ires 0.1.3 → 0.1.4

data/ext/vendor/golang.org/x/text/unicode/norm/transform.go

@@ -0,0 +1,88 @@
+// Copyright 2013 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.
+
+package norm
+
+import (
+	"unicode/utf8"
+
+	"golang.org/x/text/transform"
+)
+
+// Reset implements the Reset method of the transform.Transformer interface.
+func (Form) Reset() {}
+
+// Transform implements the Transform method of the transform.Transformer
+// interface. It may need to write segments of up to MaxSegmentSize at once.
+// Users should either catch ErrShortDst and allow dst to grow or have dst be at
+// least of size MaxTransformChunkSize to be guaranteed of progress.
+func (f Form) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
+	n := 0
+	// Cap the maximum number of src bytes to check.
+	b := src
+	eof := atEOF
+	if ns := len(dst); ns < len(b) {
+		err = transform.ErrShortDst
+		eof = false
+		b = b[:ns]
+	}
+	i, ok := formTable[f].quickSpan(inputBytes(b), n, len(b), eof)
+	n += copy(dst[n:], b[n:i])
+	if !ok {
+		nDst, nSrc, err = f.transform(dst[n:], src[n:], atEOF)
+		return nDst + n, nSrc + n, err
+	}
+	if n < len(src) && !atEOF {
+		err = transform.ErrShortSrc
+	}
+	return n, n, err
+}
+
+func flushTransform(rb *reorderBuffer) bool {
+	// Write out (must fully fit in dst, or else it is an ErrShortDst).
+	if len(rb.out) < rb.nrune*utf8.UTFMax {
+		return false
+	}
+	rb.out = rb.out[rb.flushCopy(rb.out):]
+	return true
+}
+
+var errs = []error{nil, transform.ErrShortDst, transform.ErrShortSrc}
+
+// transform implements the transform.Transformer interface. It is only called
+// when quickSpan does not pass for a given string.
+func (f Form) transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
+	// TODO: get rid of reorderBuffer. See CL 23460044.
+	rb := reorderBuffer{}
+	rb.init(f, src)
+	for {
+		// Load segment into reorder buffer.
+		rb.setFlusher(dst[nDst:], flushTransform)
+		end := decomposeSegment(&rb, nSrc, atEOF)
+		if end < 0 {
+			return nDst, nSrc, errs[-end]
+		}
+		nDst = len(dst) - len(rb.out)
+		nSrc = end
+
+		// Next quickSpan.
+		end = rb.nsrc
+		eof := atEOF
+		if n := nSrc + len(dst) - nDst; n < end {
+			err = transform.ErrShortDst
+			end = n
+			eof = false
+		}
+		end, ok := rb.f.quickSpan(rb.src, nSrc, end, eof)
+		n := copy(dst[nDst:], rb.src.bytes[nSrc:end])
+		nSrc += n
+		nDst += n
+		if ok {
+			if n < rb.nsrc && !atEOF {
+				err = transform.ErrShortSrc
+			}
+			return nDst, nSrc, err
+		}
+	}
+}

data/ext/vendor/golang.org/x/text/unicode/norm/transform_test.go

@@ -0,0 +1,101 @@
+// Copyright 2011 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.
+
+package norm
+
+import (
+	"fmt"
+	"testing"
+
+	"golang.org/x/text/transform"
+)
+
+func TestTransform(t *testing.T) {
+	tests := []struct {
+		f       Form
+		in, out string
+		eof     bool
+		dstSize int
+		err     error
+	}{
+		{NFC, "ab", "ab", true, 2, nil},
+		{NFC, "qx", "qx", true, 2, nil},
+		{NFD, "qx", "qx", true, 2, nil},
+		{NFC, "", "", true, 1, nil},
+		{NFD, "", "", true, 1, nil},
+		{NFC, "", "", false, 1, nil},
+		{NFD, "", "", false, 1, nil},
+
+		// Normalized segment does not fit in destination.
+		{NFD, "ö", "", true, 1, transform.ErrShortDst},
+		{NFD, "ö", "", true, 2, transform.ErrShortDst},
+
+		// As an artifact of the algorithm, only full segments are written.
+		// This is not strictly required, and some bytes could be written.
+		// In practice, for Transform to not block, the destination buffer
+		// should be at least MaxSegmentSize to work anyway and these edge
+		// conditions will be relatively rare.
+		{NFC, "ab", "", true, 1, transform.ErrShortDst},
+		// This is even true for inert runes.
+		{NFC, "qx", "", true, 1, transform.ErrShortDst},
+		{NFC, "a\u0300abc", "\u00e0a", true, 4, transform.ErrShortDst},
+
+		// We cannot write a segment if successive runes could still change the result.
+		{NFD, "ö", "", false, 3, transform.ErrShortSrc},
+		{NFC, "a\u0300", "", false, 4, transform.ErrShortSrc},
+		{NFD, "a\u0300", "", false, 4, transform.ErrShortSrc},
+		{NFC, "ö", "", false, 3, transform.ErrShortSrc},
+
+		{NFC, "a\u0300", "", true, 1, transform.ErrShortDst},
+		// Theoretically could fit, but won't due to simplified checks.
+		{NFC, "a\u0300", "", true, 2, transform.ErrShortDst},
+		{NFC, "a\u0300", "", true, 3, transform.ErrShortDst},
+		{NFC, "a\u0300", "\u00e0", true, 4, nil},
+
+		{NFD, "öa\u0300", "o\u0308", false, 8, transform.ErrShortSrc},
+		{NFD, "öa\u0300ö", "o\u0308a\u0300", true, 8, transform.ErrShortDst},
+		{NFD, "öa\u0300ö", "o\u0308a\u0300", false, 12, transform.ErrShortSrc},
+
+		// Illegal input is copied verbatim.
+		{NFD, "\xbd\xb2=\xbc ", "\xbd\xb2=\xbc ", true, 8, nil},
+	}
+	b := make([]byte, 100)
+	for i, tt := range tests {
+		nDst, _, err := tt.f.Transform(b[:tt.dstSize], []byte(tt.in), tt.eof)
+		out := string(b[:nDst])
+		if out != tt.out || err != tt.err {
+			t.Errorf("%d: was %+q (%v); want %+q (%v)", i, out, err, tt.out, tt.err)
+		}
+		if want := tt.f.String(tt.in)[:nDst]; want != out {
+			t.Errorf("%d: incorrect normalization: was %+q; want %+q", i, out, want)
+		}
+	}
+}
+
+var transBufSizes = []int{
+	MaxTransformChunkSize,
+	3 * MaxTransformChunkSize / 2,
+	2 * MaxTransformChunkSize,
+	3 * MaxTransformChunkSize,
+	100 * MaxTransformChunkSize,
+}
+
+func doTransNorm(f Form, buf []byte, b []byte) []byte {
+	acc := []byte{}
+	for p := 0; p < len(b); {
+		nd, ns, _ := f.Transform(buf[:], b[p:], true)
+		p += ns
+		acc = append(acc, buf[:nd]...)
+	}
+	return acc
+}
+
+func TestTransformNorm(t *testing.T) {
+	for _, sz := range transBufSizes {
+		buf := make([]byte, sz)
+		runNormTests(t, fmt.Sprintf("Transform:%d", sz), func(f Form, out []byte, s string) []byte {
+			return doTransNorm(f, buf, append(out, s...))
+		})
+	}
+}

data/ext/vendor/golang.org/x/text/unicode/norm/trie.go

@@ -0,0 +1,54 @@
+// Copyright 2011 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.
+
+package norm
+
+type valueRange struct {
+	value  uint16 // header: value:stride
+	lo, hi byte   // header: lo:n
+}
+
+type sparseBlocks struct {
+	values []valueRange
+	offset []uint16
+}
+
+var nfcSparse = sparseBlocks{
+	values: nfcSparseValues[:],
+	offset: nfcSparseOffset[:],
+}
+
+var nfkcSparse = sparseBlocks{
+	values: nfkcSparseValues[:],
+	offset: nfkcSparseOffset[:],
+}
+
+var (
+	nfcData  = newNfcTrie(0)
+	nfkcData = newNfkcTrie(0)
+)
+
+// lookupValue determines the type of block n and looks up the value for b.
+// For n < t.cutoff, the block is a simple lookup table. Otherwise, the block
+// is a list of ranges with an accompanying value. Given a matching range r,
+// the value for b is by r.value + (b - r.lo) * stride.
+func (t *sparseBlocks) lookup(n uint32, b byte) uint16 {
+	offset := t.offset[n]
+	header := t.values[offset]
+	lo := offset + 1
+	hi := lo + uint16(header.lo)
+	for lo < hi {
+		m := lo + (hi-lo)/2
+		r := t.values[m]
+		if r.lo <= b && b <= r.hi {
+			return r.value + uint16(b-r.lo)*header.value
+		}
+		if b < r.lo {
+			hi = m
+		} else {
+			lo = m + 1
+		}
+	}
+	return 0
+}

data/ext/vendor/golang.org/x/text/unicode/norm/triegen.go

@@ -0,0 +1,117 @@
+// Copyright 2011 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.
+
+// +build ignore
+
+// Trie table generator.
+// Used by make*tables tools to generate a go file with trie data structures
+// for mapping UTF-8 to a 16-bit value. All but the last byte in a UTF-8 byte
+// sequence are used to lookup offsets in the index table to be used for the
+// next byte. The last byte is used to index into a table with 16-bit values.
+
+package main
+
+import (
+	"fmt"
+	"io"
+)
+
+const maxSparseEntries = 16
+
+type normCompacter struct {
+	sparseBlocks [][]uint64
+	sparseOffset []uint16
+	sparseCount  int
+	name         string
+}
+
+func mostFrequentStride(a []uint64) int {
+	counts := make(map[int]int)
+	var v int
+	for _, x := range a {
+		if stride := int(x) - v; v != 0 && stride >= 0 {
+			counts[stride]++
+		}
+		v = int(x)
+	}
+	var maxs, maxc int
+	for stride, cnt := range counts {
+		if cnt > maxc || (cnt == maxc && stride < maxs) {
+			maxs, maxc = stride, cnt
+		}
+	}
+	return maxs
+}
+
+func countSparseEntries(a []uint64) int {
+	stride := mostFrequentStride(a)
+	var v, count int
+	for _, tv := range a {
+		if int(tv)-v != stride {
+			if tv != 0 {
+				count++
+			}
+		}
+		v = int(tv)
+	}
+	return count
+}
+
+func (c *normCompacter) Size(v []uint64) (sz int, ok bool) {
+	if n := countSparseEntries(v); n <= maxSparseEntries {
+		return (n+1)*4 + 2, true
+	}
+	return 0, false
+}
+
+func (c *normCompacter) Store(v []uint64) uint32 {
+	h := uint32(len(c.sparseOffset))
+	c.sparseBlocks = append(c.sparseBlocks, v)
+	c.sparseOffset = append(c.sparseOffset, uint16(c.sparseCount))
+	c.sparseCount += countSparseEntries(v) + 1
+	return h
+}
+
+func (c *normCompacter) Handler() string {
+	return c.name + "Sparse.lookup"
+}
+
+func (c *normCompacter) Print(w io.Writer) (retErr error) {
+	p := func(f string, x ...interface{}) {
+		if _, err := fmt.Fprintf(w, f, x...); retErr == nil && err != nil {
+			retErr = err
+		}
+	}
+
+	ls := len(c.sparseBlocks)
+	p("// %sSparseOffset: %d entries, %d bytes\n", c.name, ls, ls*2)
+	p("var %sSparseOffset = %#v\n\n", c.name, c.sparseOffset)
+
+	ns := c.sparseCount
+	p("// %sSparseValues: %d entries, %d bytes\n", c.name, ns, ns*4)
+	p("var %sSparseValues = [%d]valueRange {", c.name, ns)
+	for i, b := range c.sparseBlocks {
+		p("\n// Block %#x, offset %#x", i, c.sparseOffset[i])
+		var v int
+		stride := mostFrequentStride(b)
+		n := countSparseEntries(b)
+		p("\n{value:%#04x,lo:%#02x},", stride, uint8(n))
+		for i, nv := range b {
+			if int(nv)-v != stride {
+				if v != 0 {
+					p(",hi:%#02x},", 0x80+i-1)
+				}
+				if nv != 0 {
+					p("\n{value:%#04x,lo:%#02x", nv, 0x80+i)
+				}
+			}
+			v = int(nv)
+		}
+		if v != 0 {
+			p(",hi:%#02x},", 0x80+len(b)-1)
+		}
+	}
+	p("\n}\n\n")
+	return
+}

data/ext/vendor/golang.org/x/text/unicode/norm/ucd_test.go

@@ -0,0 +1,275 @@
+// Copyright 2011 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.
+
+package norm
+
+import (
+	"bufio"
+	"bytes"
+	"fmt"
+	"regexp"
+	"runtime"
+	"strconv"
+	"strings"
+	"sync"
+	"testing"
+	"time"
+	"unicode/utf8"
+
+	"golang.org/x/text/internal/gen"
+	"golang.org/x/text/internal/testtext"
+)
+
+var once sync.Once
+
+func skipShort(t *testing.T) {
+	testtext.SkipIfNotLong(t)
+
+	once.Do(func() { loadTestData(t) })
+}
+
+// This regression test runs the test set in NormalizationTest.txt
+// (taken from http://www.unicode.org/Public/<unicode.Version>/ucd/).
+//
+// NormalizationTest.txt has form:
+// @Part0 # Specific cases
+// #
+// 1E0A;1E0A;0044 0307;1E0A;0044 0307; # (Ḋ; Ḋ; D◌̇; Ḋ; D◌̇; ) LATIN CAPITAL LETTER D WITH DOT ABOVE
+// 1E0C;1E0C;0044 0323;1E0C;0044 0323; # (Ḍ; Ḍ; D◌̣; Ḍ; D◌̣; ) LATIN CAPITAL LETTER D WITH DOT BELOW
+//
+// Each test has 5 columns (c1, c2, c3, c4, c5), where
+// (c1, c2, c3, c4, c5) == (c1, NFC(c1), NFD(c1), NFKC(c1), NFKD(c1))
+//
+// CONFORMANCE:
+// 1. The following invariants must be true for all conformant implementations
+//
+//    NFC
+//      c2 ==  NFC(c1) ==  NFC(c2) ==  NFC(c3)
+//      c4 ==  NFC(c4) ==  NFC(c5)
+//
+//    NFD
+//      c3 ==  NFD(c1) ==  NFD(c2) ==  NFD(c3)
+//      c5 ==  NFD(c4) ==  NFD(c5)
+//
+//    NFKC
+//      c4 == NFKC(c1) == NFKC(c2) == NFKC(c3) == NFKC(c4) == NFKC(c5)
+//
+//    NFKD
+//      c5 == NFKD(c1) == NFKD(c2) == NFKD(c3) == NFKD(c4) == NFKD(c5)
+//
+// 2. For every code point X assigned in this version of Unicode that is not
+//    specifically listed in Part 1, the following invariants must be true
+//    for all conformant implementations:
+//
+//      X == NFC(X) == NFD(X) == NFKC(X) == NFKD(X)
+//
+
+// Column types.
+const (
+	cRaw = iota
+	cNFC
+	cNFD
+	cNFKC
+	cNFKD
+	cMaxColumns
+)
+
+// Holds data from NormalizationTest.txt
+var part []Part
+
+type Part struct {
+	name   string
+	number int
+	tests  []Test
+}
+
+type Test struct {
+	name   string
+	partnr int
+	number int
+	r      rune                // used for character by character test
+	cols   [cMaxColumns]string // Each has 5 entries, see below.
+}
+
+func (t Test) Name() string {
+	if t.number < 0 {
+		return part[t.partnr].name
+	}
+	return fmt.Sprintf("%s:%d", part[t.partnr].name, t.number)
+}
+
+var partRe = regexp.MustCompile(`@Part(\d) # (.*)$`)
+var testRe = regexp.MustCompile(`^` + strings.Repeat(`([\dA-F ]+);`, 5) + ` # (.*)$`)
+
+var counter int
+
+// Load the data form NormalizationTest.txt
+func loadTestData(t *testing.T) {
+	f := gen.OpenUCDFile("NormalizationTest.txt")
+	defer f.Close()
+	scanner := bufio.NewScanner(f)
+	for scanner.Scan() {
+		line := scanner.Text()
+		if len(line) == 0 || line[0] == '#' {
+			continue
+		}
+		m := partRe.FindStringSubmatch(line)
+		if m != nil {
+			if len(m) < 3 {
+				t.Fatal("Failed to parse Part: ", line)
+			}
+			i, err := strconv.Atoi(m[1])
+			if err != nil {
+				t.Fatal(err)
+			}
+			name := m[2]
+			part = append(part, Part{name: name[:len(name)-1], number: i})
+			continue
+		}
+		m = testRe.FindStringSubmatch(line)
+		if m == nil || len(m) < 7 {
+			t.Fatalf(`Failed to parse: "%s" result: %#v`, line, m)
+		}
+		test := Test{name: m[6], partnr: len(part) - 1, number: counter}
+		counter++
+		for j := 1; j < len(m)-1; j++ {
+			for _, split := range strings.Split(m[j], " ") {
+				r, err := strconv.ParseUint(split, 16, 64)
+				if err != nil {
+					t.Fatal(err)
+				}
+				if test.r == 0 {
+					// save for CharacterByCharacterTests
+					test.r = rune(r)
+				}
+				var buf [utf8.UTFMax]byte
+				sz := utf8.EncodeRune(buf[:], rune(r))
+				test.cols[j-1] += string(buf[:sz])
+			}
+		}
+		part := &part[len(part)-1]
+		part.tests = append(part.tests, test)
+	}
+	if scanner.Err() != nil {
+		t.Fatal(scanner.Err())
+	}
+}
+
+func cmpResult(t *testing.T, tc *Test, name string, f Form, gold, test, result string) {
+	if gold != result {
+		t.Errorf("%s:%s: %s(%+q)=%+q; want %+q: %s",
+			tc.Name(), name, fstr[f], test, result, gold, tc.name)
+	}
+}
+
+func cmpIsNormal(t *testing.T, tc *Test, name string, f Form, test string, result, want bool) {
+	if result != want {
+		t.Errorf("%s:%s: %s(%+q)=%v; want %v", tc.Name(), name, fstr[f], test, result, want)
+	}
+}
+
+func doTest(t *testing.T, tc *Test, f Form, gold, test string) {
+	testb := []byte(test)
+	result := f.Bytes(testb)
+	cmpResult(t, tc, "Bytes", f, gold, test, string(result))
+
+	sresult := f.String(test)
+	cmpResult(t, tc, "String", f, gold, test, sresult)
+
+	acc := []byte{}
+	i := Iter{}
+	i.InitString(f, test)
+	for !i.Done() {
+		acc = append(acc, i.Next()...)
+	}
+	cmpResult(t, tc, "Iter.Next", f, gold, test, string(acc))
+
+	buf := make([]byte, 128)
+	acc = nil
+	for p := 0; p < len(testb); {
+		nDst, nSrc, _ := f.Transform(buf, testb[p:], true)
+		acc = append(acc, buf[:nDst]...)
+		p += nSrc
+	}
+	cmpResult(t, tc, "Transform", f, gold, test, string(acc))
+
+	for i := range test {
+		out := f.Append(f.Bytes([]byte(test[:i])), []byte(test[i:])...)
+		cmpResult(t, tc, fmt.Sprintf(":Append:%d", i), f, gold, test, string(out))
+	}
+	cmpIsNormal(t, tc, "IsNormal", f, test, f.IsNormal([]byte(test)), test == gold)
+	cmpIsNormal(t, tc, "IsNormalString", f, test, f.IsNormalString(test), test == gold)
+}
+
+func doConformanceTests(t *testing.T, tc *Test, partn int) {
+	for i := 0; i <= 2; i++ {
+		doTest(t, tc, NFC, tc.cols[1], tc.cols[i])
+		doTest(t, tc, NFD, tc.cols[2], tc.cols[i])
+		doTest(t, tc, NFKC, tc.cols[3], tc.cols[i])
+		doTest(t, tc, NFKD, tc.cols[4], tc.cols[i])
+	}
+	for i := 3; i <= 4; i++ {
+		doTest(t, tc, NFC, tc.cols[3], tc.cols[i])
+		doTest(t, tc, NFD, tc.cols[4], tc.cols[i])
+		doTest(t, tc, NFKC, tc.cols[3], tc.cols[i])
+		doTest(t, tc, NFKD, tc.cols[4], tc.cols[i])
+	}
+}
+
+func TestCharacterByCharacter(t *testing.T) {
+	skipShort(t)
+	tests := part[1].tests
+	var last rune = 0
+	for i := 0; i <= len(tests); i++ { // last one is special case
+		var r rune
+		if i == len(tests) {
+			r = 0x2FA1E // Don't have to go to 0x10FFFF
+		} else {
+			r = tests[i].r
+		}
+		for last++; last < r; last++ {
+			// Check all characters that were not explicitly listed in the test.
+			tc := &Test{partnr: 1, number: -1}
+			char := string(last)
+			doTest(t, tc, NFC, char, char)
+			doTest(t, tc, NFD, char, char)
+			doTest(t, tc, NFKC, char, char)
+			doTest(t, tc, NFKD, char, char)
+		}
+		if i < len(tests) {
+			doConformanceTests(t, &tests[i], 1)
+		}
+	}
+}
+
+func TestStandardTests(t *testing.T) {
+	skipShort(t)
+	for _, j := range []int{0, 2, 3} {
+		for _, test := range part[j].tests {
+			doConformanceTests(t, &test, j)
+		}
+	}
+}
+
+// TestPerformance verifies that normalization is O(n). If any of the
+// code does not properly check for maxCombiningChars, normalization
+// may exhibit O(n**2) behavior.
+func TestPerformance(t *testing.T) {
+	skipShort(t)
+	runtime.GOMAXPROCS(2)
+	success := make(chan bool, 1)
+	go func() {
+		buf := bytes.Repeat([]byte("\u035D"), 1024*1024)
+		buf = append(buf, "\u035B"...)
+		NFC.Append(nil, buf...)
+		success <- true
+	}()
+	timeout := time.After(1 * time.Second)
+	select {
+	case <-success:
+		// test completed before the timeout
+	case <-timeout:
+		t.Errorf(`unexpectedly long time to complete PerformanceTest`)
+	}
+}