Update go dependencies (#4524)

This commit is contained in:
Manuel Alejandro de Brito Fontes 2019-09-02 21:30:28 -04:00 committed by GitHub
parent 2ba1a9e71a
commit 341d64b652
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
412 changed files with 43034 additions and 34165 deletions

712
vendor/golang.org/x/net/html/atom/gen.go generated vendored Normal file
View file

@ -0,0 +1,712 @@
// Copyright 2012 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
//go:generate go run gen.go
//go:generate go run gen.go -test
package main
import (
"bytes"
"flag"
"fmt"
"go/format"
"io/ioutil"
"math/rand"
"os"
"sort"
"strings"
)
// identifier converts s to a Go exported identifier.
// It converts "div" to "Div" and "accept-charset" to "AcceptCharset".
func identifier(s string) string {
b := make([]byte, 0, len(s))
cap := true
for _, c := range s {
if c == '-' {
cap = true
continue
}
if cap && 'a' <= c && c <= 'z' {
c -= 'a' - 'A'
}
cap = false
b = append(b, byte(c))
}
return string(b)
}
var test = flag.Bool("test", false, "generate table_test.go")
func genFile(name string, buf *bytes.Buffer) {
b, err := format.Source(buf.Bytes())
if err != nil {
fmt.Fprintln(os.Stderr, err)
os.Exit(1)
}
if err := ioutil.WriteFile(name, b, 0644); err != nil {
fmt.Fprintln(os.Stderr, err)
os.Exit(1)
}
}
func main() {
flag.Parse()
var all []string
all = append(all, elements...)
all = append(all, attributes...)
all = append(all, eventHandlers...)
all = append(all, extra...)
sort.Strings(all)
// uniq - lists have dups
w := 0
for _, s := range all {
if w == 0 || all[w-1] != s {
all[w] = s
w++
}
}
all = all[:w]
if *test {
var buf bytes.Buffer
fmt.Fprintln(&buf, "// Code generated by go generate gen.go; DO NOT EDIT.\n")
fmt.Fprintln(&buf, "//go:generate go run gen.go -test\n")
fmt.Fprintln(&buf, "package atom\n")
fmt.Fprintln(&buf, "var testAtomList = []string{")
for _, s := range all {
fmt.Fprintf(&buf, "\t%q,\n", s)
}
fmt.Fprintln(&buf, "}")
genFile("table_test.go", &buf)
return
}
// Find hash that minimizes table size.
var best *table
for i := 0; i < 1000000; i++ {
if best != nil && 1<<(best.k-1) < len(all) {
break
}
h := rand.Uint32()
for k := uint(0); k <= 16; k++ {
if best != nil && k >= best.k {
break
}
var t table
if t.init(h, k, all) {
best = &t
break
}
}
}
if best == nil {
fmt.Fprintf(os.Stderr, "failed to construct string table\n")
os.Exit(1)
}
// Lay out strings, using overlaps when possible.
layout := append([]string{}, all...)
// Remove strings that are substrings of other strings
for changed := true; changed; {
changed = false
for i, s := range layout {
if s == "" {
continue
}
for j, t := range layout {
if i != j && t != "" && strings.Contains(s, t) {
changed = true
layout[j] = ""
}
}
}
}
// Join strings where one suffix matches another prefix.
for {
// Find best i, j, k such that layout[i][len-k:] == layout[j][:k],
// maximizing overlap length k.
besti := -1
bestj := -1
bestk := 0
for i, s := range layout {
if s == "" {
continue
}
for j, t := range layout {
if i == j {
continue
}
for k := bestk + 1; k <= len(s) && k <= len(t); k++ {
if s[len(s)-k:] == t[:k] {
besti = i
bestj = j
bestk = k
}
}
}
}
if bestk > 0 {
layout[besti] += layout[bestj][bestk:]
layout[bestj] = ""
continue
}
break
}
text := strings.Join(layout, "")
atom := map[string]uint32{}
for _, s := range all {
off := strings.Index(text, s)
if off < 0 {
panic("lost string " + s)
}
atom[s] = uint32(off<<8 | len(s))
}
var buf bytes.Buffer
// Generate the Go code.
fmt.Fprintln(&buf, "// Code generated by go generate gen.go; DO NOT EDIT.\n")
fmt.Fprintln(&buf, "//go:generate go run gen.go\n")
fmt.Fprintln(&buf, "package atom\n\nconst (")
// compute max len
maxLen := 0
for _, s := range all {
if maxLen < len(s) {
maxLen = len(s)
}
fmt.Fprintf(&buf, "\t%s Atom = %#x\n", identifier(s), atom[s])
}
fmt.Fprintln(&buf, ")\n")
fmt.Fprintf(&buf, "const hash0 = %#x\n\n", best.h0)
fmt.Fprintf(&buf, "const maxAtomLen = %d\n\n", maxLen)
fmt.Fprintf(&buf, "var table = [1<<%d]Atom{\n", best.k)
for i, s := range best.tab {
if s == "" {
continue
}
fmt.Fprintf(&buf, "\t%#x: %#x, // %s\n", i, atom[s], s)
}
fmt.Fprintf(&buf, "}\n")
datasize := (1 << best.k) * 4
fmt.Fprintln(&buf, "const atomText =")
textsize := len(text)
for len(text) > 60 {
fmt.Fprintf(&buf, "\t%q +\n", text[:60])
text = text[60:]
}
fmt.Fprintf(&buf, "\t%q\n\n", text)
genFile("table.go", &buf)
fmt.Fprintf(os.Stdout, "%d atoms; %d string bytes + %d tables = %d total data\n", len(all), textsize, datasize, textsize+datasize)
}
type byLen []string
func (x byLen) Less(i, j int) bool { return len(x[i]) > len(x[j]) }
func (x byLen) Swap(i, j int) { x[i], x[j] = x[j], x[i] }
func (x byLen) Len() int { return len(x) }
// fnv computes the FNV hash with an arbitrary starting value h.
func fnv(h uint32, s string) uint32 {
for i := 0; i < len(s); i++ {
h ^= uint32(s[i])
h *= 16777619
}
return h
}
// A table represents an attempt at constructing the lookup table.
// The lookup table uses cuckoo hashing, meaning that each string
// can be found in one of two positions.
type table struct {
h0 uint32
k uint
mask uint32
tab []string
}
// hash returns the two hashes for s.
func (t *table) hash(s string) (h1, h2 uint32) {
h := fnv(t.h0, s)
h1 = h & t.mask
h2 = (h >> 16) & t.mask
return
}
// init initializes the table with the given parameters.
// h0 is the initial hash value,
// k is the number of bits of hash value to use, and
// x is the list of strings to store in the table.
// init returns false if the table cannot be constructed.
func (t *table) init(h0 uint32, k uint, x []string) bool {
t.h0 = h0
t.k = k
t.tab = make([]string, 1<<k)
t.mask = 1<<k - 1
for _, s := range x {
if !t.insert(s) {
return false
}
}
return true
}
// insert inserts s in the table.
func (t *table) insert(s string) bool {
h1, h2 := t.hash(s)
if t.tab[h1] == "" {
t.tab[h1] = s
return true
}
if t.tab[h2] == "" {
t.tab[h2] = s
return true
}
if t.push(h1, 0) {
t.tab[h1] = s
return true
}
if t.push(h2, 0) {
t.tab[h2] = s
return true
}
return false
}
// push attempts to push aside the entry in slot i.
func (t *table) push(i uint32, depth int) bool {
if depth > len(t.tab) {
return false
}
s := t.tab[i]
h1, h2 := t.hash(s)
j := h1 + h2 - i
if t.tab[j] != "" && !t.push(j, depth+1) {
return false
}
t.tab[j] = s
return true
}
// The lists of element names and attribute keys were taken from
// https://html.spec.whatwg.org/multipage/indices.html#index
// as of the "HTML Living Standard - Last Updated 16 April 2018" version.
// "command", "keygen" and "menuitem" have been removed from the spec,
// but are kept here for backwards compatibility.
var elements = []string{
"a",
"abbr",
"address",
"area",
"article",
"aside",
"audio",
"b",
"base",
"bdi",
"bdo",
"blockquote",
"body",
"br",
"button",
"canvas",
"caption",
"cite",
"code",
"col",
"colgroup",
"command",
"data",
"datalist",
"dd",
"del",
"details",
"dfn",
"dialog",
"div",
"dl",
"dt",
"em",
"embed",
"fieldset",
"figcaption",
"figure",
"footer",
"form",
"h1",
"h2",
"h3",
"h4",
"h5",
"h6",
"head",
"header",
"hgroup",
"hr",
"html",
"i",
"iframe",
"img",
"input",
"ins",
"kbd",
"keygen",
"label",
"legend",
"li",
"link",
"main",
"map",
"mark",
"menu",
"menuitem",
"meta",
"meter",
"nav",
"noscript",
"object",
"ol",
"optgroup",
"option",
"output",
"p",
"param",
"picture",
"pre",
"progress",
"q",
"rp",
"rt",
"ruby",
"s",
"samp",
"script",
"section",
"select",
"slot",
"small",
"source",
"span",
"strong",
"style",
"sub",
"summary",
"sup",
"table",
"tbody",
"td",
"template",
"textarea",
"tfoot",
"th",
"thead",
"time",
"title",
"tr",
"track",
"u",
"ul",
"var",
"video",
"wbr",
}
// https://html.spec.whatwg.org/multipage/indices.html#attributes-3
//
// "challenge", "command", "contextmenu", "dropzone", "icon", "keytype", "mediagroup",
// "radiogroup", "spellcheck", "scoped", "seamless", "sortable" and "sorted" have been removed from the spec,
// but are kept here for backwards compatibility.
var attributes = []string{
"abbr",
"accept",
"accept-charset",
"accesskey",
"action",
"allowfullscreen",
"allowpaymentrequest",
"allowusermedia",
"alt",
"as",
"async",
"autocomplete",
"autofocus",
"autoplay",
"challenge",
"charset",
"checked",
"cite",
"class",
"color",
"cols",
"colspan",
"command",
"content",
"contenteditable",
"contextmenu",
"controls",
"coords",
"crossorigin",
"data",
"datetime",
"default",
"defer",
"dir",
"dirname",
"disabled",
"download",
"draggable",
"dropzone",
"enctype",
"for",
"form",
"formaction",
"formenctype",
"formmethod",
"formnovalidate",
"formtarget",
"headers",
"height",
"hidden",
"high",
"href",
"hreflang",
"http-equiv",
"icon",
"id",
"inputmode",
"integrity",
"is",
"ismap",
"itemid",
"itemprop",
"itemref",
"itemscope",
"itemtype",
"keytype",
"kind",
"label",
"lang",
"list",
"loop",
"low",
"manifest",
"max",
"maxlength",
"media",
"mediagroup",
"method",
"min",
"minlength",
"multiple",
"muted",
"name",
"nomodule",
"nonce",
"novalidate",
"open",
"optimum",
"pattern",
"ping",
"placeholder",
"playsinline",
"poster",
"preload",
"radiogroup",
"readonly",
"referrerpolicy",
"rel",
"required",
"reversed",
"rows",
"rowspan",
"sandbox",
"spellcheck",
"scope",
"scoped",
"seamless",
"selected",
"shape",
"size",
"sizes",
"sortable",
"sorted",
"slot",
"span",
"spellcheck",
"src",
"srcdoc",
"srclang",
"srcset",
"start",
"step",
"style",
"tabindex",
"target",
"title",
"translate",
"type",
"typemustmatch",
"updateviacache",
"usemap",
"value",
"width",
"workertype",
"wrap",
}
// "onautocomplete", "onautocompleteerror", "onmousewheel",
// "onshow" and "onsort" have been removed from the spec,
// but are kept here for backwards compatibility.
var eventHandlers = []string{
"onabort",
"onautocomplete",
"onautocompleteerror",
"onauxclick",
"onafterprint",
"onbeforeprint",
"onbeforeunload",
"onblur",
"oncancel",
"oncanplay",
"oncanplaythrough",
"onchange",
"onclick",
"onclose",
"oncontextmenu",
"oncopy",
"oncuechange",
"oncut",
"ondblclick",
"ondrag",
"ondragend",
"ondragenter",
"ondragexit",
"ondragleave",
"ondragover",
"ondragstart",
"ondrop",
"ondurationchange",
"onemptied",
"onended",
"onerror",
"onfocus",
"onhashchange",
"oninput",
"oninvalid",
"onkeydown",
"onkeypress",
"onkeyup",
"onlanguagechange",
"onload",
"onloadeddata",
"onloadedmetadata",
"onloadend",
"onloadstart",
"onmessage",
"onmessageerror",
"onmousedown",
"onmouseenter",
"onmouseleave",
"onmousemove",
"onmouseout",
"onmouseover",
"onmouseup",
"onmousewheel",
"onwheel",
"onoffline",
"ononline",
"onpagehide",
"onpageshow",
"onpaste",
"onpause",
"onplay",
"onplaying",
"onpopstate",
"onprogress",
"onratechange",
"onreset",
"onresize",
"onrejectionhandled",
"onscroll",
"onsecuritypolicyviolation",
"onseeked",
"onseeking",
"onselect",
"onshow",
"onsort",
"onstalled",
"onstorage",
"onsubmit",
"onsuspend",
"ontimeupdate",
"ontoggle",
"onunhandledrejection",
"onunload",
"onvolumechange",
"onwaiting",
}
// extra are ad-hoc values not covered by any of the lists above.
var extra = []string{
"acronym",
"align",
"annotation",
"annotation-xml",
"applet",
"basefont",
"bgsound",
"big",
"blink",
"center",
"color",
"desc",
"face",
"font",
"foreignObject", // HTML is case-insensitive, but SVG-embedded-in-HTML is case-sensitive.
"foreignobject",
"frame",
"frameset",
"image",
"isindex",
"listing",
"malignmark",
"marquee",
"math",
"mglyph",
"mi",
"mn",
"mo",
"ms",
"mtext",
"nobr",
"noembed",
"noframes",
"plaintext",
"prompt",
"public",
"rb",
"rtc",
"spacer",
"strike",
"svg",
"system",
"tt",
"xmp",
}

130
vendor/golang.org/x/net/html/parse.go generated vendored
View file

@ -439,6 +439,9 @@ func (p *parser) resetInsertionMode() {
case a.Select:
if !last {
for ancestor, first := n, p.oe[0]; ancestor != first; {
if ancestor == first {
break
}
ancestor = p.oe[p.oe.index(ancestor)-1]
switch ancestor.DataAtom {
case a.Template:
@ -630,16 +633,7 @@ func inHeadIM(p *parser) bool {
p.oe.pop()
p.acknowledgeSelfClosingTag()
return true
case a.Noscript:
p.addElement()
if p.scripting {
p.setOriginalIM()
p.im = textIM
} else {
p.im = inHeadNoscriptIM
}
return true
case a.Script, a.Title, a.Noframes, a.Style:
case a.Script, a.Title, a.Noscript, a.Noframes, a.Style:
p.addElement()
p.setOriginalIM()
p.im = textIM
@ -701,49 +695,6 @@ func inHeadIM(p *parser) bool {
return false
}
// 12.2.6.4.5.
func inHeadNoscriptIM(p *parser) bool {
switch p.tok.Type {
case DoctypeToken:
// Ignore the token.
return true
case StartTagToken:
switch p.tok.DataAtom {
case a.Html:
return inBodyIM(p)
case a.Basefont, a.Bgsound, a.Link, a.Meta, a.Noframes, a.Style:
return inHeadIM(p)
case a.Head, a.Noscript:
// Ignore the token.
return true
}
case EndTagToken:
switch p.tok.DataAtom {
case a.Noscript, a.Br:
default:
// Ignore the token.
return true
}
case TextToken:
s := strings.TrimLeft(p.tok.Data, whitespace)
if len(s) == 0 {
// It was all whitespace.
return inHeadIM(p)
}
case CommentToken:
return inHeadIM(p)
}
p.oe.pop()
if p.top().DataAtom != a.Head {
panic("html: the new current node will be a head element.")
}
p.im = inHeadIM
if p.tok.DataAtom == a.Noscript {
return true
}
return false
}
// Section 12.2.6.4.6.
func afterHeadIM(p *parser) bool {
switch p.tok.Type {
@ -953,7 +904,7 @@ func inBodyIM(p *parser) bool {
case a.A:
for i := len(p.afe) - 1; i >= 0 && p.afe[i].Type != scopeMarkerNode; i-- {
if n := p.afe[i]; n.Type == ElementNode && n.DataAtom == a.A {
p.inBodyEndTagFormatting(a.A, "a")
p.inBodyEndTagFormatting(a.A)
p.oe.remove(n)
p.afe.remove(n)
break
@ -967,7 +918,7 @@ func inBodyIM(p *parser) bool {
case a.Nobr:
p.reconstructActiveFormattingElements()
if p.elementInScope(defaultScope, a.Nobr) {
p.inBodyEndTagFormatting(a.Nobr, "nobr")
p.inBodyEndTagFormatting(a.Nobr)
p.reconstructActiveFormattingElements()
}
p.addFormattingElement()
@ -1175,7 +1126,7 @@ func inBodyIM(p *parser) bool {
case a.H1, a.H2, a.H3, a.H4, a.H5, a.H6:
p.popUntil(defaultScope, a.H1, a.H2, a.H3, a.H4, a.H5, a.H6)
case a.A, a.B, a.Big, a.Code, a.Em, a.Font, a.I, a.Nobr, a.S, a.Small, a.Strike, a.Strong, a.Tt, a.U:
p.inBodyEndTagFormatting(p.tok.DataAtom, p.tok.Data)
p.inBodyEndTagFormatting(p.tok.DataAtom)
case a.Applet, a.Marquee, a.Object:
if p.popUntil(defaultScope, p.tok.DataAtom) {
p.clearActiveFormattingElements()
@ -1186,7 +1137,7 @@ func inBodyIM(p *parser) bool {
case a.Template:
return inHeadIM(p)
default:
p.inBodyEndTagOther(p.tok.DataAtom, p.tok.Data)
p.inBodyEndTagOther(p.tok.DataAtom)
}
case CommentToken:
p.addChild(&Node{
@ -1213,7 +1164,7 @@ func inBodyIM(p *parser) bool {
return true
}
func (p *parser) inBodyEndTagFormatting(tagAtom a.Atom, tagName string) {
func (p *parser) inBodyEndTagFormatting(tagAtom a.Atom) {
// This is the "adoption agency" algorithm, described at
// https://html.spec.whatwg.org/multipage/syntax.html#adoptionAgency
@ -1235,7 +1186,7 @@ func (p *parser) inBodyEndTagFormatting(tagAtom a.Atom, tagName string) {
}
}
if formattingElement == nil {
p.inBodyEndTagOther(tagAtom, tagName)
p.inBodyEndTagOther(tagAtom)
return
}
feIndex := p.oe.index(formattingElement)
@ -1340,17 +1291,9 @@ func (p *parser) inBodyEndTagFormatting(tagAtom a.Atom, tagName string) {
// inBodyEndTagOther performs the "any other end tag" algorithm for inBodyIM.
// "Any other end tag" handling from 12.2.6.5 The rules for parsing tokens in foreign content
// https://html.spec.whatwg.org/multipage/syntax.html#parsing-main-inforeign
func (p *parser) inBodyEndTagOther(tagAtom a.Atom, tagName string) {
func (p *parser) inBodyEndTagOther(tagAtom a.Atom) {
for i := len(p.oe) - 1; i >= 0; i-- {
// Two element nodes have the same tag if they have the same Data (a
// string-typed field). As an optimization, for common HTML tags, each
// Data string is assigned a unique, non-zero DataAtom (a uint32-typed
// field), since integer comparison is faster than string comparison.
// Uncommon (custom) tags get a zero DataAtom.
//
// The if condition here is equivalent to (p.oe[i].Data == tagName).
if (p.oe[i].DataAtom == tagAtom) &&
((tagAtom != 0) || (p.oe[i].Data == tagName)) {
if p.oe[i].DataAtom == tagAtom {
p.oe = p.oe[:i]
break
}
@ -1744,9 +1687,8 @@ func inCellIM(p *parser) bool {
return true
}
// Close the cell and reprocess.
if p.popUntil(tableScope, a.Td, a.Th) {
p.clearActiveFormattingElements()
}
p.popUntil(tableScope, a.Td, a.Th)
p.clearActiveFormattingElements()
p.im = inRowIM
return false
}
@ -2300,33 +2242,6 @@ func (p *parser) parse() error {
//
// The input is assumed to be UTF-8 encoded.
func Parse(r io.Reader) (*Node, error) {
return ParseWithOptions(r)
}
// ParseFragment parses a fragment of HTML and returns the nodes that were
// found. If the fragment is the InnerHTML for an existing element, pass that
// element in context.
//
// It has the same intricacies as Parse.
func ParseFragment(r io.Reader, context *Node) ([]*Node, error) {
return ParseFragmentWithOptions(r, context)
}
// ParseOption configures a parser.
type ParseOption func(p *parser)
// ParseOptionEnableScripting configures the scripting flag.
// https://html.spec.whatwg.org/multipage/webappapis.html#enabling-and-disabling-scripting
//
// By default, scripting is enabled.
func ParseOptionEnableScripting(enable bool) ParseOption {
return func(p *parser) {
p.scripting = enable
}
}
// ParseWithOptions is like Parse, with options.
func ParseWithOptions(r io.Reader, opts ...ParseOption) (*Node, error) {
p := &parser{
tokenizer: NewTokenizer(r),
doc: &Node{
@ -2336,11 +2251,6 @@ func ParseWithOptions(r io.Reader, opts ...ParseOption) (*Node, error) {
framesetOK: true,
im: initialIM,
}
for _, f := range opts {
f(p)
}
err := p.parse()
if err != nil {
return nil, err
@ -2348,8 +2258,12 @@ func ParseWithOptions(r io.Reader, opts ...ParseOption) (*Node, error) {
return p.doc, nil
}
// ParseFragmentWithOptions is like ParseFragment, with options.
func ParseFragmentWithOptions(r io.Reader, context *Node, opts ...ParseOption) ([]*Node, error) {
// ParseFragment parses a fragment of HTML and returns the nodes that were
// found. If the fragment is the InnerHTML for an existing element, pass that
// element in context.
//
// It has the same intricacies as Parse.
func ParseFragment(r io.Reader, context *Node) ([]*Node, error) {
contextTag := ""
if context != nil {
if context.Type != ElementNode {
@ -2373,10 +2287,6 @@ func ParseFragmentWithOptions(r io.Reader, context *Node, opts ...ParseOption) (
context: context,
}
for _, f := range opts {
f(p)
}
root := &Node{
Type: ElementNode,
DataAtom: a.Html,

View file

@ -643,7 +643,7 @@ func (f *Framer) WriteData(streamID uint32, endStream bool, data []byte) error {
return f.WriteDataPadded(streamID, endStream, data, nil)
}
// WriteDataPadded writes a DATA frame with optional padding.
// WriteData writes a DATA frame with optional padding.
//
// If pad is nil, the padding bit is not sent.
// The length of pad must not exceed 255 bytes.

View file

@ -52,10 +52,11 @@ import (
)
const (
prefaceTimeout = 10 * time.Second
firstSettingsTimeout = 2 * time.Second // should be in-flight with preface anyway
handlerChunkWriteSize = 4 << 10
defaultMaxStreams = 250 // TODO: make this 100 as the GFE seems to?
prefaceTimeout = 10 * time.Second
firstSettingsTimeout = 2 * time.Second // should be in-flight with preface anyway
handlerChunkWriteSize = 4 << 10
defaultMaxStreams = 250 // TODO: make this 100 as the GFE seems to?
maxQueuedControlFrames = 10000
)
var (
@ -163,6 +164,15 @@ func (s *Server) maxConcurrentStreams() uint32 {
return defaultMaxStreams
}
// maxQueuedControlFrames is the maximum number of control frames like
// SETTINGS, PING and RST_STREAM that will be queued for writing before
// the connection is closed to prevent memory exhaustion attacks.
func (s *Server) maxQueuedControlFrames() int {
// TODO: if anybody asks, add a Server field, and remember to define the
// behavior of negative values.
return maxQueuedControlFrames
}
type serverInternalState struct {
mu sync.Mutex
activeConns map[*serverConn]struct{}
@ -273,20 +283,7 @@ func ConfigureServer(s *http.Server, conf *Server) error {
if testHookOnConn != nil {
testHookOnConn()
}
// The TLSNextProto interface predates contexts, so
// the net/http package passes down its per-connection
// base context via an exported but unadvertised
// method on the Handler. This is for internal
// net/http<=>http2 use only.
var ctx context.Context
type baseContexter interface {
BaseContext() context.Context
}
if bc, ok := h.(baseContexter); ok {
ctx = bc.BaseContext()
}
conf.ServeConn(c, &ServeConnOpts{
Context: ctx,
Handler: h,
BaseConfig: hs,
})
@ -297,10 +294,6 @@ func ConfigureServer(s *http.Server, conf *Server) error {
// ServeConnOpts are options for the Server.ServeConn method.
type ServeConnOpts struct {
// Context is the base context to use.
// If nil, context.Background is used.
Context context.Context
// BaseConfig optionally sets the base configuration
// for values. If nil, defaults are used.
BaseConfig *http.Server
@ -311,13 +304,6 @@ type ServeConnOpts struct {
Handler http.Handler
}
func (o *ServeConnOpts) context() context.Context {
if o.Context != nil {
return o.Context
}
return context.Background()
}
func (o *ServeConnOpts) baseConfig() *http.Server {
if o != nil && o.BaseConfig != nil {
return o.BaseConfig
@ -463,7 +449,7 @@ func (s *Server) ServeConn(c net.Conn, opts *ServeConnOpts) {
}
func serverConnBaseContext(c net.Conn, opts *ServeConnOpts) (ctx context.Context, cancel func()) {
ctx, cancel = context.WithCancel(opts.context())
ctx, cancel = context.WithCancel(context.Background())
ctx = context.WithValue(ctx, http.LocalAddrContextKey, c.LocalAddr())
if hs := opts.baseConfig(); hs != nil {
ctx = context.WithValue(ctx, http.ServerContextKey, hs)
@ -506,6 +492,7 @@ type serverConn struct {
sawFirstSettings bool // got the initial SETTINGS frame after the preface
needToSendSettingsAck bool
unackedSettings int // how many SETTINGS have we sent without ACKs?
queuedControlFrames int // control frames in the writeSched queue
clientMaxStreams uint32 // SETTINGS_MAX_CONCURRENT_STREAMS from client (our PUSH_PROMISE limit)
advMaxStreams uint32 // our SETTINGS_MAX_CONCURRENT_STREAMS advertised the client
curClientStreams uint32 // number of open streams initiated by the client
@ -894,6 +881,14 @@ func (sc *serverConn) serve() {
}
}
// If the peer is causing us to generate a lot of control frames,
// but not reading them from us, assume they are trying to make us
// run out of memory.
if sc.queuedControlFrames > sc.srv.maxQueuedControlFrames() {
sc.vlogf("http2: too many control frames in send queue, closing connection")
return
}
// Start the shutdown timer after sending a GOAWAY. When sending GOAWAY
// with no error code (graceful shutdown), don't start the timer until
// all open streams have been completed.
@ -1093,6 +1088,14 @@ func (sc *serverConn) writeFrame(wr FrameWriteRequest) {
}
if !ignoreWrite {
if wr.isControl() {
sc.queuedControlFrames++
// For extra safety, detect wraparounds, which should not happen,
// and pull the plug.
if sc.queuedControlFrames < 0 {
sc.conn.Close()
}
}
sc.writeSched.Push(wr)
}
sc.scheduleFrameWrite()
@ -1210,10 +1213,8 @@ func (sc *serverConn) wroteFrame(res frameWriteResult) {
// If a frame is already being written, nothing happens. This will be called again
// when the frame is done being written.
//
// If a frame isn't being written we need to send one, the best frame
// to send is selected, preferring first things that aren't
// stream-specific (e.g. ACKing settings), and then finding the
// highest priority stream.
// If a frame isn't being written and we need to send one, the best frame
// to send is selected by writeSched.
//
// If a frame isn't being written and there's nothing else to send, we
// flush the write buffer.
@ -1241,6 +1242,9 @@ func (sc *serverConn) scheduleFrameWrite() {
}
if !sc.inGoAway || sc.goAwayCode == ErrCodeNo {
if wr, ok := sc.writeSched.Pop(); ok {
if wr.isControl() {
sc.queuedControlFrames--
}
sc.startFrameWrite(wr)
continue
}
@ -1533,6 +1537,8 @@ func (sc *serverConn) processSettings(f *SettingsFrame) error {
if err := f.ForeachSetting(sc.processSetting); err != nil {
return err
}
// TODO: judging by RFC 7540, Section 6.5.3 each SETTINGS frame should be
// acknowledged individually, even if multiple are received before the ACK.
sc.needToSendSettingsAck = true
sc.scheduleFrameWrite()
return nil
@ -2331,16 +2337,7 @@ type chunkWriter struct{ rws *responseWriterState }
func (cw chunkWriter) Write(p []byte) (n int, err error) { return cw.rws.writeChunk(p) }
func (rws *responseWriterState) hasTrailers() bool { return len(rws.trailers) > 0 }
func (rws *responseWriterState) hasNonemptyTrailers() bool {
for _, trailer := range rws.trailers {
if _, ok := rws.handlerHeader[trailer]; ok {
return true
}
}
return false
}
func (rws *responseWriterState) hasTrailers() bool { return len(rws.trailers) != 0 }
// declareTrailer is called for each Trailer header when the
// response header is written. It notes that a header will need to be
@ -2440,10 +2437,7 @@ func (rws *responseWriterState) writeChunk(p []byte) (n int, err error) {
rws.promoteUndeclaredTrailers()
}
// only send trailers if they have actually been defined by the
// server handler.
hasNonemptyTrailers := rws.hasNonemptyTrailers()
endStream := rws.handlerDone && !hasNonemptyTrailers
endStream := rws.handlerDone && !rws.hasTrailers()
if len(p) > 0 || endStream {
// only send a 0 byte DATA frame if we're ending the stream.
if err := rws.conn.writeDataFromHandler(rws.stream, p, endStream); err != nil {
@ -2452,7 +2446,7 @@ func (rws *responseWriterState) writeChunk(p []byte) (n int, err error) {
}
}
if rws.handlerDone && hasNonemptyTrailers {
if rws.handlerDone && rws.hasTrailers() {
err = rws.conn.writeHeaders(rws.stream, &writeResHeaders{
streamID: rws.stream.id,
h: rws.handlerHeader,

View file

@ -28,7 +28,6 @@ import (
"strconv"
"strings"
"sync"
"sync/atomic"
"time"
"golang.org/x/net/http/httpguts"
@ -200,7 +199,6 @@ type ClientConn struct {
t *Transport
tconn net.Conn // usually *tls.Conn, except specialized impls
tlsState *tls.ConnectionState // nil only for specialized impls
reused uint32 // whether conn is being reused; atomic
singleUse bool // whether being used for a single http.Request
// readLoop goroutine fields:
@ -442,8 +440,7 @@ func (t *Transport) RoundTripOpt(req *http.Request, opt RoundTripOpt) (*http.Res
t.vlogf("http2: Transport failed to get client conn for %s: %v", addr, err)
return nil, err
}
reused := !atomic.CompareAndSwapUint32(&cc.reused, 0, 1)
traceGotConn(req, cc, reused)
traceGotConn(req, cc)
res, gotErrAfterReqBodyWrite, err := cc.roundTrip(req)
if err != nil && retry <= 6 {
if req, err = shouldRetryRequest(req, err, gotErrAfterReqBodyWrite); err == nil {
@ -1414,11 +1411,7 @@ func (cc *ClientConn) encodeHeaders(req *http.Request, addGzipHeader bool, trail
// followed by the query production (see Sections 3.3 and 3.4 of
// [RFC3986]).
f(":authority", host)
m := req.Method
if m == "" {
m = http.MethodGet
}
f(":method", m)
f(":method", req.Method)
if req.Method != "CONNECT" {
f(":path", path)
f(":scheme", req.URL.Scheme)
@ -2562,15 +2555,15 @@ func traceGetConn(req *http.Request, hostPort string) {
trace.GetConn(hostPort)
}
func traceGotConn(req *http.Request, cc *ClientConn, reused bool) {
func traceGotConn(req *http.Request, cc *ClientConn) {
trace := httptrace.ContextClientTrace(req.Context())
if trace == nil || trace.GotConn == nil {
return
}
ci := httptrace.GotConnInfo{Conn: cc.tconn}
ci.Reused = reused
cc.mu.Lock()
ci.WasIdle = len(cc.streams) == 0 && reused
ci.Reused = cc.nextStreamID > 1
ci.WasIdle = len(cc.streams) == 0 && ci.Reused
if ci.WasIdle && !cc.lastActive.IsZero() {
ci.IdleTime = time.Now().Sub(cc.lastActive)
}

View file

@ -32,7 +32,7 @@ type WriteScheduler interface {
// Pop dequeues the next frame to write. Returns false if no frames can
// be written. Frames with a given wr.StreamID() are Pop'd in the same
// order they are Push'd.
// order they are Push'd. No frames should be discarded except by CloseStream.
Pop() (wr FrameWriteRequest, ok bool)
}
@ -76,6 +76,12 @@ func (wr FrameWriteRequest) StreamID() uint32 {
return wr.stream.id
}
// isControl reports whether wr is a control frame for MaxQueuedControlFrames
// purposes. That includes non-stream frames and RST_STREAM frames.
func (wr FrameWriteRequest) isControl() bool {
return wr.stream == nil
}
// DataSize returns the number of flow control bytes that must be consumed
// to write this entire frame. This is 0 for non-DATA frames.
func (wr FrameWriteRequest) DataSize() int {

View file

@ -4,16 +4,14 @@
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// +build go1.10
// Package idna implements IDNA2008 using the compatibility processing
// defined by UTS (Unicode Technical Standard) #46, which defines a standard to
// deal with the transition from IDNA2003.
//
// IDNA2008 (Internationalized Domain Names for Applications), is defined in RFC
// 5890, RFC 5891, RFC 5892, RFC 5893 and RFC 5894.
// UTS #46 is defined in https://www.unicode.org/reports/tr46.
// See https://unicode.org/cldr/utility/idna.jsp for a visualization of the
// UTS #46 is defined in http://www.unicode.org/reports/tr46.
// See http://unicode.org/cldr/utility/idna.jsp for a visualization of the
// differences between these two standards.
package idna // import "golang.org/x/net/idna"
@ -299,7 +297,7 @@ func (e runeError) Error() string {
}
// process implements the algorithm described in section 4 of UTS #46,
// see https://www.unicode.org/reports/tr46.
// see http://www.unicode.org/reports/tr46.
func (p *Profile) process(s string, toASCII bool) (string, error) {
var err error
var isBidi bool

View file

@ -1,682 +0,0 @@
// Code generated by running "go generate" in golang.org/x/text. DO NOT EDIT.
// Copyright 2016 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 !go1.10
// Package idna implements IDNA2008 using the compatibility processing
// defined by UTS (Unicode Technical Standard) #46, which defines a standard to
// deal with the transition from IDNA2003.
//
// IDNA2008 (Internationalized Domain Names for Applications), is defined in RFC
// 5890, RFC 5891, RFC 5892, RFC 5893 and RFC 5894.
// UTS #46 is defined in https://www.unicode.org/reports/tr46.
// See https://unicode.org/cldr/utility/idna.jsp for a visualization of the
// differences between these two standards.
package idna // import "golang.org/x/net/idna"
import (
"fmt"
"strings"
"unicode/utf8"
"golang.org/x/text/secure/bidirule"
"golang.org/x/text/unicode/norm"
)
// NOTE: Unlike common practice in Go APIs, the functions will return a
// sanitized domain name in case of errors. Browsers sometimes use a partially
// evaluated string as lookup.
// TODO: the current error handling is, in my opinion, the least opinionated.
// Other strategies are also viable, though:
// Option 1) Return an empty string in case of error, but allow the user to
// specify explicitly which errors to ignore.
// Option 2) Return the partially evaluated string if it is itself a valid
// string, otherwise return the empty string in case of error.
// Option 3) Option 1 and 2.
// Option 4) Always return an empty string for now and implement Option 1 as
// needed, and document that the return string may not be empty in case of
// error in the future.
// I think Option 1 is best, but it is quite opinionated.
// ToASCII is a wrapper for Punycode.ToASCII.
func ToASCII(s string) (string, error) {
return Punycode.process(s, true)
}
// ToUnicode is a wrapper for Punycode.ToUnicode.
func ToUnicode(s string) (string, error) {
return Punycode.process(s, false)
}
// An Option configures a Profile at creation time.
type Option func(*options)
// Transitional sets a Profile to use the Transitional mapping as defined in UTS
// #46. This will cause, for example, "ß" to be mapped to "ss". Using the
// transitional mapping provides a compromise between IDNA2003 and IDNA2008
// compatibility. It is used by most browsers when resolving domain names. This
// option is only meaningful if combined with MapForLookup.
func Transitional(transitional bool) Option {
return func(o *options) { o.transitional = true }
}
// VerifyDNSLength sets whether a Profile should fail if any of the IDN parts
// are longer than allowed by the RFC.
func VerifyDNSLength(verify bool) Option {
return func(o *options) { o.verifyDNSLength = verify }
}
// RemoveLeadingDots removes leading label separators. Leading runes that map to
// dots, such as U+3002 IDEOGRAPHIC FULL STOP, are removed as well.
//
// This is the behavior suggested by the UTS #46 and is adopted by some
// browsers.
func RemoveLeadingDots(remove bool) Option {
return func(o *options) { o.removeLeadingDots = remove }
}
// ValidateLabels sets whether to check the mandatory label validation criteria
// as defined in Section 5.4 of RFC 5891. This includes testing for correct use
// of hyphens ('-'), normalization, validity of runes, and the context rules.
func ValidateLabels(enable bool) Option {
return func(o *options) {
// Don't override existing mappings, but set one that at least checks
// normalization if it is not set.
if o.mapping == nil && enable {
o.mapping = normalize
}
o.trie = trie
o.validateLabels = enable
o.fromPuny = validateFromPunycode
}
}
// StrictDomainName limits the set of permissable ASCII characters to those
// allowed in domain names as defined in RFC 1034 (A-Z, a-z, 0-9 and the
// hyphen). This is set by default for MapForLookup and ValidateForRegistration.
//
// This option is useful, for instance, for browsers that allow characters
// outside this range, for example a '_' (U+005F LOW LINE). See
// http://www.rfc-editor.org/std/std3.txt for more details This option
// corresponds to the UseSTD3ASCIIRules option in UTS #46.
func StrictDomainName(use bool) Option {
return func(o *options) {
o.trie = trie
o.useSTD3Rules = use
o.fromPuny = validateFromPunycode
}
}
// NOTE: the following options pull in tables. The tables should not be linked
// in as long as the options are not used.
// BidiRule enables the Bidi rule as defined in RFC 5893. Any application
// that relies on proper validation of labels should include this rule.
func BidiRule() Option {
return func(o *options) { o.bidirule = bidirule.ValidString }
}
// ValidateForRegistration sets validation options to verify that a given IDN is
// properly formatted for registration as defined by Section 4 of RFC 5891.
func ValidateForRegistration() Option {
return func(o *options) {
o.mapping = validateRegistration
StrictDomainName(true)(o)
ValidateLabels(true)(o)
VerifyDNSLength(true)(o)
BidiRule()(o)
}
}
// MapForLookup sets validation and mapping options such that a given IDN is
// transformed for domain name lookup according to the requirements set out in
// Section 5 of RFC 5891. The mappings follow the recommendations of RFC 5894,
// RFC 5895 and UTS 46. It does not add the Bidi Rule. Use the BidiRule option
// to add this check.
//
// The mappings include normalization and mapping case, width and other
// compatibility mappings.
func MapForLookup() Option {
return func(o *options) {
o.mapping = validateAndMap
StrictDomainName(true)(o)
ValidateLabels(true)(o)
RemoveLeadingDots(true)(o)
}
}
type options struct {
transitional bool
useSTD3Rules bool
validateLabels bool
verifyDNSLength bool
removeLeadingDots bool
trie *idnaTrie
// fromPuny calls validation rules when converting A-labels to U-labels.
fromPuny func(p *Profile, s string) error
// mapping implements a validation and mapping step as defined in RFC 5895
// or UTS 46, tailored to, for example, domain registration or lookup.
mapping func(p *Profile, s string) (string, error)
// bidirule, if specified, checks whether s conforms to the Bidi Rule
// defined in RFC 5893.
bidirule func(s string) bool
}
// A Profile defines the configuration of a IDNA mapper.
type Profile struct {
options
}
func apply(o *options, opts []Option) {
for _, f := range opts {
f(o)
}
}
// New creates a new Profile.
//
// With no options, the returned Profile is the most permissive and equals the
// Punycode Profile. Options can be passed to further restrict the Profile. The
// MapForLookup and ValidateForRegistration options set a collection of options,
// for lookup and registration purposes respectively, which can be tailored by
// adding more fine-grained options, where later options override earlier
// options.
func New(o ...Option) *Profile {
p := &Profile{}
apply(&p.options, o)
return p
}
// ToASCII converts a domain or domain label to its ASCII form. For example,
// ToASCII("bücher.example.com") is "xn--bcher-kva.example.com", and
// ToASCII("golang") is "golang". If an error is encountered it will return
// an error and a (partially) processed result.
func (p *Profile) ToASCII(s string) (string, error) {
return p.process(s, true)
}
// ToUnicode converts a domain or domain label to its Unicode form. For example,
// ToUnicode("xn--bcher-kva.example.com") is "bücher.example.com", and
// ToUnicode("golang") is "golang". If an error is encountered it will return
// an error and a (partially) processed result.
func (p *Profile) ToUnicode(s string) (string, error) {
pp := *p
pp.transitional = false
return pp.process(s, false)
}
// String reports a string with a description of the profile for debugging
// purposes. The string format may change with different versions.
func (p *Profile) String() string {
s := ""
if p.transitional {
s = "Transitional"
} else {
s = "NonTransitional"
}
if p.useSTD3Rules {
s += ":UseSTD3Rules"
}
if p.validateLabels {
s += ":ValidateLabels"
}
if p.verifyDNSLength {
s += ":VerifyDNSLength"
}
return s
}
var (
// Punycode is a Profile that does raw punycode processing with a minimum
// of validation.
Punycode *Profile = punycode
// Lookup is the recommended profile for looking up domain names, according
// to Section 5 of RFC 5891. The exact configuration of this profile may
// change over time.
Lookup *Profile = lookup
// Display is the recommended profile for displaying domain names.
// The configuration of this profile may change over time.
Display *Profile = display
// Registration is the recommended profile for checking whether a given
// IDN is valid for registration, according to Section 4 of RFC 5891.
Registration *Profile = registration
punycode = &Profile{}
lookup = &Profile{options{
transitional: true,
useSTD3Rules: true,
validateLabels: true,
removeLeadingDots: true,
trie: trie,
fromPuny: validateFromPunycode,
mapping: validateAndMap,
bidirule: bidirule.ValidString,
}}
display = &Profile{options{
useSTD3Rules: true,
validateLabels: true,
removeLeadingDots: true,
trie: trie,
fromPuny: validateFromPunycode,
mapping: validateAndMap,
bidirule: bidirule.ValidString,
}}
registration = &Profile{options{
useSTD3Rules: true,
validateLabels: true,
verifyDNSLength: true,
trie: trie,
fromPuny: validateFromPunycode,
mapping: validateRegistration,
bidirule: bidirule.ValidString,
}}
// TODO: profiles
// Register: recommended for approving domain names: don't do any mappings
// but rather reject on invalid input. Bundle or block deviation characters.
)
type labelError struct{ label, code_ string }
func (e labelError) code() string { return e.code_ }
func (e labelError) Error() string {
return fmt.Sprintf("idna: invalid label %q", e.label)
}
type runeError rune
func (e runeError) code() string { return "P1" }
func (e runeError) Error() string {
return fmt.Sprintf("idna: disallowed rune %U", e)
}
// process implements the algorithm described in section 4 of UTS #46,
// see https://www.unicode.org/reports/tr46.
func (p *Profile) process(s string, toASCII bool) (string, error) {
var err error
if p.mapping != nil {
s, err = p.mapping(p, s)
}
// Remove leading empty labels.
if p.removeLeadingDots {
for ; len(s) > 0 && s[0] == '.'; s = s[1:] {
}
}
// It seems like we should only create this error on ToASCII, but the
// UTS 46 conformance tests suggests we should always check this.
if err == nil && p.verifyDNSLength && s == "" {
err = &labelError{s, "A4"}
}
labels := labelIter{orig: s}
for ; !labels.done(); labels.next() {
label := labels.label()
if label == "" {
// Empty labels are not okay. The label iterator skips the last
// label if it is empty.
if err == nil && p.verifyDNSLength {
err = &labelError{s, "A4"}
}
continue
}
if strings.HasPrefix(label, acePrefix) {
u, err2 := decode(label[len(acePrefix):])
if err2 != nil {
if err == nil {
err = err2
}
// Spec says keep the old label.
continue
}
labels.set(u)
if err == nil && p.validateLabels {
err = p.fromPuny(p, u)
}
if err == nil {
// This should be called on NonTransitional, according to the
// spec, but that currently does not have any effect. Use the
// original profile to preserve options.
err = p.validateLabel(u)
}
} else if err == nil {
err = p.validateLabel(label)
}
}
if toASCII {
for labels.reset(); !labels.done(); labels.next() {
label := labels.label()
if !ascii(label) {
a, err2 := encode(acePrefix, label)
if err == nil {
err = err2
}
label = a
labels.set(a)
}
n := len(label)
if p.verifyDNSLength && err == nil && (n == 0 || n > 63) {
err = &labelError{label, "A4"}
}
}
}
s = labels.result()
if toASCII && p.verifyDNSLength && err == nil {
// Compute the length of the domain name minus the root label and its dot.
n := len(s)
if n > 0 && s[n-1] == '.' {
n--
}
if len(s) < 1 || n > 253 {
err = &labelError{s, "A4"}
}
}
return s, err
}
func normalize(p *Profile, s string) (string, error) {
return norm.NFC.String(s), nil
}
func validateRegistration(p *Profile, s string) (string, error) {
if !norm.NFC.IsNormalString(s) {
return s, &labelError{s, "V1"}
}
for i := 0; i < len(s); {
v, sz := trie.lookupString(s[i:])
// Copy bytes not copied so far.
switch p.simplify(info(v).category()) {
// TODO: handle the NV8 defined in the Unicode idna data set to allow
// for strict conformance to IDNA2008.
case valid, deviation:
case disallowed, mapped, unknown, ignored:
r, _ := utf8.DecodeRuneInString(s[i:])
return s, runeError(r)
}
i += sz
}
return s, nil
}
func validateAndMap(p *Profile, s string) (string, error) {
var (
err error
b []byte
k int
)
for i := 0; i < len(s); {
v, sz := trie.lookupString(s[i:])
start := i
i += sz
// Copy bytes not copied so far.
switch p.simplify(info(v).category()) {
case valid:
continue
case disallowed:
if err == nil {
r, _ := utf8.DecodeRuneInString(s[start:])
err = runeError(r)
}
continue
case mapped, deviation:
b = append(b, s[k:start]...)
b = info(v).appendMapping(b, s[start:i])
case ignored:
b = append(b, s[k:start]...)
// drop the rune
case unknown:
b = append(b, s[k:start]...)
b = append(b, "\ufffd"...)
}
k = i
}
if k == 0 {
// No changes so far.
s = norm.NFC.String(s)
} else {
b = append(b, s[k:]...)
if norm.NFC.QuickSpan(b) != len(b) {
b = norm.NFC.Bytes(b)
}
// TODO: the punycode converters require strings as input.
s = string(b)
}
return s, err
}
// A labelIter allows iterating over domain name labels.
type labelIter struct {
orig string
slice []string
curStart int
curEnd int
i int
}
func (l *labelIter) reset() {
l.curStart = 0
l.curEnd = 0
l.i = 0
}
func (l *labelIter) done() bool {
return l.curStart >= len(l.orig)
}
func (l *labelIter) result() string {
if l.slice != nil {
return strings.Join(l.slice, ".")
}
return l.orig
}
func (l *labelIter) label() string {
if l.slice != nil {
return l.slice[l.i]
}
p := strings.IndexByte(l.orig[l.curStart:], '.')
l.curEnd = l.curStart + p
if p == -1 {
l.curEnd = len(l.orig)
}
return l.orig[l.curStart:l.curEnd]
}
// next sets the value to the next label. It skips the last label if it is empty.
func (l *labelIter) next() {
l.i++
if l.slice != nil {
if l.i >= len(l.slice) || l.i == len(l.slice)-1 && l.slice[l.i] == "" {
l.curStart = len(l.orig)
}
} else {
l.curStart = l.curEnd + 1
if l.curStart == len(l.orig)-1 && l.orig[l.curStart] == '.' {
l.curStart = len(l.orig)
}
}
}
func (l *labelIter) set(s string) {
if l.slice == nil {
l.slice = strings.Split(l.orig, ".")
}
l.slice[l.i] = s
}
// acePrefix is the ASCII Compatible Encoding prefix.
const acePrefix = "xn--"
func (p *Profile) simplify(cat category) category {
switch cat {
case disallowedSTD3Mapped:
if p.useSTD3Rules {
cat = disallowed
} else {
cat = mapped
}
case disallowedSTD3Valid:
if p.useSTD3Rules {
cat = disallowed
} else {
cat = valid
}
case deviation:
if !p.transitional {
cat = valid
}
case validNV8, validXV8:
// TODO: handle V2008
cat = valid
}
return cat
}
func validateFromPunycode(p *Profile, s string) error {
if !norm.NFC.IsNormalString(s) {
return &labelError{s, "V1"}
}
for i := 0; i < len(s); {
v, sz := trie.lookupString(s[i:])
if c := p.simplify(info(v).category()); c != valid && c != deviation {
return &labelError{s, "V6"}
}
i += sz
}
return nil
}
const (
zwnj = "\u200c"
zwj = "\u200d"
)
type joinState int8
const (
stateStart joinState = iota
stateVirama
stateBefore
stateBeforeVirama
stateAfter
stateFAIL
)
var joinStates = [][numJoinTypes]joinState{
stateStart: {
joiningL: stateBefore,
joiningD: stateBefore,
joinZWNJ: stateFAIL,
joinZWJ: stateFAIL,
joinVirama: stateVirama,
},
stateVirama: {
joiningL: stateBefore,
joiningD: stateBefore,
},
stateBefore: {
joiningL: stateBefore,
joiningD: stateBefore,
joiningT: stateBefore,
joinZWNJ: stateAfter,
joinZWJ: stateFAIL,
joinVirama: stateBeforeVirama,
},
stateBeforeVirama: {
joiningL: stateBefore,
joiningD: stateBefore,
joiningT: stateBefore,
},
stateAfter: {
joiningL: stateFAIL,
joiningD: stateBefore,
joiningT: stateAfter,
joiningR: stateStart,
joinZWNJ: stateFAIL,
joinZWJ: stateFAIL,
joinVirama: stateAfter, // no-op as we can't accept joiners here
},
stateFAIL: {
0: stateFAIL,
joiningL: stateFAIL,
joiningD: stateFAIL,
joiningT: stateFAIL,
joiningR: stateFAIL,
joinZWNJ: stateFAIL,
joinZWJ: stateFAIL,
joinVirama: stateFAIL,
},
}
// validateLabel validates the criteria from Section 4.1. Item 1, 4, and 6 are
// already implicitly satisfied by the overall implementation.
func (p *Profile) validateLabel(s string) error {
if s == "" {
if p.verifyDNSLength {
return &labelError{s, "A4"}
}
return nil
}
if p.bidirule != nil && !p.bidirule(s) {
return &labelError{s, "B"}
}
if !p.validateLabels {
return nil
}
trie := p.trie // p.validateLabels is only set if trie is set.
if len(s) > 4 && s[2] == '-' && s[3] == '-' {
return &labelError{s, "V2"}
}
if s[0] == '-' || s[len(s)-1] == '-' {
return &labelError{s, "V3"}
}
// TODO: merge the use of this in the trie.
v, sz := trie.lookupString(s)
x := info(v)
if x.isModifier() {
return &labelError{s, "V5"}
}
// Quickly return in the absence of zero-width (non) joiners.
if strings.Index(s, zwj) == -1 && strings.Index(s, zwnj) == -1 {
return nil
}
st := stateStart
for i := 0; ; {
jt := x.joinType()
if s[i:i+sz] == zwj {
jt = joinZWJ
} else if s[i:i+sz] == zwnj {
jt = joinZWNJ
}
st = joinStates[st][jt]
if x.isViramaModifier() {
st = joinStates[st][joinVirama]
}
if i += sz; i == len(s) {
break
}
v, sz = trie.lookupString(s[i:])
x = info(v)
}
if st == stateFAIL || st == stateAfter {
return &labelError{s, "C"}
}
return nil
}
func ascii(s string) bool {
for i := 0; i < len(s); i++ {
if s[i] >= utf8.RuneSelf {
return false
}
}
return true
}

View file

@ -1,13 +1,11 @@
// Code generated by running "go generate" in golang.org/x/text. DO NOT EDIT.
// +build go1.10,!go1.13
package idna
// UnicodeVersion is the Unicode version from which the tables in this package are derived.
const UnicodeVersion = "10.0.0"
var mappings string = "" + // Size: 8175 bytes
var mappings string = "" + // Size: 8176 bytes
"\x00\x01 \x03 ̈\x01a\x03 ̄\x012\x013\x03 ́\x03 ̧\x011\x01o\x0514\x0512" +
"\x0534\x03i̇\x03l·\x03ʼn\x01s\x03dž\x03ⱥ\x03ⱦ\x01h\x01j\x01r\x01w\x01y" +
"\x03 ̆\x03 ̇\x03 ̊\x03 ̨\x03 ̃\x03 ̋\x01l\x01x\x04̈́\x03 ι\x01;\x05 ̈́" +
@ -4556,4 +4554,4 @@ var idnaSparseValues = [1915]valueRange{
{value: 0x0040, lo: 0xb0, hi: 0xbf},
}
// Total table size 42114 bytes (41KiB); checksum: 355A58A4
// Total table size 42115 bytes (41KiB); checksum: F4A1FA4E

File diff suppressed because it is too large Load diff

File diff suppressed because it is too large Load diff

717
vendor/golang.org/x/net/publicsuffix/gen.go generated vendored Normal file
View file

@ -0,0 +1,717 @@
// Copyright 2012 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
package main
// This program generates table.go and table_test.go based on the authoritative
// public suffix list at https://publicsuffix.org/list/effective_tld_names.dat
//
// The version is derived from
// https://api.github.com/repos/publicsuffix/list/commits?path=public_suffix_list.dat
// and a human-readable form is at
// https://github.com/publicsuffix/list/commits/master/public_suffix_list.dat
//
// To fetch a particular git revision, such as 5c70ccd250, pass
// -url "https://raw.githubusercontent.com/publicsuffix/list/5c70ccd250/public_suffix_list.dat"
// and -version "an explicit version string".
import (
"bufio"
"bytes"
"flag"
"fmt"
"go/format"
"io"
"io/ioutil"
"net/http"
"os"
"regexp"
"sort"
"strings"
"golang.org/x/net/idna"
)
const (
// These sum of these four values must be no greater than 32.
nodesBitsChildren = 10
nodesBitsICANN = 1
nodesBitsTextOffset = 15
nodesBitsTextLength = 6
// These sum of these four values must be no greater than 32.
childrenBitsWildcard = 1
childrenBitsNodeType = 2
childrenBitsHi = 14
childrenBitsLo = 14
)
var (
maxChildren int
maxTextOffset int
maxTextLength int
maxHi uint32
maxLo uint32
)
func max(a, b int) int {
if a < b {
return b
}
return a
}
func u32max(a, b uint32) uint32 {
if a < b {
return b
}
return a
}
const (
nodeTypeNormal = 0
nodeTypeException = 1
nodeTypeParentOnly = 2
numNodeType = 3
)
func nodeTypeStr(n int) string {
switch n {
case nodeTypeNormal:
return "+"
case nodeTypeException:
return "!"
case nodeTypeParentOnly:
return "o"
}
panic("unreachable")
}
const (
defaultURL = "https://publicsuffix.org/list/effective_tld_names.dat"
gitCommitURL = "https://api.github.com/repos/publicsuffix/list/commits?path=public_suffix_list.dat"
)
var (
labelEncoding = map[string]uint32{}
labelsList = []string{}
labelsMap = map[string]bool{}
rules = []string{}
numICANNRules = 0
// validSuffixRE is used to check that the entries in the public suffix
// list are in canonical form (after Punycode encoding). Specifically,
// capital letters are not allowed.
validSuffixRE = regexp.MustCompile(`^[a-z0-9_\!\*\-\.]+$`)
shaRE = regexp.MustCompile(`"sha":"([^"]+)"`)
dateRE = regexp.MustCompile(`"committer":{[^{]+"date":"([^"]+)"`)
comments = flag.Bool("comments", false, "generate table.go comments, for debugging")
subset = flag.Bool("subset", false, "generate only a subset of the full table, for debugging")
url = flag.String("url", defaultURL, "URL of the publicsuffix.org list. If empty, stdin is read instead")
v = flag.Bool("v", false, "verbose output (to stderr)")
version = flag.String("version", "", "the effective_tld_names.dat version")
)
func main() {
if err := main1(); err != nil {
fmt.Fprintln(os.Stderr, err)
os.Exit(1)
}
}
func main1() error {
flag.Parse()
if nodesBitsTextLength+nodesBitsTextOffset+nodesBitsICANN+nodesBitsChildren > 32 {
return fmt.Errorf("not enough bits to encode the nodes table")
}
if childrenBitsLo+childrenBitsHi+childrenBitsNodeType+childrenBitsWildcard > 32 {
return fmt.Errorf("not enough bits to encode the children table")
}
if *version == "" {
if *url != defaultURL {
return fmt.Errorf("-version was not specified, and the -url is not the default one")
}
sha, date, err := gitCommit()
if err != nil {
return err
}
*version = fmt.Sprintf("publicsuffix.org's public_suffix_list.dat, git revision %s (%s)", sha, date)
}
var r io.Reader = os.Stdin
if *url != "" {
res, err := http.Get(*url)
if err != nil {
return err
}
if res.StatusCode != http.StatusOK {
return fmt.Errorf("bad GET status for %s: %d", *url, res.Status)
}
r = res.Body
defer res.Body.Close()
}
var root node
icann := false
br := bufio.NewReader(r)
for {
s, err := br.ReadString('\n')
if err != nil {
if err == io.EOF {
break
}
return err
}
s = strings.TrimSpace(s)
if strings.Contains(s, "BEGIN ICANN DOMAINS") {
if len(rules) != 0 {
return fmt.Errorf(`expected no rules before "BEGIN ICANN DOMAINS"`)
}
icann = true
continue
}
if strings.Contains(s, "END ICANN DOMAINS") {
icann, numICANNRules = false, len(rules)
continue
}
if s == "" || strings.HasPrefix(s, "//") {
continue
}
s, err = idna.ToASCII(s)
if err != nil {
return err
}
if !validSuffixRE.MatchString(s) {
return fmt.Errorf("bad publicsuffix.org list data: %q", s)
}
if *subset {
switch {
case s == "ac.jp" || strings.HasSuffix(s, ".ac.jp"):
case s == "ak.us" || strings.HasSuffix(s, ".ak.us"):
case s == "ao" || strings.HasSuffix(s, ".ao"):
case s == "ar" || strings.HasSuffix(s, ".ar"):
case s == "arpa" || strings.HasSuffix(s, ".arpa"):
case s == "cy" || strings.HasSuffix(s, ".cy"):
case s == "dyndns.org" || strings.HasSuffix(s, ".dyndns.org"):
case s == "jp":
case s == "kobe.jp" || strings.HasSuffix(s, ".kobe.jp"):
case s == "kyoto.jp" || strings.HasSuffix(s, ".kyoto.jp"):
case s == "om" || strings.HasSuffix(s, ".om"):
case s == "uk" || strings.HasSuffix(s, ".uk"):
case s == "uk.com" || strings.HasSuffix(s, ".uk.com"):
case s == "tw" || strings.HasSuffix(s, ".tw"):
case s == "zw" || strings.HasSuffix(s, ".zw"):
case s == "xn--p1ai" || strings.HasSuffix(s, ".xn--p1ai"):
// xn--p1ai is Russian-Cyrillic "рф".
default:
continue
}
}
rules = append(rules, s)
nt, wildcard := nodeTypeNormal, false
switch {
case strings.HasPrefix(s, "*."):
s, nt = s[2:], nodeTypeParentOnly
wildcard = true
case strings.HasPrefix(s, "!"):
s, nt = s[1:], nodeTypeException
}
labels := strings.Split(s, ".")
for n, i := &root, len(labels)-1; i >= 0; i-- {
label := labels[i]
n = n.child(label)
if i == 0 {
if nt != nodeTypeParentOnly && n.nodeType == nodeTypeParentOnly {
n.nodeType = nt
}
n.icann = n.icann && icann
n.wildcard = n.wildcard || wildcard
}
labelsMap[label] = true
}
}
labelsList = make([]string, 0, len(labelsMap))
for label := range labelsMap {
labelsList = append(labelsList, label)
}
sort.Strings(labelsList)
if err := generate(printReal, &root, "table.go"); err != nil {
return err
}
if err := generate(printTest, &root, "table_test.go"); err != nil {
return err
}
return nil
}
func generate(p func(io.Writer, *node) error, root *node, filename string) error {
buf := new(bytes.Buffer)
if err := p(buf, root); err != nil {
return err
}
b, err := format.Source(buf.Bytes())
if err != nil {
return err
}
return ioutil.WriteFile(filename, b, 0644)
}
func gitCommit() (sha, date string, retErr error) {
res, err := http.Get(gitCommitURL)
if err != nil {
return "", "", err
}
if res.StatusCode != http.StatusOK {
return "", "", fmt.Errorf("bad GET status for %s: %d", gitCommitURL, res.Status)
}
defer res.Body.Close()
b, err := ioutil.ReadAll(res.Body)
if err != nil {
return "", "", err
}
if m := shaRE.FindSubmatch(b); m != nil {
sha = string(m[1])
}
if m := dateRE.FindSubmatch(b); m != nil {
date = string(m[1])
}
if sha == "" || date == "" {
retErr = fmt.Errorf("could not find commit SHA and date in %s", gitCommitURL)
}
return sha, date, retErr
}
func printTest(w io.Writer, n *node) error {
fmt.Fprintf(w, "// generated by go run gen.go; DO NOT EDIT\n\n")
fmt.Fprintf(w, "package publicsuffix\n\nconst numICANNRules = %d\n\nvar rules = [...]string{\n", numICANNRules)
for _, rule := range rules {
fmt.Fprintf(w, "%q,\n", rule)
}
fmt.Fprintf(w, "}\n\nvar nodeLabels = [...]string{\n")
if err := n.walk(w, printNodeLabel); err != nil {
return err
}
fmt.Fprintf(w, "}\n")
return nil
}
func printReal(w io.Writer, n *node) error {
const header = `// generated by go run gen.go; DO NOT EDIT
package publicsuffix
const version = %q
const (
nodesBitsChildren = %d
nodesBitsICANN = %d
nodesBitsTextOffset = %d
nodesBitsTextLength = %d
childrenBitsWildcard = %d
childrenBitsNodeType = %d
childrenBitsHi = %d
childrenBitsLo = %d
)
const (
nodeTypeNormal = %d
nodeTypeException = %d
nodeTypeParentOnly = %d
)
// numTLD is the number of top level domains.
const numTLD = %d
`
fmt.Fprintf(w, header, *version,
nodesBitsChildren, nodesBitsICANN, nodesBitsTextOffset, nodesBitsTextLength,
childrenBitsWildcard, childrenBitsNodeType, childrenBitsHi, childrenBitsLo,
nodeTypeNormal, nodeTypeException, nodeTypeParentOnly, len(n.children))
text := combineText(labelsList)
if text == "" {
return fmt.Errorf("internal error: makeText returned no text")
}
for _, label := range labelsList {
offset, length := strings.Index(text, label), len(label)
if offset < 0 {
return fmt.Errorf("internal error: could not find %q in text %q", label, text)
}
maxTextOffset, maxTextLength = max(maxTextOffset, offset), max(maxTextLength, length)
if offset >= 1<<nodesBitsTextOffset {
return fmt.Errorf("text offset %d is too large, or nodeBitsTextOffset is too small", offset)
}
if length >= 1<<nodesBitsTextLength {
return fmt.Errorf("text length %d is too large, or nodeBitsTextLength is too small", length)
}
labelEncoding[label] = uint32(offset)<<nodesBitsTextLength | uint32(length)
}
fmt.Fprintf(w, "// Text is the combined text of all labels.\nconst text = ")
for len(text) > 0 {
n, plus := len(text), ""
if n > 64 {
n, plus = 64, " +"
}
fmt.Fprintf(w, "%q%s\n", text[:n], plus)
text = text[n:]
}
if err := n.walk(w, assignIndexes); err != nil {
return err
}
fmt.Fprintf(w, `
// nodes is the list of nodes. Each node is represented as a uint32, which
// encodes the node's children, wildcard bit and node type (as an index into
// the children array), ICANN bit and text.
//
// If the table was generated with the -comments flag, there is a //-comment
// after each node's data. In it is the nodes-array indexes of the children,
// formatted as (n0x1234-n0x1256), with * denoting the wildcard bit. The
// nodeType is printed as + for normal, ! for exception, and o for parent-only
// nodes that have children but don't match a domain label in their own right.
// An I denotes an ICANN domain.
//
// The layout within the uint32, from MSB to LSB, is:
// [%2d bits] unused
// [%2d bits] children index
// [%2d bits] ICANN bit
// [%2d bits] text index
// [%2d bits] text length
var nodes = [...]uint32{
`,
32-nodesBitsChildren-nodesBitsICANN-nodesBitsTextOffset-nodesBitsTextLength,
nodesBitsChildren, nodesBitsICANN, nodesBitsTextOffset, nodesBitsTextLength)
if err := n.walk(w, printNode); err != nil {
return err
}
fmt.Fprintf(w, `}
// children is the list of nodes' children, the parent's wildcard bit and the
// parent's node type. If a node has no children then their children index
// will be in the range [0, 6), depending on the wildcard bit and node type.
//
// The layout within the uint32, from MSB to LSB, is:
// [%2d bits] unused
// [%2d bits] wildcard bit
// [%2d bits] node type
// [%2d bits] high nodes index (exclusive) of children
// [%2d bits] low nodes index (inclusive) of children
var children=[...]uint32{
`,
32-childrenBitsWildcard-childrenBitsNodeType-childrenBitsHi-childrenBitsLo,
childrenBitsWildcard, childrenBitsNodeType, childrenBitsHi, childrenBitsLo)
for i, c := range childrenEncoding {
s := "---------------"
lo := c & (1<<childrenBitsLo - 1)
hi := (c >> childrenBitsLo) & (1<<childrenBitsHi - 1)
if lo != hi {
s = fmt.Sprintf("n0x%04x-n0x%04x", lo, hi)
}
nodeType := int(c>>(childrenBitsLo+childrenBitsHi)) & (1<<childrenBitsNodeType - 1)
wildcard := c>>(childrenBitsLo+childrenBitsHi+childrenBitsNodeType) != 0
if *comments {
fmt.Fprintf(w, "0x%08x, // c0x%04x (%s)%s %s\n",
c, i, s, wildcardStr(wildcard), nodeTypeStr(nodeType))
} else {
fmt.Fprintf(w, "0x%x,\n", c)
}
}
fmt.Fprintf(w, "}\n\n")
fmt.Fprintf(w, "// max children %d (capacity %d)\n", maxChildren, 1<<nodesBitsChildren-1)
fmt.Fprintf(w, "// max text offset %d (capacity %d)\n", maxTextOffset, 1<<nodesBitsTextOffset-1)
fmt.Fprintf(w, "// max text length %d (capacity %d)\n", maxTextLength, 1<<nodesBitsTextLength-1)
fmt.Fprintf(w, "// max hi %d (capacity %d)\n", maxHi, 1<<childrenBitsHi-1)
fmt.Fprintf(w, "// max lo %d (capacity %d)\n", maxLo, 1<<childrenBitsLo-1)
return nil
}
type node struct {
label string
nodeType int
icann bool
wildcard bool
// nodesIndex and childrenIndex are the index of this node in the nodes
// and the index of its children offset/length in the children arrays.
nodesIndex, childrenIndex int
// firstChild is the index of this node's first child, or zero if this
// node has no children.
firstChild int
// children are the node's children, in strictly increasing node label order.
children []*node
}
func (n *node) walk(w io.Writer, f func(w1 io.Writer, n1 *node) error) error {
if err := f(w, n); err != nil {
return err
}
for _, c := range n.children {
if err := c.walk(w, f); err != nil {
return err
}
}
return nil
}
// child returns the child of n with the given label. The child is created if
// it did not exist beforehand.
func (n *node) child(label string) *node {
for _, c := range n.children {
if c.label == label {
return c
}
}
c := &node{
label: label,
nodeType: nodeTypeParentOnly,
icann: true,
}
n.children = append(n.children, c)
sort.Sort(byLabel(n.children))
return c
}
type byLabel []*node
func (b byLabel) Len() int { return len(b) }
func (b byLabel) Swap(i, j int) { b[i], b[j] = b[j], b[i] }
func (b byLabel) Less(i, j int) bool { return b[i].label < b[j].label }
var nextNodesIndex int
// childrenEncoding are the encoded entries in the generated children array.
// All these pre-defined entries have no children.
var childrenEncoding = []uint32{
0 << (childrenBitsLo + childrenBitsHi), // Without wildcard bit, nodeTypeNormal.
1 << (childrenBitsLo + childrenBitsHi), // Without wildcard bit, nodeTypeException.
2 << (childrenBitsLo + childrenBitsHi), // Without wildcard bit, nodeTypeParentOnly.
4 << (childrenBitsLo + childrenBitsHi), // With wildcard bit, nodeTypeNormal.
5 << (childrenBitsLo + childrenBitsHi), // With wildcard bit, nodeTypeException.
6 << (childrenBitsLo + childrenBitsHi), // With wildcard bit, nodeTypeParentOnly.
}
var firstCallToAssignIndexes = true
func assignIndexes(w io.Writer, n *node) error {
if len(n.children) != 0 {
// Assign nodesIndex.
n.firstChild = nextNodesIndex
for _, c := range n.children {
c.nodesIndex = nextNodesIndex
nextNodesIndex++
}
// The root node's children is implicit.
if firstCallToAssignIndexes {
firstCallToAssignIndexes = false
return nil
}
// Assign childrenIndex.
maxChildren = max(maxChildren, len(childrenEncoding))
if len(childrenEncoding) >= 1<<nodesBitsChildren {
return fmt.Errorf("children table size %d is too large, or nodeBitsChildren is too small", len(childrenEncoding))
}
n.childrenIndex = len(childrenEncoding)
lo := uint32(n.firstChild)
hi := lo + uint32(len(n.children))
maxLo, maxHi = u32max(maxLo, lo), u32max(maxHi, hi)
if lo >= 1<<childrenBitsLo {
return fmt.Errorf("children lo %d is too large, or childrenBitsLo is too small", lo)
}
if hi >= 1<<childrenBitsHi {
return fmt.Errorf("children hi %d is too large, or childrenBitsHi is too small", hi)
}
enc := hi<<childrenBitsLo | lo
enc |= uint32(n.nodeType) << (childrenBitsLo + childrenBitsHi)
if n.wildcard {
enc |= 1 << (childrenBitsLo + childrenBitsHi + childrenBitsNodeType)
}
childrenEncoding = append(childrenEncoding, enc)
} else {
n.childrenIndex = n.nodeType
if n.wildcard {
n.childrenIndex += numNodeType
}
}
return nil
}
func printNode(w io.Writer, n *node) error {
for _, c := range n.children {
s := "---------------"
if len(c.children) != 0 {
s = fmt.Sprintf("n0x%04x-n0x%04x", c.firstChild, c.firstChild+len(c.children))
}
encoding := labelEncoding[c.label]
if c.icann {
encoding |= 1 << (nodesBitsTextLength + nodesBitsTextOffset)
}
encoding |= uint32(c.childrenIndex) << (nodesBitsTextLength + nodesBitsTextOffset + nodesBitsICANN)
if *comments {
fmt.Fprintf(w, "0x%08x, // n0x%04x c0x%04x (%s)%s %s %s %s\n",
encoding, c.nodesIndex, c.childrenIndex, s, wildcardStr(c.wildcard),
nodeTypeStr(c.nodeType), icannStr(c.icann), c.label,
)
} else {
fmt.Fprintf(w, "0x%x,\n", encoding)
}
}
return nil
}
func printNodeLabel(w io.Writer, n *node) error {
for _, c := range n.children {
fmt.Fprintf(w, "%q,\n", c.label)
}
return nil
}
func icannStr(icann bool) string {
if icann {
return "I"
}
return " "
}
func wildcardStr(wildcard bool) string {
if wildcard {
return "*"
}
return " "
}
// combineText combines all the strings in labelsList to form one giant string.
// Overlapping strings will be merged: "arpa" and "parliament" could yield
// "arparliament".
func combineText(labelsList []string) string {
beforeLength := 0
for _, s := range labelsList {
beforeLength += len(s)
}
text := crush(removeSubstrings(labelsList))
if *v {
fmt.Fprintf(os.Stderr, "crushed %d bytes to become %d bytes\n", beforeLength, len(text))
}
return text
}
type byLength []string
func (s byLength) Len() int { return len(s) }
func (s byLength) Swap(i, j int) { s[i], s[j] = s[j], s[i] }
func (s byLength) Less(i, j int) bool { return len(s[i]) < len(s[j]) }
// removeSubstrings returns a copy of its input with any strings removed
// that are substrings of other provided strings.
func removeSubstrings(input []string) []string {
// Make a copy of input.
ss := append(make([]string, 0, len(input)), input...)
sort.Sort(byLength(ss))
for i, shortString := range ss {
// For each string, only consider strings higher than it in sort order, i.e.
// of equal length or greater.
for _, longString := range ss[i+1:] {
if strings.Contains(longString, shortString) {
ss[i] = ""
break
}
}
}
// Remove the empty strings.
sort.Strings(ss)
for len(ss) > 0 && ss[0] == "" {
ss = ss[1:]
}
return ss
}
// crush combines a list of strings, taking advantage of overlaps. It returns a
// single string that contains each input string as a substring.
func crush(ss []string) string {
maxLabelLen := 0
for _, s := range ss {
if maxLabelLen < len(s) {
maxLabelLen = len(s)
}
}
for prefixLen := maxLabelLen; prefixLen > 0; prefixLen-- {
prefixes := makePrefixMap(ss, prefixLen)
for i, s := range ss {
if len(s) <= prefixLen {
continue
}
mergeLabel(ss, i, prefixLen, prefixes)
}
}
return strings.Join(ss, "")
}
// mergeLabel merges the label at ss[i] with the first available matching label
// in prefixMap, where the last "prefixLen" characters in ss[i] match the first
// "prefixLen" characters in the matching label.
// It will merge ss[i] repeatedly until no more matches are available.
// All matching labels merged into ss[i] are replaced by "".
func mergeLabel(ss []string, i, prefixLen int, prefixes prefixMap) {
s := ss[i]
suffix := s[len(s)-prefixLen:]
for _, j := range prefixes[suffix] {
// Empty strings mean "already used." Also avoid merging with self.
if ss[j] == "" || i == j {
continue
}
if *v {
fmt.Fprintf(os.Stderr, "%d-length overlap at (%4d,%4d): %q and %q share %q\n",
prefixLen, i, j, ss[i], ss[j], suffix)
}
ss[i] += ss[j][prefixLen:]
ss[j] = ""
// ss[i] has a new suffix, so merge again if possible.
// Note: we only have to merge again at the same prefix length. Shorter
// prefix lengths will be handled in the next iteration of crush's for loop.
// Can there be matches for longer prefix lengths, introduced by the merge?
// I believe that any such matches would by necessity have been eliminated
// during substring removal or merged at a higher prefix length. For
// instance, in crush("abc", "cde", "bcdef"), combining "abc" and "cde"
// would yield "abcde", which could be merged with "bcdef." However, in
// practice "cde" would already have been elimintated by removeSubstrings.
mergeLabel(ss, i, prefixLen, prefixes)
return
}
}
// prefixMap maps from a prefix to a list of strings containing that prefix. The
// list of strings is represented as indexes into a slice of strings stored
// elsewhere.
type prefixMap map[string][]int
// makePrefixMap constructs a prefixMap from a slice of strings.
func makePrefixMap(ss []string, prefixLen int) prefixMap {
prefixes := make(prefixMap)
for i, s := range ss {
// We use < rather than <= because if a label matches on a prefix equal to
// its full length, that's actually a substring match handled by
// removeSubstrings.
if prefixLen < len(s) {
prefix := s[:prefixLen]
prefixes[prefix] = append(prefixes[prefix], i)
}
}
return prefixes
}

View file

@ -165,10 +165,6 @@ func nodeLabel(i uint32) string {
// EffectiveTLDPlusOne returns the effective top level domain plus one more
// label. For example, the eTLD+1 for "foo.bar.golang.org" is "golang.org".
func EffectiveTLDPlusOne(domain string) (string, error) {
if strings.HasPrefix(domain, ".") || strings.HasSuffix(domain, ".") || strings.Contains(domain, "..") {
return "", fmt.Errorf("publicsuffix: empty label in domain %q", domain)
}
suffix, _ := PublicSuffix(domain)
if len(domain) <= len(suffix) {
return "", fmt.Errorf("publicsuffix: cannot derive eTLD+1 for domain %q", domain)

File diff suppressed because it is too large Load diff