00
|
|
|
|
for i := 0; i < n; i++ {
|
|
|
|
|
|
y[i] = 'a'
|
|
|
|
|
|
}
|
2015-07-02 11:43:46 -07:00
|
|
|
|
|
2016-05-26 10:54:25 +02:00
|
|
|
|
b.ResetTimer()
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
Compare(x, y)
|
|
|
|
|
|
}
|
|
|
|
|
|
})
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
bytes, strings: optimize for ASCII sets
In a large codebase within Google, there are thousands of uses of:
ContainsAny|IndexAny|LastIndexAny|Trim|TrimLeft|TrimRight
An analysis of their usage shows that over 97% of them only use character
sets consisting of only ASCII symbols.
Uses of ContainsAny|IndexAny|LastIndexAny:
6% are 1 character (e.g., "\n" or " ")
58% are 2-4 characters (e.g., "<>" or "\r\n\t ")
24% are 5-9 characters (e.g., "()[]*^$")
10% are 10+ characters (e.g., "+-=&|><!(){}[]^\"~*?:\\/ ")
We optimize for ASCII sets, which are commonly used to search for
"control" characters in some string. We don't optimize for the
single character scenario since IndexRune or IndexByte could be used.
Uses of Trim|TrimLeft|TrimRight:
71% are 1 character (e.g., "\n" or " ")
14% are 2 characters (e.g., "\r\n")
10% are 3-4 characters (e.g., " \t\r\n")
5% are 10+ characters (e.g., "0123456789abcdefABCDEF")
We optimize for the single character case with a simple closured function
that only checks for that character's value. We optimize for the medium
and larger sets using a 16-byte bit-map representing a set of ASCII characters.
The benchmarks below have the following suffix name "%d:%d" where the first
number is the length of the input and the second number is the length
of the charset.
== bytes package ==
benchmark old ns/op new ns/op delta
BenchmarkIndexAnyASCII/1:1-4 5.09 5.23 +2.75%
BenchmarkIndexAnyASCII/1:2-4 5.81 5.85 +0.69%
BenchmarkIndexAnyASCII/1:4-4 7.22 7.50 +3.88%
BenchmarkIndexAnyASCII/1:8-4 11.0 11.1 +0.91%
BenchmarkIndexAnyASCII/1:16-4 17.5 17.8 +1.71%
BenchmarkIndexAnyASCII/16:1-4 36.0 34.0 -5.56%
BenchmarkIndexAnyASCII/16:2-4 46.6 36.5 -21.67%
BenchmarkIndexAnyASCII/16:4-4 78.0 40.4 -48.21%
BenchmarkIndexAnyASCII/16:8-4 136 47.4 -65.15%
BenchmarkIndexAnyASCII/16:16-4 254 61.5 -75.79%
BenchmarkIndexAnyASCII/256:1-4 542 388 -28.41%
BenchmarkIndexAnyASCII/256:2-4 705 382 -45.82%
BenchmarkIndexAnyASCII/256:4-4 1089 386 -64.55%
BenchmarkIndexAnyASCII/256:8-4 1994 394 -80.24%
BenchmarkIndexAnyASCII/256:16-4 3843 411 -89.31%
BenchmarkIndexAnyASCII/4096:1-4 8522 5873 -31.08%
BenchmarkIndexAnyASCII/4096:2-4 11253 5861 -47.92%
BenchmarkIndexAnyASCII/4096:4-4 17824 5883 -66.99%
BenchmarkIndexAnyASCII/4096:8-4 32053 5871 -81.68%
BenchmarkIndexAnyASCII/4096:16-4 60512 5888 -90.27%
BenchmarkTrimASCII/1:1-4 79.5 70.8 -10.94%
BenchmarkTrimASCII/1:2-4 79.0 105 +32.91%
BenchmarkTrimASCII/1:4-4 79.6 109 +36.93%
BenchmarkTrimASCII/1:8-4 78.8 118 +49.75%
BenchmarkTrimASCII/1:16-4 80.2 132 +64.59%
BenchmarkTrimASCII/16:1-4 243 116 -52.26%
BenchmarkTrimASCII/16:2-4 243 171 -29.63%
BenchmarkTrimASCII/16:4-4 243 176 -27.57%
BenchmarkTrimASCII/16:8-4 241 184 -23.65%
BenchmarkTrimASCII/16:16-4 238 199 -16.39%
BenchmarkTrimASCII/256:1-4 2580 840 -67.44%
BenchmarkTrimASCII/256:2-4 2603 1175 -54.86%
BenchmarkTrimASCII/256:4-4 2572 1188 -53.81%
BenchmarkTrimASCII/256:8-4 2550 1191 -53.29%
BenchmarkTrimASCII/256:16-4 2585 1208 -53.27%
BenchmarkTrimASCII/4096:1-4 39773 12181 -69.37%
BenchmarkTrimASCII/4096:2-4 39946 17231 -56.86%
BenchmarkTrimASCII/4096:4-4 39641 17179 -56.66%
BenchmarkTrimASCII/4096:8-4 39835 17175 -56.88%
BenchmarkTrimASCII/4096:16-4 40229 17215 -57.21%
== strings package ==
benchmark old ns/op new ns/op delta
BenchmarkIndexAnyASCII/1:1-4 5.94 4.97 -16.33%
BenchmarkIndexAnyASCII/1:2-4 5.94 5.55 -6.57%
BenchmarkIndexAnyASCII/1:4-4 7.45 7.21 -3.22%
BenchmarkIndexAnyASCII/1:8-4 10.8 10.6 -1.85%
BenchmarkIndexAnyASCII/1:16-4 17.4 17.2 -1.15%
BenchmarkIndexAnyASCII/16:1-4 36.4 32.2 -11.54%
BenchmarkIndexAnyASCII/16:2-4 49.6 34.6 -30.24%
BenchmarkIndexAnyASCII/16:4-4 77.5 37.9 -51.10%
BenchmarkIndexAnyASCII/16:8-4 138 45.5 -67.03%
BenchmarkIndexAnyASCII/16:16-4 241 59.1 -75.48%
BenchmarkIndexAnyASCII/256:1-4 509 378 -25.74%
BenchmarkIndexAnyASCII/256:2-4 720 381 -47.08%
BenchmarkIndexAnyASCII/256:4-4 1142 384 -66.37%
BenchmarkIndexAnyASCII/256:8-4 1999 391 -80.44%
BenchmarkIndexAnyASCII/256:16-4 3735 403 -89.21%
BenchmarkIndexAnyASCII/4096:1-4 7973 5824 -26.95%
BenchmarkIndexAnyASCII/4096:2-4 11432 5809 -49.19%
BenchmarkIndexAnyASCII/4096:4-4 18327 5819 -68.25%
BenchmarkIndexAnyASCII/4096:8-4 33059 5828 -82.37%
BenchmarkIndexAnyASCII/4096:16-4 59703 5817 -90.26%
BenchmarkTrimASCII/1:1-4 71.9 71.8 -0.14%
BenchmarkTrimASCII/1:2-4 73.3 103 +40.52%
BenchmarkTrimASCII/1:4-4 71.8 106 +47.63%
BenchmarkTrimASCII/1:8-4 71.2 113 +58.71%
BenchmarkTrimASCII/1:16-4 71.6 128 +78.77%
BenchmarkTrimASCII/16:1-4 152 116 -23.68%
BenchmarkTrimASCII/16:2-4 160 168 +5.00%
BenchmarkTrimASCII/16:4-4 172 170 -1.16%
BenchmarkTrimASCII/16:8-4 200 177 -11.50%
BenchmarkTrimASCII/16:16-4 254 193 -24.02%
BenchmarkTrimASCII/256:1-4 1438 864 -39.92%
BenchmarkTrimASCII/256:2-4 1551 1195 -22.95%
BenchmarkTrimASCII/256:4-4 1770 1200 -32.20%
BenchmarkTrimASCII/256:8-4 2195 1216 -44.60%
BenchmarkTrimASCII/256:16-4 3054 1224 -59.92%
BenchmarkTrimASCII/4096:1-4 21726 12557 -42.20%
BenchmarkTrimASCII/4096:2-4 23586 17508 -25.77%
BenchmarkTrimASCII/4096:4-4 26898 17510 -34.90%
BenchmarkTrimASCII/4096:8-4 33714 17595 -47.81%
BenchmarkTrimASCII/4096:16-4 47429 17700 -62.68%
The benchmarks added test the worst case. For IndexAny, that is when the
charset matches none of the input. For Trim, it is when the charset matches
all of the input.
Change-Id: I970874d101a96b33528fc99b165379abe58cf6ea
Reviewed-on: https://go-review.googlesource.com/31593
Run-TryBot: Joe Tsai <thebrokentoaster@gmail.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Brad Fitzpatrick <bradfitz@golang.org>
Reviewed-by: Martin Möhrmann <martisch@uos.de>
2016-10-20 03:16:22 -07:00
|
|
|
|
|
|
|
|
|
|
func BenchmarkIndexAnyASCII(b *testing.B) {
|
|
|
|
|
|
x := Repeat([]byte{'#'}, 4096) // Never matches set
|
|
|
|
|
|
cs := "0123456789abcdef"
|
|
|
|
|
|
for k := 1; k <= 4096; k <<= 4 {
|
|
|
|
|
|
for j := 1; j <= 16; j <<= 1 {
|
|
|
|
|
|
b.Run(fmt.Sprintf("%d:%d", k, j), func(b *testing.B) {
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
IndexAny(x[:k], cs[:j])
|
|
|
|
|
|
}
|
|
|
|
|
|
})
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
func BenchmarkTrimASCII(b *testing.B) {
|
|
|
|
|
|
cs := "0123456789abcdef"
|
|
|
|
|
|
for k := 1; k <= 4096; k <<= 4 {
|
|
|
|
|
|
for j := 1; j <= 16; j <<= 1 {
|
|
|
|
|
|
b.Run(fmt.Sprintf("%d:%d", k, j), func(b *testing.B) {
|
|
|
|
|
|
x := Repeat([]byte(cs[:j]), k) // Always matches set
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
Trim(x[:k], cs[:j])
|
|
|
|
|
|
}
|
|
|
|
|
|
})
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|
2017-11-04 10:19:53 -07:00
|
|
|
|
|
|
|
|
|
|
func BenchmarkIndexPeriodic(b *testing.B) {
|
|
|
|
|
|
key := []byte{1, 1}
|
|
|
|
|
|
for _, skip := range [...]int{2, 4, 8, 16, 32, 64} {
|
|
|
|
|
|
b.Run(fmt.Sprintf("IndexPeriodic%d", skip), func(b *testing.B) {
|
|
|
|
|
|
buf := make([]byte, 1<<16)
|
|
|
|
|
|
for i := 0; i < len(buf); i += skip {
|
|
|
|
|
|
buf[i] = 1
|
|
|
|
|
|
}
|
|
|
|
|
|
for i := 0; i < b.N; i++ {
|
|
|
|
|
|
Index(buf, key)
|
|
|
|
|
|
}
|
|
|
|
|
|
})
|
|
|
|
|
|
}
|
|
|
|
|
|
}
|