Bump to v1.0.1

.gitignore

@@ -1,3 +1,3 @@
 /.idea/
 .DS_Store
-/oddb-go
+/od-database-crawler

README.md

@@ -1,2 +1,7 @@
-# oddb Go crawler
+# od-database Go crawler 🚀
 > by terorie 2018 :P
+
+ * Crawler for [__OD-Database__](https://github.com/simon987/od-database)
+ * Crawls HTTP open directories (standard Web Server Listings)
+ * Gets name, path, size and modification time of all files
+ * Lightweight and fast: __over 9000 requests per second__ on a standard laptop

config.go

@@ -1,42 +1,64 @@
 package main
 
 import (
+	"bufio"
 	"fmt"
 	"github.com/sirupsen/logrus"
 	"github.com/spf13/viper"
+	"io"
 	"os"
+	"strings"
 	"time"
 )
 
 var config struct {
 	ServerUrl  string
 	Token      string
+	ServerTimeout time.Duration
+	Recheck    time.Duration
+	ChunkSize  int64
 	Retries    int
 	Workers    int
+	UserAgent  string
+	Timeout    time.Duration
 	Tasks      int32
 	CrawlStats time.Duration
 	AllocStats time.Duration
 	Verbose    bool
+	PrintHTTP  bool
 }
 
 const (
 	ConfServerUrl  = "server.url"
 	ConfToken      = "server.token"
+	ConfServerTimeout = "server.timeout"
+	ConfRecheck    = "server.recheck"
+	ConfChunkSize  = "server.upload_chunk"
 	ConfTasks      = "crawl.tasks"
 	ConfRetries    = "crawl.retries"
 	ConfWorkers    = "crawl.connections"
+	ConfUserAgent  = "crawl.user-agent"
+	ConfTimeout    = "crawl.timeout"
 	ConfCrawlStats = "output.crawl_stats"
 	ConfAllocStats = "output.resource_stats"
 	ConfVerbose    = "output.verbose"
+	ConfPrintHTTP  = "output.http"
+	ConfLogFile    = "output.log"
 )
 
 func prepareConfig() {
 	viper.SetDefault(ConfRetries, 5)
 	viper.SetDefault(ConfWorkers, 2)
 	viper.SetDefault(ConfTasks, 3)
+	viper.SetDefault(ConfUserAgent, "")
+	viper.SetDefault(ConfTimeout, 10 * time.Second)
 	viper.SetDefault(ConfCrawlStats, 3 * time.Second)
 	viper.SetDefault(ConfAllocStats, 0)
 	viper.SetDefault(ConfVerbose, false)
+	viper.SetDefault(ConfPrintHTTP, false)
+	viper.SetDefault(ConfLogFile, "")
+	viper.SetDefault(ConfRecheck, 3 * time.Second)
+	viper.SetDefault(ConfChunkSize, "1 MB")
 }
 
 func readConfig() {
@@ -49,14 +71,24 @@ func readConfig() {
 	}
 
 	config.ServerUrl = viper.GetString(ConfServerUrl)
-	//if config.ServerUrl == "" {
+	if config.ServerUrl == "" {
-	//	configMissing(ConfServerUrl)
+		configMissing(ConfServerUrl)
-	//}
+	}
+	config.ServerUrl = strings.TrimRight(config.ServerUrl, "/")
 
 	config.Token = viper.GetString(ConfToken)
-	//if config.Token == "" {
+	if config.Token == "" {
-	//	configMissing(ConfToken)
+		configMissing(ConfToken)
-	//}
+	}
+
+	config.ServerTimeout = viper.GetDuration(ConfServerTimeout)
+
+	config.Recheck = viper.GetDuration(ConfRecheck)
+
+	config.ChunkSize = int64(viper.GetSizeInBytes(ConfChunkSize))
+	if config.ChunkSize < 100 {
+		configOOB(ConfChunkSize, config.ChunkSize)
+	}
 
 	config.Retries = viper.GetInt(ConfRetries)
 	if config.Retries < 0 {
@@ -73,6 +105,10 @@ func readConfig() {
 		configOOB(ConfTasks, int(config.Tasks))
 	}
 
+	config.UserAgent = viper.GetString(ConfUserAgent)
+
+	config.Timeout = viper.GetDuration(ConfTimeout)
+
 	config.CrawlStats = viper.GetDuration(ConfCrawlStats)
 
 	config.AllocStats = viper.GetDuration(ConfAllocStats)
@@ -81,6 +117,19 @@ func readConfig() {
 	if config.Verbose {
 		logrus.SetLevel(logrus.DebugLevel)
 	}
+
+	if filePath := viper.GetString(ConfLogFile); filePath != "" {
+		f, err := os.OpenFile(filePath, os.O_CREATE | os.O_WRONLY | os.O_APPEND, 0644)
+		bufWriter := bufio.NewWriter(f)
+		if err != nil { panic(err) }
+		exitHooks.Add(func() {
+			bufWriter.Flush()
+			f.Close()
+		})
+		logrus.SetOutput(io.MultiWriter(os.Stdout, bufWriter))
+	}
+
+	config.PrintHTTP = viper.GetBool(ConfPrintHTTP)
 }
 
 func configMissing(key string) {
@@ -88,7 +137,7 @@ func configMissing(key string) {
 	os.Exit(1)
 }
 
-func configOOB(key string, v int) {
+func configOOB(key string, v interface{}) {
-	fmt.Fprintf(os.Stderr, "config: illegal value %d for %key!\n", v, key)
+	fmt.Fprintf(os.Stderr, "config: illegal value %v for key %s!\n", v, key)
 	os.Exit(1)
 }

config.yml

@@ -1,26 +1,60 @@
 # OD-Database server settings
 server:
   # Connection URL
-  url: localhost:6969
+  url: http://od-db.mine.terorie.com/api
+
   # Server auth token
   token:
 
+  # Request timeout
+  timeout: 60s
+
+  # Recheck interval
+  # The crawler periodically asks the server
+  # for new jobs. Sets the minimum wait time
+  # between /task/get requests to the server.
+  recheck: 1s
+
+  # Upload chunk size
+  # If the value is too high, the upload fails.
+  upload_chunk: 1 MB
+
 # Log output settings
 output:
   # Crawl statistics
   crawl_stats: 1s
+
   # CPU/RAM/Job queue stats
-  resource_stats: 1s
+  resource_stats: 10s
+
   # More output? (Every listed dir)
   verbose: false
 
+  # Print HTTP errors (Super spammy)
+  http: false
+
+  # Log file
+  # If empty, no log file is created.
+  log: crawler.log
+
 # Crawler settings
 crawl:
-  # Number of sites that can be
+  # Number of sites that can be processed at once
-  # processed at once
+  tasks: 100
-  tasks: 3
+
   # Number of connections per site
-  connections: 2
+  # Please be careful with this setting!
+  # The crawler fires fast and more than
+  # ten connections can overwhelm a server.
+  connections: 10
+
   # How often to retry getting data
   # from the site before giving up
   retries: 5
+
+  # Time before discarding a network request
+  timeout: 10s
+
+  # Crawler User-Agent
+  # If empty, no User-Agent header is sent.
+  user-agent: "Mozilla/5.0 (X11; od-database-crawler) Gecko/20100101 Firefox/52.0"

crawl.go

@@ -2,103 +2,102 @@ package main
 
 import (
 	"bytes"
-	"encoding/base64"
+	"crypto/tls"
-	"fmt"
+	"github.com/terorie/od-database-crawler/ds/redblackhash"
-	"github.com/sirupsen/logrus"
+	"github.com/terorie/od-database-crawler/fasturl"
 	"github.com/valyala/fasthttp"
 	"golang.org/x/crypto/blake2b"
 	"golang.org/x/net/html"
-	"golang.org/x/net/html/atom"
-	"net/url"
 	"path"
 	"strconv"
 	"strings"
 	"time"
 )
 
-var client fasthttp.Client
+var client = fasthttp.Client {
+	TLSConfig: &tls.Config{
+		InsecureSkipVerify: true,
+	},
+}
 
-func GetDir(j *Job, f *File) (links []url.URL, err error) {
+func GetDir(j *Job, f *File) (links []fasturl.URL, err error) {
 	f.IsDir = true
 	f.Name = path.Base(j.Uri.Path)
 
 	req := fasthttp.AcquireRequest()
-	req.SetRequestURI(j.Uri.String())
+	if config.UserAgent != "" {
+		req.Header.SetUserAgent(config.UserAgent)
+	}
+	req.SetRequestURI(j.UriStr)
 
 	res := fasthttp.AcquireResponse()
 	defer fasthttp.ReleaseResponse(res)
 
-	err = client.Do(req, res)
+	err = client.DoTimeout(req, res, config.Timeout)
 	fasthttp.ReleaseRequest(req)
 
 	if err != nil {
-		logrus.Error(err)
 		return
 	}
 
 	err = checkStatusCode(res.StatusCode())
-	if err != nil { return }
+	if err != nil {
+		return
+	}
 
 	body := res.Body()
 	doc := html.NewTokenizer(bytes.NewReader(body))
 
 	var linkHref string
-	var linkTexts []string
 	for {
 		tokenType := doc.Next()
-		token := doc.Token()
 		if tokenType == html.ErrorToken {
 			break
 		}
 
 		switch tokenType {
 		case html.StartTagToken:
-			if token.DataAtom == atom.A {
+			name, hasAttr := doc.TagName()
-				for _, attr := range token.Attr {
+			if len(name) == 1 && name[0] == 'a' {
-					if attr.Key == "href" {
+				for hasAttr {
-						linkHref = attr.Val
+					var ks, vs []byte
+					ks, vs, hasAttr = doc.TagAttr()
+					if bytes.Equal(ks, []byte("href")) {
+						// TODO Check escape
+						linkHref = string(vs)
 						break
 					}
 				}
 			}
 
-		case html.TextToken:
-			if linkHref != "" {
-				linkTexts = append(linkTexts, token.Data)
-			}
-
 		case html.EndTagToken:
-			if linkHref != "" && token.DataAtom == atom.A {
+			name, _ := doc.TagName()
+			if len(name) == 1 && name[0] == 'a' {
 				// Copy params
 				href := linkHref
-				linkText := strings.Join(linkTexts, " ")
 
 				// Reset params
 				linkHref = ""
-				linkTexts = nil
 
-				// TODO Optimized decision tree
+				if strings.LastIndexByte(href, '?') != -1 {
-				for _, entry := range urlBlackList {
-					if href == entry {
 					goto nextToken
 				}
-				}
+
-				for _, entry := range urlPartBlackList {
+				switch href {
-					if strings.Contains(href, entry) {
+				case "", " ", ".", "..", "/":
 					goto nextToken
 				}
-				}
+
-				for _, entry := range fileNameBlackList {
+				if strings.Contains(href, "../") {
-					if strings.Contains(linkText, entry) {
 					goto nextToken
 				}
+
+				var link fasturl.URL
+				err = j.Uri.ParseRel(&link, href)
+				if err != nil {
+					continue
 				}
 
-				subref, err := url.Parse(href)
-				if err != nil { continue }
-
-				link := *j.Uri.ResolveReference(subref)
-
 				if link.Scheme != j.Uri.Scheme ||
 					link.Host != j.Uri.Host ||
 					link.Path == j.Uri.Path ||
@@ -116,36 +115,42 @@ func GetDir(j *Job, f *File) (links []url.URL, err error) {
 	return
 }
 
-func GetFile(u url.URL, f *File) (err error) {
+func GetFile(u fasturl.URL, f *File) (err error) {
 	f.IsDir = false
 	u.Path = path.Clean(u.Path)
 	f.Name = path.Base(u.Path)
-	f.Path = strings.Trim(u.Path, "/")
+	f.Path = strings.Trim(path.Dir(u.Path), "/")
 
 	req := fasthttp.AcquireRequest()
 	req.Header.SetMethod("HEAD")
+	if config.UserAgent != "" {
+		req.Header.SetUserAgent(config.UserAgent)
+	}
 	req.SetRequestURI(u.String())
 
 	res := fasthttp.AcquireResponse()
 	res.SkipBody = true
 	defer fasthttp.ReleaseResponse(res)
 
-	err = client.Do(req, res)
+	err = client.DoTimeout(req, res, config.Timeout)
 	fasthttp.ReleaseRequest(req)
 
-	if err != nil { return }
+	if err != nil {
+		return
+	}
 
 	err = checkStatusCode(res.StatusCode())
-	if err != nil { return }
+	if err != nil {
+		return
+	}
 
-	// TODO Inefficient af
+	f.applyContentLength(string(res.Header.Peek("content-length")))
-	header := res.Header.Header()
+	f.applyLastModified(string(res.Header.Peek("last-modified")))
-	f.ParseHeader(header)
 
 	return nil
 }
 
-func (f *File) HashDir(links []url.URL) string {
+func (f *File) HashDir(links []fasturl.URL) (o redblackhash.Key) {
 	h, _ := blake2b.New256(nil)
 	h.Write([]byte(f.Name))
 	for _, link := range links {
@@ -153,67 +158,46 @@ func (f *File) HashDir(links []url.URL) string {
 		h.Write([]byte(fileName))
 	}
 	sum := h.Sum(nil)
-	b64sum := base64.StdEncoding.EncodeToString(sum)
+	copy(o[:redblackhash.KeySize], sum)
-	return b64sum
+	return
 }
 
-func (f *File) ParseHeader(h []byte) {
+func (f *File) applyContentLength(v string) {
-	var k1, k2 int
+	if v == "" {
-	var v1, v2 int
+		return
-
-	// Simple finite state machine
-	state := 0
-	for i, b := range h {
-		switch state {
-		case 0:
-			if b == byte(':') {
-				state = 1
-				k2 = i
 	}
-
-		case 1:
-			state = 2
-
-		case 2:
-			state = 3
-			v1 = i
-
-		case 3:
-			if b == byte('\r') {
-				state = 4
-			}
-
-		case 4:
-			state = 0
-			v2 = i - 1
-
-			key := string(h[k1:k2])
-			val := string(h[v1:v2])
-			k1 = i
-
-			f.applyHeader(key, val)
-		}
-	}
-
-}
-
-func (f *File) applyHeader(k, v string) {
-	switch k {
-	case "content-length":
 	size, err := strconv.ParseInt(v, 10, 64)
-		if err != nil { break }
+	if err != nil {
-		if size < 0 { break }
+		return
+	}
+	if size < 0 {
+		return
+	}
 	f.Size = size
+}
 
-	case "last-modified":
+// TODO Cleanup
+func (f *File) applyLastModified(v string) {
+	if v == "" {
+		return
+	}
+	var t time.Time
 	var err error
-		f.MTime, err = time.Parse(time.RFC1123, v)
+	t, err = time.Parse(time.RFC1123, v)
-		if err == nil { break }
+	if err == nil {
-		f.MTime, err = time.Parse(time.RFC850, v)
+		f.MTime = t.Unix()
-		if err == nil { break }
+		return
+	}
+	t, err = time.Parse(time.RFC850, v)
+	if err == nil {
+		f.MTime = t.Unix()
+		return
+	}
 	// TODO Parse asctime
-		f.MTime, err = time.Parse("2006-01-02", v[:10])
+	t, err = time.Parse("2006-01-02", v[:10])
-		if err == nil { break }
+	if err == nil {
+		f.MTime = t.Unix()
+		return
 	}
 }
 
@@ -221,53 +205,7 @@ func checkStatusCode(status int) error {
 	switch status {
 	case fasthttp.StatusOK:
 		return nil
-
-	case fasthttp.StatusTooManyRequests:
-		return ErrRateLimit
-
-	case fasthttp.StatusForbidden,
-		fasthttp.StatusUnauthorized:
-		return ErrForbidden
-
 	default:
-		return fmt.Errorf("got HTTP status %d", status)
+		return &HttpError{status}
 	}
 }
-
-var urlBlackList = [...]string {
-	"",
-	" ",
-	".",
-	"..",
-	"/",
-}
-
-var urlPartBlackList = [...]string {
-	"?C=N&O=D",
-	"?C=M&O=A",
-	"?C=S&O=A",
-	"?C=D&O=A",
-	"?C=N;O=D",
-	"?C=M;O=A",
-	"?C=M&O=D",
-	"?C=S;O=A",
-	"?C=S&O=D",
-	"?C=D;O=A",
-	"?MA",
-	"?SA",
-	"?DA",
-	"?ND",
-	"?C=N&O=A",
-	"?C=N&O=A",
-	"?M=A",
-	"?N=D",
-	"?S=A",
-	"?D=A",
-}
-
-var fileNameBlackList = [...]string {
-	"Parent Directory",
-	" Parent Directory",
-	"../",
-}
-

ds/redblackhash/redblack.go

@@ -0,0 +1,489 @@
+// Copyright (c) 2015, Emir Pasic. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Modifications by terorie
+
+// Package redblacktree implements a red-black tree.
+//
+// Used by TreeSet and TreeMap.
+//
+// Structure is not thread safe.
+//
+// References: http://en.wikipedia.org/wiki/Red%E2%80%93black_tree
+package redblackhash
+
+import (
+	"bytes"
+	"fmt"
+	"sync"
+)
+
+const (
+	black, red color = true, false
+	KeySize = 64
+)
+
+type color bool
+type Key [KeySize]byte
+
+// Tree holds elements of the red-black tree
+type Tree struct {
+	sync.Mutex
+	Root       *Node
+	size       int
+}
+
+// Node is a single element within the tree
+type Node struct {
+	Key    Key
+	color  color
+	Left   *Node
+	Right  *Node
+	Parent *Node
+}
+
+func (k *Key) Compare(o *Key) int {
+	return bytes.Compare(k[:], o[:])
+}
+
+// Put inserts node into the tree.
+// Key should adhere to the comparator's type assertion, otherwise method panics.
+func (tree *Tree) Put(key *Key) {
+	var insertedNode *Node
+	if tree.Root == nil {
+		// Assert key is of comparator's type for initial tree
+		tree.Root = &Node{Key: *key, color: red}
+		insertedNode = tree.Root
+	} else {
+		node := tree.Root
+		loop := true
+		for loop {
+			compare := key.Compare(&node.Key)
+			switch {
+			case compare == 0:
+				node.Key = *key
+				return
+			case compare < 0:
+				if node.Left == nil {
+					node.Left = &Node{Key: *key, color: red}
+					insertedNode = node.Left
+					loop = false
+				} else {
+					node = node.Left
+				}
+			case compare > 0:
+				if node.Right == nil {
+					node.Right = &Node{Key: *key, color: red}
+					insertedNode = node.Right
+					loop = false
+				} else {
+					node = node.Right
+				}
+			}
+		}
+		insertedNode.Parent = node
+	}
+	tree.insertCase1(insertedNode)
+	tree.size++
+}
+
+// Get searches the node in the tree by key and returns its value or nil if key is not found in tree.
+// Second return parameter is true if key was found, otherwise false.
+// Key should adhere to the comparator's type assertion, otherwise method panics.
+func (tree *Tree) Get(key *Key) (found bool) {
+	node := tree.lookup(key)
+	return node != nil
+}
+
+// Remove remove the node from the tree by key.
+// Key should adhere to the comparator's type assertion, otherwise method panics.
+func (tree *Tree) Remove(key *Key) {
+	var child *Node
+	node := tree.lookup(key)
+	if node == nil {
+		return
+	}
+	if node.Left != nil && node.Right != nil {
+		pred := node.Left.maximumNode()
+		node.Key = pred.Key
+		node = pred
+	}
+	if node.Left == nil || node.Right == nil {
+		if node.Right == nil {
+			child = node.Left
+		} else {
+			child = node.Right
+		}
+		if node.color == black {
+			node.color = nodeColor(child)
+			tree.deleteCase1(node)
+		}
+		tree.replaceNode(node, child)
+		if node.Parent == nil && child != nil {
+			child.color = black
+		}
+	}
+	tree.size--
+}
+
+// Empty returns true if tree does not contain any nodes
+func (tree *Tree) Empty() bool {
+	return tree.size == 0
+}
+
+// Size returns number of nodes in the tree.
+func (tree *Tree) Size() int {
+	return tree.size
+}
+
+// Left returns the left-most (min) node or nil if tree is empty.
+func (tree *Tree) Left() *Node {
+	var parent *Node
+	current := tree.Root
+	for current != nil {
+		parent = current
+		current = current.Left
+	}
+	return parent
+}
+
+// Right returns the right-most (max) node or nil if tree is empty.
+func (tree *Tree) Right() *Node {
+	var parent *Node
+	current := tree.Root
+	for current != nil {
+		parent = current
+		current = current.Right
+	}
+	return parent
+}
+
+// Floor Finds floor node of the input key, return the floor node or nil if no floor is found.
+// Second return parameter is true if floor was found, otherwise false.
+//
+// Floor node is defined as the largest node that is smaller than or equal to the given node.
+// A floor node may not be found, either because the tree is empty, or because
+// all nodes in the tree are larger than the given node.
+//
+// Key should adhere to the comparator's type assertion, otherwise method panics.
+func (tree *Tree) Floor(key *Key) (floor *Node, found bool) {
+	found = false
+	node := tree.Root
+	for node != nil {
+		compare := key.Compare(&node.Key)
+		switch {
+		case compare == 0:
+			return node, true
+		case compare < 0:
+			node = node.Left
+		case compare > 0:
+			floor, found = node, true
+			node = node.Right
+		}
+	}
+	if found {
+		return floor, true
+	}
+	return nil, false
+}
+
+// Ceiling finds ceiling node of the input key, return the ceiling node or nil if no ceiling is found.
+// Second return parameter is true if ceiling was found, otherwise false.
+//
+// Ceiling node is defined as the smallest node that is larger than or equal to the given node.
+// A ceiling node may not be found, either because the tree is empty, or because
+// all nodes in the tree are smaller than the given node.
+//
+// Key should adhere to the comparator's type assertion, otherwise method panics.
+func (tree *Tree) Ceiling(key *Key) (ceiling *Node, found bool) {
+	found = false
+	node := tree.Root
+	for node != nil {
+		compare := key.Compare(&node.Key)
+		switch {
+		case compare == 0:
+			return node, true
+		case compare < 0:
+			ceiling, found = node, true
+			node = node.Left
+		case compare > 0:
+			node = node.Right
+		}
+	}
+	if found {
+		return ceiling, true
+	}
+	return nil, false
+}
+
+// Clear removes all nodes from the tree.
+func (tree *Tree) Clear() {
+	tree.Root = nil
+	tree.size = 0
+}
+
+// String returns a string representation of container
+func (tree *Tree) String() string {
+	str := "RedBlackTree\n"
+	if !tree.Empty() {
+		output(tree.Root, "", true, &str)
+	}
+	return str
+}
+
+func (node *Node) String() string {
+	return fmt.Sprintf("%v", node.Key)
+}
+
+func output(node *Node, prefix string, isTail bool, str *string) {
+	if node.Right != nil {
+		newPrefix := prefix
+		if isTail {
+			newPrefix += "│   "
+		} else {
+			newPrefix += "    "
+		}
+		output(node.Right, newPrefix, false, str)
+	}
+	*str += prefix
+	if isTail {
+		*str += "└── "
+	} else {
+		*str += "┌── "
+	}
+	*str += node.String() + "\n"
+	if node.Left != nil {
+		newPrefix := prefix
+		if isTail {
+			newPrefix += "    "
+		} else {
+			newPrefix += "│   "
+		}
+		output(node.Left, newPrefix, true, str)
+	}
+}
+
+func (tree *Tree) lookup(key *Key) *Node {
+	node := tree.Root
+	for node != nil {
+		compare := key.Compare(&node.Key)
+		switch {
+		case compare == 0:
+			return node
+		case compare < 0:
+			node = node.Left
+		case compare > 0:
+			node = node.Right
+		}
+	}
+	return nil
+}
+
+func (node *Node) grandparent() *Node {
+	if node != nil && node.Parent != nil {
+		return node.Parent.Parent
+	}
+	return nil
+}
+
+func (node *Node) uncle() *Node {
+	if node == nil || node.Parent == nil || node.Parent.Parent == nil {
+		return nil
+	}
+	return node.Parent.sibling()
+}
+
+func (node *Node) sibling() *Node {
+	if node == nil || node.Parent == nil {
+		return nil
+	}
+	if node == node.Parent.Left {
+		return node.Parent.Right
+	}
+	return node.Parent.Left
+}
+
+func (tree *Tree) rotateLeft(node *Node) {
+	right := node.Right
+	tree.replaceNode(node, right)
+	node.Right = right.Left
+	if right.Left != nil {
+		right.Left.Parent = node
+	}
+	right.Left = node
+	node.Parent = right
+}
+
+func (tree *Tree) rotateRight(node *Node) {
+	left := node.Left
+	tree.replaceNode(node, left)
+	node.Left = left.Right
+	if left.Right != nil {
+		left.Right.Parent = node
+	}
+	left.Right = node
+	node.Parent = left
+}
+
+func (tree *Tree) replaceNode(old *Node, new *Node) {
+	if old.Parent == nil {
+		tree.Root = new
+	} else {
+		if old == old.Parent.Left {
+			old.Parent.Left = new
+		} else {
+			old.Parent.Right = new
+		}
+	}
+	if new != nil {
+		new.Parent = old.Parent
+	}
+}
+
+func (tree *Tree) insertCase1(node *Node) {
+	if node.Parent == nil {
+		node.color = black
+	} else {
+		tree.insertCase2(node)
+	}
+}
+
+func (tree *Tree) insertCase2(node *Node) {
+	if nodeColor(node.Parent) == black {
+		return
+	}
+	tree.insertCase3(node)
+}
+
+func (tree *Tree) insertCase3(node *Node) {
+	uncle := node.uncle()
+	if nodeColor(uncle) == red {
+		node.Parent.color = black
+		uncle.color = black
+		node.grandparent().color = red
+		tree.insertCase1(node.grandparent())
+	} else {
+		tree.insertCase4(node)
+	}
+}
+
+func (tree *Tree) insertCase4(node *Node) {
+	grandparent := node.grandparent()
+	if node == node.Parent.Right && node.Parent == grandparent.Left {
+		tree.rotateLeft(node.Parent)
+		node = node.Left
+	} else if node == node.Parent.Left && node.Parent == grandparent.Right {
+		tree.rotateRight(node.Parent)
+		node = node.Right
+	}
+	tree.insertCase5(node)
+}
+
+func (tree *Tree) insertCase5(node *Node) {
+	node.Parent.color = black
+	grandparent := node.grandparent()
+	grandparent.color = red
+	if node == node.Parent.Left && node.Parent == grandparent.Left {
+		tree.rotateRight(grandparent)
+	} else if node == node.Parent.Right && node.Parent == grandparent.Right {
+		tree.rotateLeft(grandparent)
+	}
+}
+
+func (node *Node) maximumNode() *Node {
+	if node == nil {
+		return nil
+	}
+	for node.Right != nil {
+		node = node.Right
+	}
+	return node
+}
+
+func (tree *Tree) deleteCase1(node *Node) {
+	if node.Parent == nil {
+		return
+	}
+	tree.deleteCase2(node)
+}
+
+func (tree *Tree) deleteCase2(node *Node) {
+	sibling := node.sibling()
+	if nodeColor(sibling) == red {
+		node.Parent.color = red
+		sibling.color = black
+		if node == node.Parent.Left {
+			tree.rotateLeft(node.Parent)
+		} else {
+			tree.rotateRight(node.Parent)
+		}
+	}
+	tree.deleteCase3(node)
+}
+
+func (tree *Tree) deleteCase3(node *Node) {
+	sibling := node.sibling()
+	if nodeColor(node.Parent) == black &&
+		nodeColor(sibling) == black &&
+		nodeColor(sibling.Left) == black &&
+		nodeColor(sibling.Right) == black {
+		sibling.color = red
+		tree.deleteCase1(node.Parent)
+	} else {
+		tree.deleteCase4(node)
+	}
+}
+
+func (tree *Tree) deleteCase4(node *Node) {
+	sibling := node.sibling()
+	if nodeColor(node.Parent) == red &&
+		nodeColor(sibling) == black &&
+		nodeColor(sibling.Left) == black &&
+		nodeColor(sibling.Right) == black {
+		sibling.color = red
+		node.Parent.color = black
+	} else {
+		tree.deleteCase5(node)
+	}
+}
+
+func (tree *Tree) deleteCase5(node *Node) {
+	sibling := node.sibling()
+	if node == node.Parent.Left &&
+		nodeColor(sibling) == black &&
+		nodeColor(sibling.Left) == red &&
+		nodeColor(sibling.Right) == black {
+		sibling.color = red
+		sibling.Left.color = black
+		tree.rotateRight(sibling)
+	} else if node == node.Parent.Right &&
+		nodeColor(sibling) == black &&
+		nodeColor(sibling.Right) == red &&
+		nodeColor(sibling.Left) == black {
+		sibling.color = red
+		sibling.Right.color = black
+		tree.rotateLeft(sibling)
+	}
+	tree.deleteCase6(node)
+}
+
+func (tree *Tree) deleteCase6(node *Node) {
+	sibling := node.sibling()
+	sibling.color = nodeColor(node.Parent)
+	node.Parent.color = black
+	if node == node.Parent.Left && nodeColor(sibling.Right) == red {
+		sibling.Right.color = black
+		tree.rotateLeft(node.Parent)
+	} else if nodeColor(sibling.Left) == red {
+		sibling.Left.color = black
+		tree.rotateRight(node.Parent)
+	}
+}
+
+func nodeColor(node *Node) color {
+	if node == nil {
+		return black
+	}
+	return node.color
+}

errors.go

@@ -1,8 +1,17 @@
 package main
 
-import "errors"
+import (
+	"errors"
+	"fmt"
+)
 
 var ErrRateLimit = errors.New("too many requests")
-var ErrForbidden = errors.New("access denied")
 var ErrKnown     = errors.New("already crawled")
 
+type HttpError struct {
+	code int
+}
+
+func (e HttpError) Error() string {
+	return fmt.Sprintf("http status %d", e.code)
+}

fasturl/url.go

@@ -0,0 +1,869 @@
+// 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.
+
+// Package fasturl parses URLs and implements query escaping.
+package fasturl
+
+// Modifications by terorie
+
+// See RFC 3986. This package generally follows RFC 3986, except where
+// it deviates for compatibility reasons. When sending changes, first
+// search old issues for history on decisions. Unit tests should also
+// contain references to issue numbers with details.
+
+import (
+	"errors"
+	"fmt"
+	"strconv"
+	"strings"
+)
+
+type Scheme int
+const (
+	SchemeInvalid = iota
+	SchemeHTTP
+	SchemeHTTPS
+	SchemeCount
+)
+
+var Schemes = [SchemeCount]string {
+	"",
+	"http",
+	"https",
+}
+
+var ErrUnknownScheme = errors.New("unknown protocol scheme")
+
+// Error reports an error and the operation and URL that caused it.
+type Error struct {
+	Op  string
+	URL string
+	Err error
+}
+
+func (e *Error) Error() string { return e.Op + " " + e.URL + ": " + e.Err.Error() }
+
+type timeout interface {
+	Timeout() bool
+}
+
+func (e *Error) Timeout() bool {
+	t, ok := e.Err.(timeout)
+	return ok && t.Timeout()
+}
+
+type temporary interface {
+	Temporary() bool
+}
+
+func (e *Error) Temporary() bool {
+	t, ok := e.Err.(temporary)
+	return ok && t.Temporary()
+}
+
+func ishex(c byte) bool {
+	switch {
+	case '0' <= c && c <= '9':
+		return true
+	case 'a' <= c && c <= 'f':
+		return true
+	case 'A' <= c && c <= 'F':
+		return true
+	}
+	return false
+}
+
+func unhex(c byte) byte {
+	switch {
+	case '0' <= c && c <= '9':
+		return c - '0'
+	case 'a' <= c && c <= 'f':
+		return c - 'a' + 10
+	case 'A' <= c && c <= 'F':
+		return c - 'A' + 10
+	}
+	return 0
+}
+
+type encoding int
+
+const (
+	encodePath encoding = 1 + iota
+	encodePathSegment
+	encodeHost
+	encodeZone
+	encodeUserPassword
+	encodeQueryComponent
+	encodeFragment
+)
+
+type EscapeError string
+
+func (e EscapeError) Error() string {
+	return "invalid URL escape " + strconv.Quote(string(e))
+}
+
+type InvalidHostError string
+
+func (e InvalidHostError) Error() string {
+	return "invalid character " + strconv.Quote(string(e)) + " in host name"
+}
+
+// Return true if the specified character should be escaped when
+// appearing in a URL string, according to RFC 3986.
+//
+// Please be informed that for now shouldEscape does not check all
+// reserved characters correctly. See golang.org/issue/5684.
+func shouldEscape(c byte, mode encoding) bool {
+	// §2.3 Unreserved characters (alphanum)
+	if 'A' <= c && c <= 'Z' || 'a' <= c && c <= 'z' || '0' <= c && c <= '9' {
+		return false
+	}
+
+	if mode == encodeHost || mode == encodeZone {
+		// §3.2.2 Host allows
+		//	sub-delims = "!" / "$" / "&" / "'" / "(" / ")" / "*" / "+" / "," / ";" / "="
+		// as part of reg-name.
+		// We add : because we include :port as part of host.
+		// We add [ ] because we include [ipv6]:port as part of host.
+		// We add < > because they're the only characters left that
+		// we could possibly allow, and Parse will reject them if we
+		// escape them (because hosts can't use %-encoding for
+		// ASCII bytes).
+		switch c {
+		case '!', '$', '&', '\'', '(', ')', '*', '+', ',', ';', '=', ':', '[', ']', '<', '>', '"':
+			return false
+		}
+	}
+
+	switch c {
+	case '-', '_', '.', '~': // §2.3 Unreserved characters (mark)
+		return false
+
+	case '$', '&', '+', ',', '/', ':', ';', '=', '?', '@': // §2.2 Reserved characters (reserved)
+		// Different sections of the URL allow a few of
+		// the reserved characters to appear unescaped.
+		switch mode {
+		case encodePath: // §3.3
+			// The RFC allows : @ & = + $ but saves / ; , for assigning
+			// meaning to individual path segments. This package
+			// only manipulates the path as a whole, so we allow those
+			// last three as well. That leaves only ? to escape.
+			return c == '?'
+
+		case encodePathSegment: // §3.3
+			// The RFC allows : @ & = + $ but saves / ; , for assigning
+			// meaning to individual path segments.
+			return c == '/' || c == ';' || c == ',' || c == '?'
+
+		case encodeUserPassword: // §3.2.1
+			// The RFC allows ';', ':', '&', '=', '+', '$', and ',' in
+			// userinfo, so we must escape only '@', '/', and '?'.
+			// The parsing of userinfo treats ':' as special so we must escape
+			// that too.
+			return c == '@' || c == '/' || c == '?' || c == ':'
+
+		case encodeQueryComponent: // §3.4
+			// The RFC reserves (so we must escape) everything.
+			return true
+
+		case encodeFragment: // §4.1
+			// The RFC text is silent but the grammar allows
+			// everything, so escape nothing.
+			return false
+		}
+	}
+
+	if mode == encodeFragment {
+		// RFC 3986 §2.2 allows not escaping sub-delims. A subset of sub-delims are
+		// included in reserved from RFC 2396 §2.2. The remaining sub-delims do not
+		// need to be escaped. To minimize potential breakage, we apply two restrictions:
+		// (1) we always escape sub-delims outside of the fragment, and (2) we always
+		// escape single quote to avoid breaking callers that had previously assumed that
+		// single quotes would be escaped. See issue #19917.
+		switch c {
+		case '!', '(', ')', '*':
+			return false
+		}
+	}
+
+	// Everything else must be escaped.
+	return true
+}
+
+// unescape unescapes a string; the mode specifies
+// which section of the URL string is being unescaped.
+func unescape(s string, mode encoding) (string, error) {
+	// Count %, check that they're well-formed.
+	n := 0
+	hasPlus := false
+	for i := 0; i < len(s); {
+		switch s[i] {
+		case '%':
+			n++
+			if i+2 >= len(s) || !ishex(s[i+1]) || !ishex(s[i+2]) {
+				s = s[i:]
+				if len(s) > 3 {
+					s = s[:3]
+				}
+				return "", EscapeError(s)
+			}
+			// Per https://tools.ietf.org/html/rfc3986#page-21
+			// in the host component %-encoding can only be used
+			// for non-ASCII bytes.
+			// But https://tools.ietf.org/html/rfc6874#section-2
+			// introduces %25 being allowed to escape a percent sign
+			// in IPv6 scoped-address literals. Yay.
+			if mode == encodeHost && unhex(s[i+1]) < 8 && s[i:i+3] != "%25" {
+				return "", EscapeError(s[i : i+3])
+			}
+			if mode == encodeZone {
+				// RFC 6874 says basically "anything goes" for zone identifiers
+				// and that even non-ASCII can be redundantly escaped,
+				// but it seems prudent to restrict %-escaped bytes here to those
+				// that are valid host name bytes in their unescaped form.
+				// That is, you can use escaping in the zone identifier but not
+				// to introduce bytes you couldn't just write directly.
+				// But Windows puts spaces here! Yay.
+				v := unhex(s[i+1])<<4 | unhex(s[i+2])
+				if s[i:i+3] != "%25" && v != ' ' && shouldEscape(v, encodeHost) {
+					return "", EscapeError(s[i : i+3])
+				}
+			}
+			i += 3
+		case '+':
+			hasPlus = mode == encodeQueryComponent
+			i++
+		default:
+			if (mode == encodeHost || mode == encodeZone) && s[i] < 0x80 && shouldEscape(s[i], mode) {
+				return "", InvalidHostError(s[i : i+1])
+			}
+			i++
+		}
+	}
+
+	if n == 0 && !hasPlus {
+		return s, nil
+	}
+
+	t := make([]byte, len(s)-2*n)
+	j := 0
+	for i := 0; i < len(s); {
+		switch s[i] {
+		case '%':
+			t[j] = unhex(s[i+1])<<4 | unhex(s[i+2])
+			j++
+			i += 3
+		case '+':
+			if mode == encodeQueryComponent {
+				t[j] = ' '
+			} else {
+				t[j] = '+'
+			}
+			j++
+			i++
+		default:
+			t[j] = s[i]
+			j++
+			i++
+		}
+	}
+	return string(t), nil
+}
+
+func escape(s string, mode encoding) string {
+	spaceCount, hexCount := 0, 0
+	for i := 0; i < len(s); i++ {
+		c := s[i]
+		if shouldEscape(c, mode) {
+			if c == ' ' && mode == encodeQueryComponent {
+				spaceCount++
+			} else {
+				hexCount++
+			}
+		}
+	}
+
+	if spaceCount == 0 && hexCount == 0 {
+		return s
+	}
+
+	t := make([]byte, len(s)+2*hexCount)
+	j := 0
+	for i := 0; i < len(s); i++ {
+		switch c := s[i]; {
+		case c == ' ' && mode == encodeQueryComponent:
+			t[j] = '+'
+			j++
+		case shouldEscape(c, mode):
+			t[j] = '%'
+			t[j+1] = "0123456789ABCDEF"[c>>4]
+			t[j+2] = "0123456789ABCDEF"[c&15]
+			j += 3
+		default:
+			t[j] = s[i]
+			j++
+		}
+	}
+	return string(t)
+}
+
+// A URL represents a parsed URL (technically, a URI reference).
+//
+// The general form represented is:
+//
+//	[scheme:][//[userinfo@]host][/]path[?query][#fragment]
+//
+// URLs that do not start with a slash after the scheme are interpreted as:
+//
+//	scheme:opaque[?query][#fragment]
+//
+// Note that the Path field is stored in decoded form: /%47%6f%2f becomes /Go/.
+// A consequence is that it is impossible to tell which slashes in the Path were
+// slashes in the raw URL and which were %2f. This distinction is rarely important,
+// but when it is, code must not use Path directly.
+// The Parse function sets both Path and RawPath in the URL it returns,
+// and URL's String method uses RawPath if it is a valid encoding of Path,
+// by calling the EscapedPath method.
+type URL struct {
+	Scheme     Scheme
+	Host       string    // host or host:port
+	Path       string    // path (relative paths may omit leading slash)
+}
+
+// Maybe rawurl is of the form scheme:path.
+// (Scheme must be [a-zA-Z][a-zA-Z0-9+-.]*)
+// If so, return scheme, path; else return "", rawurl.
+func getscheme(rawurl string) (scheme Scheme, path string, err error) {
+	for i := 0; i < len(rawurl); i++ {
+		c := rawurl[i]
+		switch {
+		case 'a' <= c && c <= 'z' || 'A' <= c && c <= 'Z':
+			// do nothing
+		case '0' <= c && c <= '9' || c == '+' || c == '-' || c == '.':
+			if i == 0 {
+				return SchemeInvalid, rawurl, nil
+			}
+		case c == ':':
+			if i == 0 {
+				return SchemeInvalid, "", errors.New("missing protocol scheme")
+			}
+			switch rawurl[:i] {
+			case "http":
+				scheme = SchemeHTTP
+			case "https":
+				scheme = SchemeHTTPS
+			default:
+				return SchemeInvalid, "", ErrUnknownScheme
+			}
+
+			path = rawurl[i+1:]
+			return
+		default:
+			// we have encountered an invalid character,
+			// so there is no valid scheme
+			return SchemeInvalid, rawurl, nil
+		}
+	}
+	return SchemeInvalid, rawurl, nil
+}
+
+// Maybe s is of the form t c u.
+// If so, return t, c u (or t, u if cutc == true).
+// If not, return s, "".
+func split(s string, c string, cutc bool) (string, string) {
+	i := strings.Index(s, c)
+	if i < 0 {
+		return s, ""
+	}
+	if cutc {
+		return s[:i], s[i+len(c):]
+	}
+	return s[:i], s[i:]
+}
+
+// Parse parses rawurl into a URL structure.
+//
+// The rawurl may be relative (a path, without a host) or absolute
+// (starting with a scheme). Trying to parse a hostname and path
+// without a scheme is invalid but may not necessarily return an
+// error, due to parsing ambiguities.
+func (u *URL) Parse(rawurl string) error {
+	// Cut off #frag
+	s, frag := split(rawurl, "#", true)
+	err := u.parse(s, false)
+	if err != nil {
+		return &Error{"parse", s, err}
+	}
+	if frag == "" {
+		return nil
+	}
+	return nil
+}
+
+// ParseRequestURI parses rawurl into a URL structure. It assumes that
+// rawurl was received in an HTTP request, so the rawurl is interpreted
+// only as an absolute URI or an absolute path.
+// The string rawurl is assumed not to have a #fragment suffix.
+// (Web browsers strip #fragment before sending the URL to a web server.)
+func (u *URL) ParseRequestURI(rawurl string) error {
+	err := u.parse(rawurl, true)
+	if err != nil {
+		return &Error{"parse", rawurl, err}
+	}
+	return nil
+}
+
+// parse parses a URL from a string in one of two contexts. If
+// viaRequest is true, the URL is assumed to have arrived via an HTTP request,
+// in which case only absolute URLs or path-absolute relative URLs are allowed.
+// If viaRequest is false, all forms of relative URLs are allowed.
+func (u *URL) parse(rawurl string, viaRequest bool) error {
+	var rest string
+	var err error
+
+	if rawurl == "" && viaRequest {
+		return errors.New("empty url")
+	}
+
+	if rawurl == "*" {
+		u.Path = "*"
+		return nil
+	}
+
+	// Split off possible leading "http:", "mailto:", etc.
+	// Cannot contain escaped characters.
+	if u.Scheme, rest, err = getscheme(rawurl); err != nil {
+		return err
+	}
+
+	if strings.HasSuffix(rest, "?") && strings.Count(rest, "?") == 1 {
+		rest = rest[:len(rest)-1]
+	} else {
+		rest, _ = split(rest, "?", true)
+	}
+
+	if !strings.HasPrefix(rest, "/") {
+		if u.Scheme != SchemeInvalid {
+			// We consider rootless paths per RFC 3986 as opaque.
+			return nil
+		}
+		if viaRequest {
+			return errors.New("invalid URI for request")
+		}
+
+		// Avoid confusion with malformed schemes, like cache_object:foo/bar.
+		// See golang.org/issue/16822.
+		//
+		// RFC 3986, §3.3:
+		// In addition, a URI reference (Section 4.1) may be a relative-path reference,
+		// in which case the first path segment cannot contain a colon (":") character.
+		colon := strings.Index(rest, ":")
+		slash := strings.Index(rest, "/")
+		if colon >= 0 && (slash < 0 || colon < slash) {
+			// First path segment has colon. Not allowed in relative URL.
+			return errors.New("first path segment in URL cannot contain colon")
+		}
+	}
+
+	if (u.Scheme != SchemeInvalid || !viaRequest && !strings.HasPrefix(rest, "///")) && strings.HasPrefix(rest, "//") {
+		var authority string
+		authority, rest = split(rest[2:], "/", false)
+		u.Host, err = parseAuthority(authority)
+		if err != nil {
+			return err
+		}
+	}
+	u.Path = rest
+	return nil
+}
+
+func parseAuthority(authority string) (host string, err error) {
+	i := strings.LastIndex(authority, "@")
+	if i < 0 {
+		host, err = parseHost(authority)
+	} else {
+		host, err = parseHost(authority[i+1:])
+	}
+	if err != nil {
+		return "", err
+	}
+	if i < 0 {
+		return host, nil
+	}
+	userinfo := authority[:i]
+	if !validUserinfo(userinfo) {
+		return "", errors.New("fasturl: invalid userinfo")
+	}
+	return host, nil
+}
+
+// parseHost parses host as an authority without user
+// information. That is, as host[:port].
+func parseHost(host string) (string, error) {
+	if strings.HasPrefix(host, "[") {
+		// Parse an IP-Literal in RFC 3986 and RFC 6874.
+		// E.g., "[fe80::1]", "[fe80::1%25en0]", "[fe80::1]:80".
+		i := strings.LastIndex(host, "]")
+		if i < 0 {
+			return "", errors.New("missing ']' in host")
+		}
+		colonPort := host[i+1:]
+		if !validOptionalPort(colonPort) {
+			return "", fmt.Errorf("invalid port %q after host", colonPort)
+		}
+
+		// RFC 6874 defines that %25 (%-encoded percent) introduces
+		// the zone identifier, and the zone identifier can use basically
+		// any %-encoding it likes. That's different from the host, which
+		// can only %-encode non-ASCII bytes.
+		// We do impose some restrictions on the zone, to avoid stupidity
+		// like newlines.
+		zone := strings.Index(host[:i], "%25")
+		if zone >= 0 {
+			host1, err := unescape(host[:zone], encodeHost)
+			if err != nil {
+				return "", err
+			}
+			host2, err := unescape(host[zone:i], encodeZone)
+			if err != nil {
+				return "", err
+			}
+			host3, err := unescape(host[i:], encodeHost)
+			if err != nil {
+				return "", err
+			}
+			return host1 + host2 + host3, nil
+		}
+	}
+
+	var err error
+	if host, err = unescape(host, encodeHost); err != nil {
+		return "", err
+	}
+	return host, nil
+}
+
+// validOptionalPort reports whether port is either an empty string
+// or matches /^:\d*$/
+func validOptionalPort(port string) bool {
+	if port == "" {
+		return true
+	}
+	if port[0] != ':' {
+		return false
+	}
+	for _, b := range port[1:] {
+		if b < '0' || b > '9' {
+			return false
+		}
+	}
+	return true
+}
+
+// String reassembles the URL into a valid URL string.
+// The general form of the result is one of:
+//
+//	scheme:opaque?query#fragment
+//	scheme://userinfo@host/path?query#fragment
+//
+// If u.Opaque is non-empty, String uses the first form;
+// otherwise it uses the second form.
+// To obtain the path, String uses u.EscapedPath().
+//
+// In the second form, the following rules apply:
+//	- if u.Scheme is empty, scheme: is omitted.
+//	- if u.User is nil, userinfo@ is omitted.
+//	- if u.Host is empty, host/ is omitted.
+//	- if u.Scheme and u.Host are empty and u.User is nil,
+//	   the entire scheme://userinfo@host/ is omitted.
+//	- if u.Host is non-empty and u.Path begins with a /,
+//	   the form host/path does not add its own /.
+//	- if u.RawQuery is empty, ?query is omitted.
+//	- if u.Fragment is empty, #fragment is omitted.
+func (u *URL) String() string {
+	var buf strings.Builder
+	if u.Scheme != SchemeInvalid {
+		buf.WriteString(Schemes[u.Scheme])
+		buf.WriteByte(':')
+	}
+	if u.Scheme != SchemeInvalid || u.Host != "" {
+		if u.Host != "" || u.Path != "" {
+			buf.WriteString("//")
+		}
+		if h := u.Host; h != "" {
+			buf.WriteString(escape(h, encodeHost))
+		}
+	}
+	path := u.Path
+	if path != "" && path[0] != '/' && u.Host != "" {
+		buf.WriteByte('/')
+	}
+	if buf.Len() == 0 {
+		// RFC 3986 §4.2
+		// A path segment that contains a colon character (e.g., "this:that")
+		// cannot be used as the first segment of a relative-path reference, as
+		// it would be mistaken for a scheme name. Such a segment must be
+		// preceded by a dot-segment (e.g., "./this:that") to make a relative-
+		// path reference.
+		if i := strings.IndexByte(path, ':'); i > -1 && strings.IndexByte(path[:i], '/') == -1 {
+			buf.WriteString("./")
+		}
+	}
+	buf.WriteString(path)
+	return buf.String()
+}
+
+func isRunesDot(r []rune) bool {
+	return len(r) == 1 && r[0] == '.'
+}
+
+func isRunesDoubleDot(r []rune) bool {
+	return len(r) == 2 && r[0] == '.' && r[1] == '.'
+}
+
+// resolvePath applies special path segments from refs and applies
+// them to base, per RFC 3986.
+func resolvePath(base, ref string) string {
+	var full string
+	if ref == "" {
+		full = base
+	} else if ref[0] != '/' {
+		i := strings.LastIndex(base, "/")
+		full = base[:i+1] + ref
+	} else {
+		full = ref
+	}
+	if full == "" {
+		return ""
+	} else if full == "/" {
+		return "/"
+	}
+
+	dst := make([]rune, len(full))
+	dst = dst[0:0]
+
+	start := 0
+	rs := []rune(full)
+	if len(rs) != 0 && rs[0] == '/' {
+		rs = rs[1:]
+	}
+	var stack []int
+	stack = append(stack, 0)
+	for i, c := range rs {
+		if i == len(rs) - 1 {
+			closingSlash := false
+			part := rs[start:]
+			if len(part) == 0 {
+				dst = append(dst, '/')
+			} else if part[len(part)-1] == '/' {
+				part = part[:len(part)-1]
+				closingSlash = true
+			}
+			switch {
+			case isRunesDot(part):
+				dst = append(dst, '/')
+			case isRunesDoubleDot(part):
+				// Cut to the last slash
+				start = i+1
+				dst = dst[:stack[len(stack)-1]]
+				if len(stack) != 1 {
+					stack = stack[:len(stack)-1]
+				}
+				dst = append(dst, '/')
+			default:
+				dst = append(dst, '/')
+				dst = append(dst, part...)
+			}
+			if closingSlash && len(dst) != 0 && dst[len(dst)-1] != '/' {
+				dst = append(dst, '/')
+			}
+		} else if c == '/' {
+			part := rs[start:i]
+			switch {
+			case isRunesDot(part):
+				start = i+1
+			case isRunesDoubleDot(part):
+				// Cut to the last slash
+				start = i+1
+				dst = dst[:stack[len(stack)-1]]
+				if len(stack) != 1 {
+					stack = stack[:len(stack)-1]
+				}
+			default:
+				start = i+1
+				stack = append(stack, len(dst))
+				dst = append(dst, '/')
+				dst = append(dst, part...)
+			}
+		}
+	}
+	return string(dst)
+
+	/*var dst []string
+	src := strings.Split(full, "/")
+	for _, elem := range src {
+		switch elem {
+		case ".":
+			// drop
+		case "..":
+			if len(dst) > 0 {
+				dst = dst[:len(dst)-1]
+			}
+		default:
+			dst = append(dst, elem)
+		}
+	}
+	if last := src[len(src)-1]; last == "." || last == ".." {
+		// Add final slash to the joined path.
+		dst = append(dst, "")
+	}
+	return "/" + strings.TrimPrefix(strings.Join(dst, "/"), "/")*/
+}
+
+// IsAbs reports whether the URL is absolute.
+// Absolute means that it has a non-empty scheme.
+func (u *URL) IsAbs() bool {
+	return u.Scheme != SchemeInvalid
+}
+
+// ParseRel parses a URL in the context of the receiver. The provided URL
+// may be relative or absolute. Parse returns nil, err on parse
+// failure, otherwise its return value is the same as ResolveReference.
+func (u *URL) ParseRel(out *URL, ref string) error {
+	var refurl URL
+
+	err := refurl.Parse(ref)
+	if err != nil {
+		return err
+	}
+
+	u.ResolveReference(out, &refurl)
+	return nil
+}
+
+// ResolveReference resolves a URI reference to an absolute URI from
+// an absolute base URI u, per RFC 3986 Section 5.2. The URI reference
+// may be relative or absolute. ResolveReference always returns a new
+// URL instance, even if the returned URL is identical to either the
+// base or reference. If ref is an absolute URL, then ResolveReference
+// ignores base and returns a copy of ref.
+func (u *URL) ResolveReference(url *URL, ref *URL) {
+	*url = *ref
+	if ref.Scheme == SchemeInvalid {
+		url.Scheme = u.Scheme
+	}
+	if ref.Scheme != SchemeInvalid || ref.Host != "" {
+		// The "absoluteURI" or "net_path" cases.
+		// We can ignore the error from setPath since we know we provided a
+		// validly-escaped path.
+		url.Path = resolvePath(ref.Path, "")
+		return
+	}
+	// The "abs_path" or "rel_path" cases.
+	url.Host = u.Host
+	url.Path = resolvePath(u.Path, ref.Path)
+	return
+}
+
+// Marshaling interface implementations.
+// Would like to implement MarshalText/UnmarshalText but that will change the JSON representation of URLs.
+
+func (u *URL) MarshalBinary() (text []byte, err error) {
+	return []byte(u.String()), nil
+}
+
+func (u *URL) UnmarshalBinary(text []byte) error {
+	var u1 URL
+	err := u1.Parse(string(text))
+	if err != nil {
+		return err
+	}
+	*u = u1
+	return nil
+}
+
+// validUserinfo reports whether s is a valid userinfo string per RFC 3986
+// Section 3.2.1:
+//     userinfo    = *( unreserved / pct-encoded / sub-delims / ":" )
+//     unreserved  = ALPHA / DIGIT / "-" / "." / "_" / "~"
+//     sub-delims  = "!" / "$" / "&" / "'" / "(" / ")"
+//                   / "*" / "+" / "," / ";" / "="
+//
+// It doesn't validate pct-encoded. The caller does that via func unescape.
+func validUserinfo(s string) bool {
+	for _, r := range s {
+		if 'A' <= r && r <= 'Z' {
+			continue
+		}
+		if 'a' <= r && r <= 'z' {
+			continue
+		}
+		if '0' <= r && r <= '9' {
+			continue
+		}
+		switch r {
+		case '-', '.', '_', ':', '~', '!', '$', '&', '\'',
+			'(', ')', '*', '+', ',', ';', '=', '%', '@':
+			continue
+		default:
+			return false
+		}
+	}
+	return true
+}
+
+func PathUnescape(s string) string {
+	newStr, err := pathUnescape(s)
+	if err != nil {
+		return s
+	} else {
+		return newStr
+	}
+}
+
+func pathUnescape(s string) (string, error) {
+	// Count %, check that they're well-formed.
+	n := 0
+	for i := 0; i < len(s); {
+		switch s[i] {
+		case '%':
+			n++
+			if i+2 >= len(s) || !ishex(s[i+1]) || !ishex(s[i+2]) {
+				s = s[i:]
+				if len(s) > 3 {
+					s = s[:3]
+				}
+				return "", EscapeError(s)
+			}
+			i += 3
+		default:
+			i++
+		}
+	}
+
+	if n == 0 {
+		return s, nil
+	}
+
+	t := make([]byte, len(s)-2*n)
+	j := 0
+	for i := 0; i < len(s); {
+		switch s[i] {
+		case '%':
+			t[j] = unhex(s[i+1])<<4 | unhex(s[i+2])
+			j++
+			i += 3
+		case '+':
+			t[j] = '+'
+			j++
+			i++
+		default:
+			t[j] = s[i]
+			j++
+			i++
+		}
+	}
+	return string(t), nil
+}

fasturl/url_test.go

@@ -0,0 +1,897 @@
+// 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.
+
+package fasturl
+
+import (
+	"bytes"
+	encodingPkg "encoding"
+	"encoding/gob"
+	"encoding/json"
+	"fmt"
+	"io"
+	"net"
+	"reflect"
+	"testing"
+)
+
+type URLTest struct {
+	in        string
+	out       *URL   // expected parse; RawPath="" means same as Path
+	roundtrip string // expected result of reserializing the URL; empty means same as "in".
+}
+
+var urltests = []URLTest{
+	// no path
+	{
+		"http://www.google.com",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "www.google.com",
+		},
+		"",
+	},
+	// path
+	{
+		"http://www.google.com/",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "www.google.com",
+			Path:   "/",
+		},
+		"",
+	},
+	// %20 outside query
+	{
+		"http://www.google.com/a%20b",
+		&URL{
+			Scheme:   SchemeHTTP,
+			Host:     "www.google.com",
+			Path:     "/a%20b",
+		},
+		"",
+	},
+	// leading // without scheme should create an authority
+	{
+		"//foo",
+		&URL{
+			Host: "foo",
+		},
+		"",
+	},
+	// Three leading slashes isn't an authority, but doesn't return an error.
+	// (We can't return an error, as this code is also used via
+	// ServeHTTP -> ReadRequest -> Parse, which is arguably a
+	// different URL parsing context, but currently shares the
+	// same codepath)
+	{
+		"///threeslashes",
+		&URL{
+			Path: "///threeslashes",
+		},
+		"",
+	},
+	// unescaped @ in username should not confuse host
+	{
+		"http://j@ne:password@google.com",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "google.com",
+		},
+		"http://google.com",
+	},
+	// unescaped @ in password should not confuse host
+	{
+		"http://jane:p@ssword@google.com",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "google.com",
+		},
+		"http://google.com",
+	},
+	// Relative path
+	{
+		"a/b/c",
+		&URL{
+			Path: "a/b/c",
+		},
+		"a/b/c",
+	},
+	// host subcomponent; IPv4 address in RFC 3986
+	{
+		"http://192.168.0.1/",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "192.168.0.1",
+			Path:   "/",
+		},
+		"",
+	},
+	// host and port subcomponents; IPv4 address in RFC 3986
+	{
+		"http://192.168.0.1:8080/",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "192.168.0.1:8080",
+			Path:   "/",
+		},
+		"",
+	},
+	// host subcomponent; IPv6 address in RFC 3986
+	{
+		"http://[fe80::1]/",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "[fe80::1]",
+			Path:   "/",
+		},
+		"",
+	},
+	// host and port subcomponents; IPv6 address in RFC 3986
+	{
+		"http://[fe80::1]:8080/",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "[fe80::1]:8080",
+			Path:   "/",
+		},
+		"",
+	},
+	// host subcomponent; IPv6 address with zone identifier in RFC 6874
+	{
+		"http://[fe80::1%25en0]/", // alphanum zone identifier
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "[fe80::1%en0]",
+			Path:   "/",
+		},
+		"",
+	},
+	// host and port subcomponents; IPv6 address with zone identifier in RFC 6874
+	{
+		"http://[fe80::1%25en0]:8080/", // alphanum zone identifier
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "[fe80::1%en0]:8080",
+			Path:   "/",
+		},
+		"",
+	},
+	// host subcomponent; IPv6 address with zone identifier in RFC 6874
+	{
+		"http://[fe80::1%25%65%6e%301-._~]/", // percent-encoded+unreserved zone identifier
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "[fe80::1%en01-._~]",
+			Path:   "/",
+		},
+		"http://[fe80::1%25en01-._~]/",
+	},
+	// host and port subcomponents; IPv6 address with zone identifier in RFC 6874
+	{
+		"http://[fe80::1%25%65%6e%301-._~]:8080/", // percent-encoded+unreserved zone identifier
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "[fe80::1%en01-._~]:8080",
+			Path:   "/",
+		},
+		"http://[fe80::1%25en01-._~]:8080/",
+	},
+	// golang.org/issue/12200 (colon with empty port)
+	{
+		"http://192.168.0.2:8080/foo",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "192.168.0.2:8080",
+			Path:   "/foo",
+		},
+		"",
+	},
+	{
+		"http://192.168.0.2:/foo",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "192.168.0.2:",
+			Path:   "/foo",
+		},
+		"",
+	},
+	{
+		// Malformed IPv6 but still accepted.
+		"http://2b01:e34:ef40:7730:8e70:5aff:fefe:edac:8080/foo",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "2b01:e34:ef40:7730:8e70:5aff:fefe:edac:8080",
+			Path:   "/foo",
+		},
+		"",
+	},
+	{
+		// Malformed IPv6 but still accepted.
+		"http://2b01:e34:ef40:7730:8e70:5aff:fefe:edac:/foo",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "2b01:e34:ef40:7730:8e70:5aff:fefe:edac:",
+			Path:   "/foo",
+		},
+		"",
+	},
+	{
+		"http://[2b01:e34:ef40:7730:8e70:5aff:fefe:edac]:8080/foo",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "[2b01:e34:ef40:7730:8e70:5aff:fefe:edac]:8080",
+			Path:   "/foo",
+		},
+		"",
+	},
+	{
+		"http://[2b01:e34:ef40:7730:8e70:5aff:fefe:edac]:/foo",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "[2b01:e34:ef40:7730:8e70:5aff:fefe:edac]:",
+			Path:   "/foo",
+		},
+		"",
+	},
+	// golang.org/issue/7991 and golang.org/issue/12719 (non-ascii %-encoded in host)
+	{
+		"http://hello.世界.com/foo",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "hello.世界.com",
+			Path:   "/foo",
+		},
+		"http://hello.%E4%B8%96%E7%95%8C.com/foo",
+	},
+	{
+		"http://hello.%e4%b8%96%e7%95%8c.com/foo",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "hello.世界.com",
+			Path:   "/foo",
+		},
+		"http://hello.%E4%B8%96%E7%95%8C.com/foo",
+	},
+	{
+		"http://hello.%E4%B8%96%E7%95%8C.com/foo",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "hello.世界.com",
+			Path:   "/foo",
+		},
+		"",
+	},
+	// golang.org/issue/10433 (path beginning with //)
+	{
+		"http://example.com//foo",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "example.com",
+			Path:   "//foo",
+		},
+		"",
+	},
+	// test that we can reparse the host names we accept.
+	{
+		"http://authority<\"hi\">/foo",
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "authority<\"hi\">",
+			Path:   "/foo",
+		},
+		"",
+	},
+}
+
+// more useful string for debugging than fmt's struct printer
+func ufmt(u *URL) string {
+	return fmt.Sprintf("scheme=%q, host=%q, path=%q",
+		Schemes[u.Scheme], u.Host, u.Path)
+}
+
+func BenchmarkString(b *testing.B) {
+	b.StopTimer()
+	b.ReportAllocs()
+	for _, tt := range urltests {
+		var u URL
+		err := u.Parse(tt.in)
+		if err != nil {
+			b.Errorf("Parse(%q) returned error %s", tt.in, err)
+			continue
+		}
+		if tt.roundtrip == "" {
+			continue
+		}
+		b.StartTimer()
+		var g string
+		for i := 0; i < b.N; i++ {
+			g = u.String()
+		}
+		b.StopTimer()
+		if w := tt.roundtrip; b.N > 0 && g != w {
+			b.Errorf("Parse(%q).String() == %q, want %q", tt.in, g, w)
+		}
+	}
+}
+
+func TestParse(t *testing.T) {
+	for _, tt := range urltests {
+		var u URL
+		err := u.Parse(tt.in)
+		if err != nil {
+			t.Errorf("Parse(%q) returned error %v", tt.in, err)
+			continue
+		}
+		if !reflect.DeepEqual(&u, tt.out) {
+			t.Errorf("Parse(%q):\n\tgot  %v\n\twant %v\n", tt.in, ufmt(&u), ufmt(tt.out))
+		}
+	}
+}
+
+const pathThatLooksSchemeRelative = "//not.a.user@not.a.host/just/a/path"
+
+var parseRequestURLTests = []struct {
+	url           string
+	expectedValid bool
+}{
+	{"http://foo.com", true},
+	{"http://foo.com/", true},
+	{"http://foo.com/path", true},
+	{"/", true},
+	{pathThatLooksSchemeRelative, true},
+	{"//not.a.user@%66%6f%6f.com/just/a/path/also", true},
+	{"*", true},
+	{"http://192.168.0.1/", true},
+	{"http://192.168.0.1:8080/", true},
+	{"http://[fe80::1]/", true},
+	{"http://[fe80::1]:8080/", true},
+
+	// Tests exercising RFC 6874 compliance:
+	{"http://[fe80::1%25en0]/", true},                 // with alphanum zone identifier
+	{"http://[fe80::1%25en0]:8080/", true},            // with alphanum zone identifier
+	{"http://[fe80::1%25%65%6e%301-._~]/", true},      // with percent-encoded+unreserved zone identifier
+	{"http://[fe80::1%25%65%6e%301-._~]:8080/", true}, // with percent-encoded+unreserved zone identifier
+
+	{"foo.html", false},
+	{"../dir/", false},
+	{"http://192.168.0.%31/", false},
+	{"http://192.168.0.%31:8080/", false},
+	{"http://[fe80::%31]/", false},
+	{"http://[fe80::%31]:8080/", false},
+	{"http://[fe80::%31%25en0]/", false},
+	{"http://[fe80::%31%25en0]:8080/", false},
+
+	// These two cases are valid as textual representations as
+	// described in RFC 4007, but are not valid as address
+	// literals with IPv6 zone identifiers in URIs as described in
+	// RFC 6874.
+	{"http://[fe80::1%en0]/", false},
+	{"http://[fe80::1%en0]:8080/", false},
+}
+
+func TestParseRequestURI(t *testing.T) {
+	for _, test := range parseRequestURLTests {
+		var u URL
+		err := u.ParseRequestURI(test.url)
+		if test.expectedValid && err != nil {
+			t.Errorf("ParseRequestURI(%q) gave err %v; want no error", test.url, err)
+		} else if !test.expectedValid && err == nil {
+			t.Errorf("ParseRequestURI(%q) gave nil error; want some error", test.url)
+		}
+	}
+
+	var url URL
+	err := url.ParseRequestURI(pathThatLooksSchemeRelative)
+	if err != nil {
+		t.Fatalf("Unexpected error %v", err)
+	}
+	if url.Path != pathThatLooksSchemeRelative {
+		t.Errorf("ParseRequestURI path:\ngot  %q\nwant %q", url.Path, pathThatLooksSchemeRelative)
+	}
+}
+
+var stringURLTests = []struct {
+	url  URL
+	want string
+}{
+	// No leading slash on path should prepend slash on String() call
+	{
+		url: URL{
+			Scheme: SchemeHTTP,
+			Host:   "www.google.com",
+			Path:   "search",
+		},
+		want: "http://www.google.com/search",
+	},
+	// Relative path with first element containing ":" should be prepended with "./", golang.org/issue/17184
+	{
+		url: URL{
+			Path: "this:that",
+		},
+		want: "./this:that",
+	},
+	// Relative path with second element containing ":" should not be prepended with "./"
+	{
+		url: URL{
+			Path: "here/this:that",
+		},
+		want: "here/this:that",
+	},
+	// Non-relative path with first element containing ":" should not be prepended with "./"
+	{
+		url: URL{
+			Scheme: SchemeHTTP,
+			Host:   "www.google.com",
+			Path:   "this:that",
+		},
+		want: "http://www.google.com/this:that",
+	},
+}
+
+func TestURLString(t *testing.T) {
+	for _, tt := range urltests {
+		var u URL
+		err := u.Parse(tt.in)
+		if err != nil {
+			t.Errorf("Parse(%q) returned error %s", tt.in, err)
+			continue
+		}
+		expected := tt.in
+		if tt.roundtrip != "" {
+			expected = tt.roundtrip
+		}
+		s := u.String()
+		if s != expected {
+			t.Errorf("Parse(%q).String() == %q (expected %q)", tt.in, s, expected)
+		}
+	}
+
+	for _, tt := range stringURLTests {
+		if got := tt.url.String(); got != tt.want {
+			t.Errorf("%+v.String() = %q; want %q", tt.url, got, tt.want)
+		}
+	}
+}
+
+var resolvePathTests = []struct {
+	base, ref, expected string
+}{
+	{"a/b", ".", "/a/"},
+	{"a/b", "c", "/a/c"},
+	{"a/b", "..", "/"},
+	{"a/", "..", "/"},
+	{"a/", "../..", "/"},
+	{"a/b/c", "..", "/a/"},
+	{"a/b/c", "../d", "/a/d"},
+	{"a/b/c", ".././d", "/a/d"},
+	{"a/b", "./..", "/"},
+	{"a/./b", ".", "/a/"},
+	{"a/../", ".", "/"},
+	{"a/.././b", "c", "/c"},
+}
+
+func TestResolvePath(t *testing.T) {
+	for _, test := range resolvePathTests {
+		got := resolvePath(test.base, test.ref)
+		if got != test.expected {
+			t.Errorf("For %q + %q got %q; expected %q", test.base, test.ref, got, test.expected)
+		}
+	}
+}
+
+var resolveReferenceTests = []struct {
+	base, rel, expected string
+}{
+	// Absolute URL references
+	{"http://foo.com?a=b", "https://bar.com/", "https://bar.com/"},
+	{"http://foo.com/", "https://bar.com/?a=b", "https://bar.com/"},
+	{"http://foo.com/", "https://bar.com/?", "https://bar.com/"},
+
+	// Path-absolute references
+	{"http://foo.com/bar", "/baz", "http://foo.com/baz"},
+	{"http://foo.com/bar?a=b#f", "/baz", "http://foo.com/baz"},
+	{"http://foo.com/bar?a=b", "/baz?", "http://foo.com/baz"},
+	{"http://foo.com/bar?a=b", "/baz?c=d", "http://foo.com/baz"},
+
+	// Multiple slashes
+	{"http://foo.com/bar", "http://foo.com//baz", "http://foo.com//baz"},
+	{"http://foo.com/bar", "http://foo.com///baz/quux", "http://foo.com///baz/quux"},
+
+	// Scheme-relative
+	{"https://foo.com/bar?a=b", "//bar.com/quux", "https://bar.com/quux"},
+
+	// Path-relative references:
+
+	// ... current directory
+	{"http://foo.com", ".", "http://foo.com/"},
+	{"http://foo.com/bar", ".", "http://foo.com/"},
+	{"http://foo.com/bar/", ".", "http://foo.com/bar/"},
+
+	// ... going down
+	{"http://foo.com", "bar", "http://foo.com/bar"},
+	{"http://foo.com/", "bar", "http://foo.com/bar"},
+	{"http://foo.com/bar/baz", "quux", "http://foo.com/bar/quux"},
+
+	// ... going up
+	{"http://foo.com/bar/baz", "../quux", "http://foo.com/quux"},
+	{"http://foo.com/bar/baz", "../../../../../quux", "http://foo.com/quux"},
+	{"http://foo.com/bar", "..", "http://foo.com/"},
+	{"http://foo.com/bar/baz", "./..", "http://foo.com/"},
+	// ".." in the middle (issue 3560)
+	{"http://foo.com/bar/baz", "quux/dotdot/../tail", "http://foo.com/bar/quux/tail"},
+	{"http://foo.com/bar/baz", "quux/./dotdot/../tail", "http://foo.com/bar/quux/tail"},
+	{"http://foo.com/bar/baz", "quux/./dotdot/.././tail", "http://foo.com/bar/quux/tail"},
+	{"http://foo.com/bar/baz", "quux/./dotdot/./../tail", "http://foo.com/bar/quux/tail"},
+	{"http://foo.com/bar/baz", "quux/./dotdot/dotdot/././../../tail", "http://foo.com/bar/quux/tail"},
+	{"http://foo.com/bar/baz", "quux/./dotdot/dotdot/./.././../tail", "http://foo.com/bar/quux/tail"},
+	{"http://foo.com/bar/baz", "quux/./dotdot/dotdot/dotdot/./../../.././././tail", "http://foo.com/bar/quux/tail"},
+	{"http://foo.com/bar/baz", "quux/./dotdot/../dotdot/../dot/./tail/..", "http://foo.com/bar/quux/dot/"},
+
+	// Remove any dot-segments prior to forming the target URI.
+	// http://tools.ietf.org/html/rfc3986#section-5.2.4
+	{"http://foo.com/dot/./dotdot/../foo/bar", "../baz", "http://foo.com/dot/baz"},
+
+	// Triple dot isn't special
+	{"http://foo.com/bar", "...", "http://foo.com/..."},
+
+	// Fragment
+	{"http://foo.com/bar", ".#frag", "http://foo.com/"},
+	{"http://example.org/", "#!$&%27()*+,;=", "http://example.org/"},
+
+	// Paths with escaping (issue 16947).
+	{"http://foo.com/foo%2fbar/", "../baz", "http://foo.com/baz"},
+	{"http://foo.com/1/2%2f/3%2f4/5", "../../a/b/c", "http://foo.com/1/a/b/c"},
+	{"http://foo.com/1/2/3", "./a%2f../../b/..%2fc", "http://foo.com/1/2/b/..%2fc"},
+	{"http://foo.com/1/2%2f/3%2f4/5", "./a%2f../b/../c", "http://foo.com/1/2%2f/3%2f4/a%2f../c"},
+	{"http://foo.com/foo%20bar/", "../baz", "http://foo.com/baz"},
+	{"http://foo.com/foo", "../bar%2fbaz", "http://foo.com/bar%2fbaz"},
+	{"http://foo.com/foo%2dbar/", "./baz-quux", "http://foo.com/foo%2dbar/baz-quux"},
+
+	// RFC 3986: Normal Examples
+	// http://tools.ietf.org/html/rfc3986#section-5.4.1
+	{"http://a/b/c/d;p?q", "g", "http://a/b/c/g"},
+	{"http://a/b/c/d;p?q", "./g", "http://a/b/c/g"},
+	{"http://a/b/c/d;p?q", "g/", "http://a/b/c/g/"},
+	{"http://a/b/c/d;p?q", "/g", "http://a/g"},
+	{"http://a/b/c/d;p?q", "//g", "http://g"},
+	{"http://a/b/c/d;p?q", "?y", "http://a/b/c/d;p"},
+	{"http://a/b/c/d;p?q", "g?y", "http://a/b/c/g"},
+	{"http://a/b/c/d;p?q", "#s", "http://a/b/c/d;p"},
+	{"http://a/b/c/d;p?q", "g#s", "http://a/b/c/g"},
+	{"http://a/b/c/d;p?q", "g?y#s", "http://a/b/c/g"},
+	{"http://a/b/c/d;p?q", ";x", "http://a/b/c/;x"},
+	{"http://a/b/c/d;p?q", "g;x", "http://a/b/c/g;x"},
+	{"http://a/b/c/d;p?q", "g;x?y#s", "http://a/b/c/g;x"},
+	{"http://a/b/c/d;p?q", "", "http://a/b/c/d;p"},
+	{"http://a/b/c/d;p?q", ".", "http://a/b/c/"},
+	{"http://a/b/c/d;p?q", "./", "http://a/b/c/"},
+	{"http://a/b/c/d;p?q", "..", "http://a/b/"},
+	{"http://a/b/c/d;p?q", "../", "http://a/b/"},
+	{"http://a/b/c/d;p?q", "../g", "http://a/b/g"},
+	{"http://a/b/c/d;p?q", "../..", "http://a/"},
+	{"http://a/b/c/d;p?q", "../../", "http://a/"},
+	{"http://a/b/c/d;p?q", "../../g", "http://a/g"},
+
+	// RFC 3986: Abnormal Examples
+	// http://tools.ietf.org/html/rfc3986#section-5.4.2
+	{"http://a/b/c/d;p?q", "../../../g", "http://a/g"},
+	{"http://a/b/c/d;p?q", "../../../../g", "http://a/g"},
+	{"http://a/b/c/d;p?q", "/./g", "http://a/g"},
+	{"http://a/b/c/d;p?q", "/../g", "http://a/g"},
+	{"http://a/b/c/d;p?q", "g.", "http://a/b/c/g."},
+	{"http://a/b/c/d;p?q", ".g", "http://a/b/c/.g"},
+	{"http://a/b/c/d;p?q", "g..", "http://a/b/c/g.."},
+	{"http://a/b/c/d;p?q", "..g", "http://a/b/c/..g"},
+	{"http://a/b/c/d;p?q", "./../g", "http://a/b/g"},
+	{"http://a/b/c/d;p?q", "./g/.", "http://a/b/c/g/"},
+	{"http://a/b/c/d;p?q", "g/./h", "http://a/b/c/g/h"},
+	{"http://a/b/c/d;p?q", "g/../h", "http://a/b/c/h"},
+	{"http://a/b/c/d;p?q", "g;x=1/./y", "http://a/b/c/g;x=1/y"},
+	{"http://a/b/c/d;p?q", "g;x=1/../y", "http://a/b/c/y"},
+	{"http://a/b/c/d;p?q", "g?y/./x", "http://a/b/c/g"},
+	{"http://a/b/c/d;p?q", "g?y/../x", "http://a/b/c/g"},
+	{"http://a/b/c/d;p?q", "g#s/./x", "http://a/b/c/g"},
+	{"http://a/b/c/d;p?q", "g#s/../x", "http://a/b/c/g"},
+
+	// Extras.
+	{"https://a/b/c/d;p?q", "//g?q", "https://g"},
+	{"https://a/b/c/d;p?q", "//g#s", "https://g"},
+	{"https://a/b/c/d;p?q", "//g/d/e/f?y#s", "https://g/d/e/f"},
+	{"https://a/b/c/d;p#s", "?y", "https://a/b/c/d;p"},
+	{"https://a/b/c/d;p?q#s", "?y", "https://a/b/c/d;p"},
+}
+
+func TestResolveReference(t *testing.T) {
+	mustParse := func(url string) *URL {
+		u := new(URL)
+		err := u.Parse(url)
+		if err != nil {
+			t.Fatalf("Parse(%q) got err %v", url, err)
+		}
+		return u
+	}
+	for _, test := range resolveReferenceTests {
+		base := mustParse(test.base)
+		rel := mustParse(test.rel)
+		var url URL
+		base.ResolveReference(&url, rel)
+		if got := url.String(); got != test.expected {
+			t.Errorf("URL(%q).ResolveReference(%q)\ngot  %q\nwant %q", test.base, test.rel, got, test.expected)
+		}
+	}
+}
+
+type RequestURITest struct {
+	url *URL
+	out string
+}
+
+var requritests = []RequestURITest{
+	{
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "example.com",
+			Path:   "",
+		},
+		"/",
+	},
+	{
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "example.com",
+			Path:   "/a b",
+		},
+		"/a%20b",
+	},
+	{
+		&URL{
+			Scheme: SchemeHTTP,
+			Host:   "example.com",
+			Path:   "//foo",
+		},
+		"//foo",
+	},
+}
+
+func TestParseErrors(t *testing.T) {
+	tests := []struct {
+		in      string
+		wantErr bool
+	}{
+		{"http://[::1]", false},
+		{"http://[::1]:80", false},
+		{"http://[::1]:namedport", true}, // rfc3986 3.2.3
+		{"http://[::1]/", false},
+		{"http://[::1]a", true},
+		{"http://[::1]%23", true},
+		{"http://[::1%25en0]", false},     // valid zone id
+		{"http://[::1]:", false},          // colon, but no port OK
+		{"http://[::1]:%38%30", true},     // not allowed: % encoding only for non-ASCII
+		{"http://[::1%25%41]", false},     // RFC 6874 allows over-escaping in zone
+		{"http://[%10::1]", true},         // no %xx escapes in IP address
+		{"http://[::1]/%48", false},       // %xx in path is fine
+		{"http://%41:8080/", true},        // not allowed: % encoding only for non-ASCII
+
+		{"http://[]%20%48%54%54%50%2f%31%2e%31%0a%4d%79%48%65%61%64%65%72%3a%20%31%32%33%0a%0a/", true}, // golang.org/issue/11208
+		{"http://a b.com/", true},    // no space in host name please
+	}
+	for _, tt := range tests {
+		var u URL
+		err := u.Parse(tt.in)
+		if tt.wantErr {
+			if err == nil {
+				t.Errorf("Parse(%q) = %#v; want an error", tt.in, u)
+			}
+			continue
+		}
+		if err != nil {
+			t.Logf("Parse(%q) = %v; want no error", tt.in, err)
+		}
+	}
+}
+
+type shouldEscapeTest struct {
+	in     byte
+	mode   encoding
+	escape bool
+}
+
+var shouldEscapeTests = []shouldEscapeTest{
+	// Unreserved characters (§2.3)
+	{'a', encodePath, false},
+	{'a', encodeUserPassword, false},
+	{'a', encodeQueryComponent, false},
+	{'a', encodeFragment, false},
+	{'a', encodeHost, false},
+	{'z', encodePath, false},
+	{'A', encodePath, false},
+	{'Z', encodePath, false},
+	{'0', encodePath, false},
+	{'9', encodePath, false},
+	{'-', encodePath, false},
+	{'-', encodeUserPassword, false},
+	{'-', encodeQueryComponent, false},
+	{'-', encodeFragment, false},
+	{'.', encodePath, false},
+	{'_', encodePath, false},
+	{'~', encodePath, false},
+
+	// User information (§3.2.1)
+	{':', encodeUserPassword, true},
+	{'/', encodeUserPassword, true},
+	{'?', encodeUserPassword, true},
+	{'@', encodeUserPassword, true},
+	{'$', encodeUserPassword, false},
+	{'&', encodeUserPassword, false},
+	{'+', encodeUserPassword, false},
+	{',', encodeUserPassword, false},
+	{';', encodeUserPassword, false},
+	{'=', encodeUserPassword, false},
+
+	// Host (IP address, IPv6 address, registered name, port suffix; §3.2.2)
+	{'!', encodeHost, false},
+	{'$', encodeHost, false},
+	{'&', encodeHost, false},
+	{'\'', encodeHost, false},
+	{'(', encodeHost, false},
+	{')', encodeHost, false},
+	{'*', encodeHost, false},
+	{'+', encodeHost, false},
+	{',', encodeHost, false},
+	{';', encodeHost, false},
+	{'=', encodeHost, false},
+	{':', encodeHost, false},
+	{'[', encodeHost, false},
+	{']', encodeHost, false},
+	{'0', encodeHost, false},
+	{'9', encodeHost, false},
+	{'A', encodeHost, false},
+	{'z', encodeHost, false},
+	{'_', encodeHost, false},
+	{'-', encodeHost, false},
+	{'.', encodeHost, false},
+}
+
+func TestShouldEscape(t *testing.T) {
+	for _, tt := range shouldEscapeTests {
+		if shouldEscape(tt.in, tt.mode) != tt.escape {
+			t.Errorf("shouldEscape(%q, %v) returned %v; expected %v", tt.in, tt.mode, !tt.escape, tt.escape)
+		}
+	}
+}
+
+type timeoutError struct {
+	timeout bool
+}
+
+func (e *timeoutError) Error() string { return "timeout error" }
+func (e *timeoutError) Timeout() bool { return e.timeout }
+
+type temporaryError struct {
+	temporary bool
+}
+
+func (e *temporaryError) Error() string   { return "temporary error" }
+func (e *temporaryError) Temporary() bool { return e.temporary }
+
+type timeoutTemporaryError struct {
+	timeoutError
+	temporaryError
+}
+
+func (e *timeoutTemporaryError) Error() string { return "timeout/temporary error" }
+
+var netErrorTests = []struct {
+	err       error
+	timeout   bool
+	temporary bool
+}{{
+	err:       &Error{"Get", "http://google.com/", &timeoutError{timeout: true}},
+	timeout:   true,
+	temporary: false,
+}, {
+	err:       &Error{"Get", "http://google.com/", &timeoutError{timeout: false}},
+	timeout:   false,
+	temporary: false,
+}, {
+	err:       &Error{"Get", "http://google.com/", &temporaryError{temporary: true}},
+	timeout:   false,
+	temporary: true,
+}, {
+	err:       &Error{"Get", "http://google.com/", &temporaryError{temporary: false}},
+	timeout:   false,
+	temporary: false,
+}, {
+	err:       &Error{"Get", "http://google.com/", &timeoutTemporaryError{timeoutError{timeout: true}, temporaryError{temporary: true}}},
+	timeout:   true,
+	temporary: true,
+}, {
+	err:       &Error{"Get", "http://google.com/", &timeoutTemporaryError{timeoutError{timeout: false}, temporaryError{temporary: true}}},
+	timeout:   false,
+	temporary: true,
+}, {
+	err:       &Error{"Get", "http://google.com/", &timeoutTemporaryError{timeoutError{timeout: true}, temporaryError{temporary: false}}},
+	timeout:   true,
+	temporary: false,
+}, {
+	err:       &Error{"Get", "http://google.com/", &timeoutTemporaryError{timeoutError{timeout: false}, temporaryError{temporary: false}}},
+	timeout:   false,
+	temporary: false,
+}, {
+	err:       &Error{"Get", "http://google.com/", io.EOF},
+	timeout:   false,
+	temporary: false,
+}}
+
+// Test that url.Error implements net.Error and that it forwards
+func TestURLErrorImplementsNetError(t *testing.T) {
+	for i, tt := range netErrorTests {
+		err, ok := tt.err.(net.Error)
+		if !ok {
+			t.Errorf("%d: %T does not implement net.Error", i+1, tt.err)
+			continue
+		}
+		if err.Timeout() != tt.timeout {
+			t.Errorf("%d: err.Timeout(): got %v, want %v", i+1, err.Timeout(), tt.timeout)
+			continue
+		}
+		if err.Temporary() != tt.temporary {
+			t.Errorf("%d: err.Temporary(): got %v, want %v", i+1, err.Temporary(), tt.temporary)
+		}
+	}
+}
+
+var _ encodingPkg.BinaryMarshaler = (*URL)(nil)
+var _ encodingPkg.BinaryUnmarshaler = (*URL)(nil)
+
+func TestJSON(t *testing.T) {
+	var u URL
+	err := u.Parse("https://www.google.com/x?y=z")
+	if err != nil {
+		t.Fatal(err)
+	}
+	js, err := json.Marshal(&u)
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	// If only we could implement TextMarshaler/TextUnmarshaler,
+	// this would work:
+	//
+	// if string(js) != strconv.Quote(u.String()) {
+	// 	t.Errorf("json encoding: %s\nwant: %s\n", js, strconv.Quote(u.String()))
+	// }
+
+	u1 := new(URL)
+	err = json.Unmarshal(js, u1)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if u1.String() != u.String() {
+		t.Errorf("json decoded to: %s\nwant: %s\n", u1, &u)
+	}
+}
+
+func TestGob(t *testing.T) {
+	var u URL
+	err := u.Parse("https://www.google.com/x?y=z")
+	if err != nil {
+		t.Fatal(err)
+	}
+	var w bytes.Buffer
+	err = gob.NewEncoder(&w).Encode(&u)
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	u1 := new(URL)
+	err = gob.NewDecoder(&w).Decode(u1)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if u1.String() != u.String() {
+		t.Errorf("json decoded to: %s\nwant: %s\n", u1, &u)
+	}
+}

main.go

@@ -3,85 +3,141 @@ package main
 import (
 	"context"
 	"github.com/sirupsen/logrus"
+	"github.com/terorie/od-database-crawler/fasturl"
 	"github.com/urfave/cli"
-	"log"
-	"net/http"
-	_ "net/http/pprof"
-	"net/url"
 	"os"
 	"strings"
+	"sync/atomic"
 	"time"
 )
 
 var app = cli.App {
-	Name:         "oddb-go",
+	Name:         "od-database-crawler",
 	Usage:        "OD-Database Go crawler",
-	Version:      "0.1",
+	Version:      "1.0.1",
 	BashComplete: cli.DefaultAppComplete,
 	Writer:       os.Stdout,
-	Compiled:     buildDate,
+	Action:       cmdBase,
 	Commands:     []cli.Command{
 		{
 			Name:      "crawl",
 			Usage:     "Crawl a list of URLs",
-			ArgsUsage: "[site, site, ...]",
+			ArgsUsage: "<site>",
 			Action:    cmdCrawler,
 		},
 	},
+	After: func(i *cli.Context) error {
+		exitHooks.Execute()
+		return nil
+	},
 }
 
+var exitHooks Hooks
+
 func init() {
 	prepareConfig()
 }
 
 func main() {
-	go func() {
+	err := os.MkdirAll("crawled", 0755)
-		log.Println(http.ListenAndServe("localhost:42069", nil))
+	if err != nil {
-	}()
+		panic(err)
+	}
+
+	readConfig()
 	app.Run(os.Args)
 }
 
-func cmdCrawler(clic *cli.Context) error {
+func cmdBase(_ *cli.Context) error {
-	readConfig()
+	// TODO Graceful shutdown
-
+	appCtx := context.Background()
-	if clic.NArg() == 0 {
+	forceCtx := context.Background()
-		cli.ShowCommandHelpAndExit(clic, "crawl", 1)
-	}
-
-	args := clic.Args()
-	remotes := make([]*OD, len(args))
-	for i, arg := range args {
-		// https://github.com/golang/go/issues/19779
-		if !strings.Contains(arg, "://") {
-			arg = "http://" + arg
-		}
-		u, err := url.Parse(arg)
-		if !strings.HasSuffix(u.Path, "/") {
-			u.Path += "/"
-		}
-		if err != nil { return err }
-		remotes[i] = &OD{ BaseUri: *u }
-	}
-
-	c := context.Background()
 
 	inRemotes := make(chan *OD)
-	go Schedule(c, inRemotes)
+	go Schedule(forceCtx, inRemotes)
 
-	for _, remote := range remotes {
+	ticker := time.NewTicker(config.Recheck)
-		globalWait.Add(1)
+	defer ticker.Stop()
-		inRemotes <- remote
+	for {
+		select {
+		case <-appCtx.Done():
+			return nil
+		case <-ticker.C:
+			t, err := FetchTask()
+			if err != nil {
+				logrus.WithError(err).
+					Error("Failed to get new task")
+				time.Sleep(30 * time.Second)
+				continue
+			}
+			if t == nil {
+				// No new task
+				if atomic.LoadInt32(&numActiveTasks) == 0 {
+					logrus.Info("Waiting …")
+				}
+				continue
 			}
 
-	// Wait for all jobs to finish
+			var baseUri fasturl.URL
-	globalWait.Wait()
+			err = baseUri.Parse(t.Url)
+			if urlErr, ok := err.(*fasturl.Error); ok && urlErr.Err == fasturl.ErrUnknownScheme {
+				// Not an error
+				err = nil
 
-	logrus.Info("All dirs processed!")
+				// Give back task
+				//err2 := CancelTask(t.WebsiteId)
+				//if err2 != nil {
+				//	logrus.Error(err2)
+				//}
+
+				continue
+			} else if err != nil {
+				logrus.WithError(err).
+					Error("Failed to get new task")
+				time.Sleep(30 * time.Second)
+				continue
+			}
+			ScheduleTask(inRemotes, t, &baseUri)
+		}
+	}
 
 	return nil
 }
 
-var buildDate = time.Date(
+func cmdCrawler(clic *cli.Context) error {
-	2018, 10, 28,
+	if clic.NArg() != 1 {
-	17, 10, 0, 0,
+		cli.ShowCommandHelpAndExit(clic, "crawl", 1)
-	time.UTC)
+	}
+
+	arg := clic.Args()[0]
+	// https://github.com/golang/go/issues/19779
+	if !strings.Contains(arg, "://") {
+		arg = "http://" + arg
+	}
+	var u fasturl.URL
+	err := u.Parse(arg)
+	if !strings.HasSuffix(u.Path, "/") {
+		u.Path += "/"
+	}
+	if err != nil { return err }
+
+	// TODO Graceful shutdown
+	forceCtx := context.Background()
+
+	inRemotes := make(chan *OD)
+	go Schedule(forceCtx, inRemotes)
+
+	ticker := time.NewTicker(3 * time.Second)
+	defer ticker.Stop()
+
+	task := Task {
+		WebsiteId: 0,
+		Url: u.String(),
+	}
+	ScheduleTask(inRemotes, &task, &u)
+
+	// Wait for all jobs to finish
+	globalWait.Wait()
+
+	return nil
+}

model.go

@@ -1,32 +1,59 @@
 package main
 
 import (
-	"net/url"
+	"github.com/terorie/od-database-crawler/ds/redblackhash"
+	"github.com/terorie/od-database-crawler/fasturl"
 	"sync"
 	"time"
 )
 
+type Task struct {
+	WebsiteId uint64 `json:"website_id"`
+	Url       string `json:"url"`
+}
+
+type TaskResult struct {
+	StatusCode    string    `json:"status_code"`
+	FileCount     uint64    `json:"file_count"`
+	ErrorCount    uint64    `json:"-"`
+	StartTime     time.Time `json:"-"`
+	StartTimeUnix int64     `json:"start_time"`
+	EndTimeUnix   int64     `json:"end_time"`
+	WebsiteId     uint64    `json:"website_id"`
+}
+
 type Job struct {
 	OD        *OD
-	Uri       url.URL
+	Uri       fasturl.URL
 	UriStr    string
 	Fails     int
 	LastError error
 }
 
 type OD struct {
+	Task    Task
+	Result  TaskResult
 	Wait    sync.WaitGroup
-	BaseUri url.URL
+	BaseUri fasturl.URL
-	lock    sync.Mutex
-	Files   []File
 	WCtx    WorkerContext
-	Scanned sync.Map
+	Scanned redblackhash.Tree
 }
 
 type File struct {
 	Name  string `json:"name"`
 	Size  int64  `json:"size"`
-	MTime time.Time `json:"mtime"`
+	MTime int64  `json:"mtime"`
 	Path  string `json:"path"`
 	IsDir bool   `json:"-"`
 }
+
+func (o *OD) LoadOrStoreKey(k *redblackhash.Key) (exists bool) {
+	o.Scanned.Lock()
+	defer o.Scanned.Unlock()
+
+	exists = o.Scanned.Get(k)
+	if exists { return true }
+
+	o.Scanned.Put(k)
+	return false
+}

release.sh

@@ -0,0 +1,25 @@
+#!/usr/bin/env bash
+
+appname="od-database-crawler"
+tag=$1
+[ -z "$tag" ] && echo "Usage: build <version>" && exit 1
+
+name=${appname}-${tag}-windows.exe
+GOOS="windows" GOARCH="amd64" go build -ldflags="-s -w" -o $name
+gzip -f $name
+echo $name
+
+name=${appname}-${tag}-linux
+GOOS="linux" GOARCH="amd64" go build -ldflags="-s -w" -o $name
+gzip -f $name
+echo $name
+
+name=${appname}-${tag}-mac
+GOOS="darwin" GOARCH="amd64" go build -ldflags="-s -w" -o $name
+gzip -f $name
+echo $name
+
+name=${appname}-${tag}-freebsd
+GOOS="freebsd" GOARCH="amd64" go build -ldflags="-s -w" -o $name
+gzip -f $name
+echo $name

scheduler.go

@@ -2,55 +2,201 @@ package main
 
 import (
 	"context"
+	"encoding/json"
+	"fmt"
 	"github.com/sirupsen/logrus"
+	"github.com/terorie/od-database-crawler/fasturl"
+	"os"
+	"path"
+	"sync"
 	"sync/atomic"
+	"time"
 )
 
-var activeTasks int32
+var activeTasksLock sync.Mutex
+var activeTasks = make(map[uint64]bool)
+var numActiveTasks int32
 var totalBuffered int64
 
 func Schedule(c context.Context, remotes <-chan *OD) {
 	go Stats(c)
 
-	for {
+	for remote := range remotes {
-		select {
-		case <-c.Done():
-			return
-
-		case remote := <-remotes:
 		logrus.WithField("url", remote.BaseUri.String()).
 			Info("Starting crawler")
 
+		// Collect results
+		results := make(chan File)
+
 		// Spawn workers
 		remote.WCtx.in, remote.WCtx.out = makeJobBuffer(c)
 		for i := 0; i < config.Workers; i++ {
-				go remote.WCtx.Worker()
+			go remote.WCtx.Worker(results)
 		}
 
 		// Enqueue initial job
-			atomic.AddInt32(&activeTasks, 1)
+		atomic.AddInt32(&numActiveTasks, 1)
 		remote.WCtx.queueJob(Job{
 			OD:     remote,
 			Uri:    remote.BaseUri,
 			UriStr: remote.BaseUri.String(),
 			Fails:  0,
 		})
-			globalWait.Done()
 
 		// Upload result when ready
-			go remote.Watch()
+		go remote.Watch(results)
+
+		// Sleep if max number of tasks are active
+		for atomic.LoadInt32(&numActiveTasks) > config.Tasks {
+			select {
+			case <-c.Done():
+				return
+			case <-time.After(time.Second):
+				continue
+			}
 		}
 	}
 }
 
-func (r *OD) Watch() {
+func ScheduleTask(remotes chan<- *OD, t *Task, u *fasturl.URL) {
+	if !t.register() {
+		return
+	}
+
+	globalWait.Add(1)
+	now := time.Now()
+	od := &OD {
+		Task: *t,
+		BaseUri: *u,
+		Result: TaskResult {
+			WebsiteId: t.WebsiteId,
+			StartTime: now,
+			StartTimeUnix: now.Unix(),
+		},
+	}
+	remotes <- od
+}
+
+func (t *Task) register() bool {
+	activeTasksLock.Lock()
+	defer activeTasksLock.Unlock()
+
+	if _, known := activeTasks[t.WebsiteId]; known {
+		return false
+	} else {
+		activeTasks[t.WebsiteId] = true
+		return true
+	}
+}
+
+func (t *Task) unregister() {
+	activeTasksLock.Lock()
+	delete(activeTasks, t.WebsiteId)
+	activeTasksLock.Unlock()
+}
+
+func (o *OD) Watch(results chan File) {
+	// Mark job as completely done
+	defer globalWait.Done()
+	defer o.Task.unregister()
+
+	filePath := path.Join("crawled", fmt.Sprintf("%d.json", o.Task.WebsiteId))
+
+	// Open crawl results file
+	f, err := os.OpenFile(
+		filePath,
+		os.O_CREATE | os.O_RDWR | os.O_TRUNC,
+		0644,
+	)
+	if err != nil {
+		logrus.WithError(err).
+			Error("Failed saving crawl results")
+		return
+	}
+	defer f.Close()
+	defer os.Remove(filePath)
+
+	// Listen for exit code of Collect()
+	collectErrC := make(chan error)
+
+	// Block until all results are written
+	// (closes results channel)
+	o.handleCollect(results, f, collectErrC)
+
+	// Exit code of Collect()
+	err = <-collectErrC
+	close(collectErrC)
+	if err != nil {
+		logrus.WithError(err).
+			Error("Failed saving crawl results")
+		return
+	}
+
+	// Upload results
+	err = PushResult(&o.Result, f)
+	if err != nil {
+		logrus.WithError(err).
+			Error("Failed uploading crawl results")
+		return
+	}
+}
+
+func (o *OD) handleCollect(results chan File, f *os.File, collectErrC chan error) {
+	// Begin collecting results
+	go o.Task.Collect(results, f, collectErrC)
+	defer close(results)
+
 	// Wait for all jobs on remote to finish
-	r.Wait.Wait()
+	o.Wait.Wait()
-	close(r.WCtx.in)
+	close(o.WCtx.in)
-	atomic.AddInt32(&activeTasks, -1)
+	atomic.AddInt32(&numActiveTasks, -1)
 
-	logrus.WithField("url", r.BaseUri.String()).
+	// Log finish
-		Info("Crawler finished")
+
+	logrus.WithFields(logrus.Fields{
+		"id":  o.Task.WebsiteId,
+		"url": o.BaseUri.String(),
+		"duration": time.Since(o.Result.StartTime),
+	}).Info("Crawler finished")
+
+	// Set status code
+	now := time.Now()
+	o.Result.EndTimeUnix = now.Unix()
+	fileCount := atomic.LoadUint64(&o.Result.FileCount)
+	if fileCount == 0 {
+		errorCount := atomic.LoadUint64(&o.Result.ErrorCount)
+		if errorCount == 0 {
+			o.Result.StatusCode = "empty"
+		} else {
+			o.Result.StatusCode = "directory listing failed"
+		}
+	} else {
+		o.Result.StatusCode = "success"
+	}
+}
+
+func (t *Task) Collect(results chan File, f *os.File, errC chan<- error) {
+	err := t.collect(results, f)
+	if err != nil {
+		logrus.WithError(err).
+			Error("Failed saving crawl results")
+	}
+	errC <- err
+}
+
+func (t *Task) collect(results chan File, f *os.File) error {
+	for result := range results {
+		result.Path = fasturl.PathUnescape(result.Path)
+		result.Name = fasturl.PathUnescape(result.Name)
+		resJson, err := json.Marshal(result)
+		if err != nil { panic(err) }
+		_, err = f.Write(resJson)
+		if err != nil { return err }
+		_, err = f.Write([]byte{'\n'})
+		if err != nil { return err }
+	}
+
+	return nil
 }
 
 func makeJobBuffer(c context.Context) (chan<- Job, <-chan Job) {

server.go

@@ -10,34 +10,27 @@ import (
 	"net/http"
 	"net/url"
 	"os"
-	"path/filepath"
 	"strconv"
-	"strings"
 )
 
-const (
+var serverClient = http.Client {
-	fileListChunkSize int64 = 5000000 // 5 mb
+	Timeout: config.ServerTimeout,
-)
+}
-
-var serverClient = http.DefaultClient
 
 func FetchTask() (t *Task, err error) {
-	escToken, _ := json.Marshal(config.Token)
+	res, err := serverClient.PostForm(
-	payload := `{"token":` + string(escToken) + `}`
-
-	req, err := http.NewRequest(
-		http.MethodPost,
 		config.ServerUrl + "/task/get",
-		strings.NewReader(payload))
+		url.Values{ "token": {config.Token} })
-	if err != nil { return }
-
-	res, err := serverClient.Do(req)
 	if err != nil { return }
 	defer res.Body.Close()
 
-	if res.StatusCode != 200 {
+	switch res.StatusCode {
-		err = fmt.Errorf("http %s", res.Status)
+	case 200:
-		return
+		break
+	case 404, 500:
+		return nil, nil
+	default:
+		return nil, fmt.Errorf("http %s", res.Status)
 	}
 
 	t = new(Task)
@@ -47,21 +40,17 @@ func FetchTask() (t *Task, err error) {
 	return
 }
 
-func PushResult(result *TaskResult) (err error) {
+func PushResult(result *TaskResult, f *os.File) (err error) {
-	filePath := filepath.Join(
+	if result.WebsiteId == 0 {
-		".", "crawled",
+		// Not a real result, don't push
-		fmt.Sprintf("%d.json", result.WebsiteId))
+		return nil
+	}
 
-	defer os.Remove(filePath)
+	// Rewind to the beginning of the file
-
+	_, err = f.Seek(0, 0)
-	f, err := os.Open(filePath)
+	if err != nil {
-	if os.IsNotExist(err) {
-		err = fmt.Errorf("cannot upload result: %s does not exist", filePath)
-		return
-	} else if err != nil {
 		return
 	}
-	defer f.Close()
 
 	err = uploadChunks(result.WebsiteId, f)
 	if err != nil {
@@ -73,47 +62,50 @@ func PushResult(result *TaskResult) (err error) {
 		return
 	}
 
-	err = uploadResult(result)
+	// Upload result ignoring errors
-	if err != nil {
+	uploadResult(result)
-		logrus.Errorf("Failed to upload result: %s", err)
-		err2 := CancelTask(result.WebsiteId)
-		if err2 != nil {
-			logrus.Error(err2)
-		}
-		return
-	}
 
 	return
 }
 
-func uploadChunks(websiteId uint64, f *os.File) (err error) {
+func uploadChunks(websiteId uint64, f *os.File) error {
-	for iter := 1; iter > 0; iter++ {
+	eof := false
+	for iter := 1; !eof; iter++ {
 		// TODO Stream with io.Pipe?
 		var b bytes.Buffer
 
 		multi := multipart.NewWriter(&b)
 
 		// Set upload fields
+		var err error
 		err = multi.WriteField("token", config.Token)
-		if err != nil { return }
+		if err != nil { return err }
 		err = multi.WriteField("website_id", fmt.Sprintf("%d", websiteId))
-		if err != nil { return }
+		if err != nil { return err }
 
 		// Copy chunk to file_list
 		formFile, err := multi.CreateFormFile("file_list", "file_list")
-		_, err = io.CopyN(formFile, f, fileListChunkSize)
+		var n int64
-		if err == io.EOF {
+		n, err = io.CopyN(formFile, f, config.ChunkSize)
-			break
+		if err != io.EOF && err != nil {
-		} else if err == io.ErrUnexpectedEOF {
+			return err
+		}
+		if n == 0 {
+			// Don't upload, no content
+			return nil
+		} else if n < config.ChunkSize {
 			err = nil
 			// Break at end of iteration
-			iter = -420
+			eof = true
 		}
 
+		multi.Close()
+
 		req, err := http.NewRequest(
 			http.MethodPost,
 			config.ServerUrl + "/task/upload",
 			&b)
+		req.Header.Set("content-type", multi.FormDataContentType())
 		if err != nil { return err }
 
 		res, err := serverClient.Do(req)
@@ -125,52 +117,42 @@ func uploadChunks(websiteId uint64, f *os.File) (err error) {
 				iter, res.Status)
 		}
 
-		logrus.Infof("Uploading file list part %d: %s",
+		logrus.WithField("id", websiteId).
-			iter, res.Status)
+			WithField("part", iter).
+			Infof("Uploading files chunk")
 	}
-	return
+	return nil
 }
 
 func uploadResult(result *TaskResult) (err error) {
 	resultEnc, err := json.Marshal(result)
 	if err != nil { panic(err) }
 
-	payload := url.Values {
+	res, err := serverClient.PostForm(
+		config.ServerUrl + "/task/complete",
+		url.Values {
 			"token": {config.Token},
 			"result": {string(resultEnc)},
-	}.Encode()
+		},
-
+	)
-	req, err := http.NewRequest(
-		http.MethodPost,
-		config.ServerUrl + "/task/complete",
-		strings.NewReader(payload))
-	if err != nil { return }
-
-	res, err := serverClient.Do(req)
 	if err != nil { return }
 	res.Body.Close()
 
 	if res.StatusCode != http.StatusOK {
-		return fmt.Errorf("failed to cancel task: %s", res.Status)
+		return HttpError{res.StatusCode}
 	}
 
 	return
 }
 
 func CancelTask(websiteId uint64) (err error) {
-	form := url.Values{
+	res, err := serverClient.PostForm(
+		config.ServerUrl + "/task/cancel",
+		url.Values{
 			"token": {config.Token},
 			"website_id": {strconv.FormatUint(websiteId, 10)},
-	}
+		},
-	encForm := form.Encode()
+	)
-
-	req, err := http.NewRequest(
-		http.MethodPost,
-		config.ServerUrl + "/task/cancel",
-		strings.NewReader(encForm))
-	if err != nil { return }
-
-	res, err := serverClient.Do(req)
 	if err != nil { return }
 	res.Body.Close()
 

stats.go

@@ -39,6 +39,10 @@ func Stats(c context.Context) {
 			perSecond = math.Round(perSecond)
 			perSecond /= 2
 
+			if perSecond <= 0 {
+				continue
+			}
+
 			logrus.WithFields(logrus.Fields{
 				"per_second": perSecond,
 				"done":    atomic.LoadUint64(&totalDone),
@@ -53,7 +57,7 @@ func Stats(c context.Context) {
 			runtime.ReadMemStats(&mem)
 
 			logrus.WithFields(logrus.Fields{
-				"queue_count": totalBuffered,
+				"queue_count": atomic.LoadInt64(&totalBuffered),
 				"heap": FormatByteCount(mem.Alloc),
 				"objects": mem.HeapObjects,
 				"num_gc": mem.NumGC,

tasks.go

@@ -1,16 +0,0 @@
-package main
-
-import "time"
-
-type Task struct {
-	WebsiteId int `json:"website_id"`
-	Url string `json:"url"`
-}
-
-type TaskResult struct {
-	StatusCode int `json:"status_code"`
-	FileCount uint64 `json:"file_count"`
-	StartTime time.Time `json:"start_time"`
-	EndTime time.Time `json:"end_time"`
-	WebsiteId uint64 `json:"website_id"`
-}

util.go

@@ -1,6 +1,9 @@
 package main
 
-import "fmt"
+import (
+	"fmt"
+	"sync"
+)
 
 // https://programming.guide/go/formatting-byte-size-to-human-readable-format.html
 func FormatByteCount(b uint64) string {
@@ -16,3 +19,20 @@ func FormatByteCount(b uint64) string {
 		return fmt.Sprintf("%.1f %ciB", float64(b)/float64(div), "KMGTPE"[exp])
 	}
 }
+
+type Hooks struct {
+	m sync.Mutex
+	l []func()
+}
+
+func (h *Hooks) Add(hook func()) {
+	h.m.Lock()
+	h.l = append(h.l, hook)
+	h.m.Unlock()
+}
+
+func (h *Hooks) Execute() {
+	for _, hook := range h.l {
+		hook()
+	}
+}

worker.go

@@ -2,7 +2,9 @@ package main
 
 import (
 	"github.com/sirupsen/logrus"
+	"github.com/valyala/fasthttp"
 	"math"
+	"sort"
 	"strings"
 	"sync"
 	"sync/atomic"
@@ -18,13 +20,13 @@ type WorkerContext struct {
 	numRateLimits int
 }
 
-func (w WorkerContext) Worker() {
+func (w WorkerContext) Worker(results chan<- File) {
 	for job := range w.out {
-		w.step(job)
+		w.step(results, job)
 	}
 }
 
-func (w WorkerContext) step(job Job) {
+func (w WorkerContext) step(results chan<- File, job Job) {
 	defer w.finishJob(&job)
 
 	var f File
@@ -38,10 +40,15 @@ func (w WorkerContext) step(job Job) {
 	if err != nil {
 		job.Fails++
 
-		if err == ErrForbidden {
+		if httpErr, ok := err.(*HttpError); ok {
-			// Don't attempt crawling again
+			switch httpErr.code {
+			case fasthttp.StatusTooManyRequests:
+				err = ErrRateLimit
+			default:
+				// Don't retry HTTP error codes
 				return
 			}
+		}
 
 		if job.Fails > config.Retries {
 			atomic.AddUint64(&totalAborted, 1)
@@ -63,17 +70,22 @@ func (w WorkerContext) step(job Job) {
 		w.queueJob(job)
 	}
 
-	job.OD.Files = append(job.OD.Files, f)
+	if !f.IsDir {
+		results <- f
+	}
 }
 
 func DoJob(job *Job, f *File) (newJobs []Job, err error) {
-	if strings.HasSuffix(job.Uri.Path, "/") {
+	if len(job.Uri.Path) == 0 { return }
+	if job.Uri.Path[len(job.Uri.Path)-1] == '/' {
 		// Load directory
 		links, err := GetDir(job, f)
 		if err != nil {
+			if !isErrSilent(err) {
 				logrus.WithError(err).
-				WithField("url", job.Uri.String()).
+					WithField("url", job.UriStr).
-				Error("Failed getting dir")
+					Error("Failed to crawl dir")
+			}
 			return nil, err
 		}
 
@@ -81,43 +93,59 @@ func DoJob(job *Job, f *File) (newJobs []Job, err error) {
 		hash := f.HashDir(links)
 
 		// Skip symlinked dirs
-		if _, old := job.OD.Scanned.LoadOrStore(hash, true); old {
+		if job.OD.LoadOrStoreKey(&hash) {
 			return nil, ErrKnown
 		}
 
+		// Sort by path
+		sort.Slice(links, func(i, j int) bool {
+			return strings.Compare(links[i].Path, links[j].Path) < 0
+		})
+
+		var newJobCount int
+		var lastLink string
 		for _, link := range links {
-			// Skip already queued links
+			uriStr := link.String()
-			if _, old := job.OD.Scanned.LoadOrStore(link, true); old {
+
+			// Ignore dupes
+			if uriStr == lastLink {
 				continue
 			}
-			job.OD.Wait.Add(1)
+			lastLink = uriStr
+
 			newJobs = append(newJobs, Job{
 				OD:     job.OD,
 				Uri:    link,
-				UriStr: link.String(),
+				UriStr: uriStr,
 				Fails:  0,
 			})
+
+			newJobCount++
 		}
+		if config.Verbose {
 			logrus.WithFields(logrus.Fields{
 				"url":   job.UriStr,
-			"files": len(links),
+				"files": newJobCount,
 			}).Debug("Listed")
+		}
 	} else {
 		// Load file
 		err := GetFile(job.Uri, f)
 		if err != nil {
+			if !isErrSilent(err) {
 				logrus.WithError(err).
-				WithField("url", job.Uri.String()).
+					WithField("url", job.UriStr).
-				Error("Failed getting file")
+					Error("Failed to crawl file")
+			}
 			return nil, err
 		}
+		atomic.AddUint64(&job.OD.Result.FileCount, 1)
 	}
 	return
 }
 
 func (w WorkerContext) queueJob(job Job) {
 	job.OD.Wait.Add(1)
-	globalWait.Add(1)
 
 	if w.numRateLimits > 0 {
 		if time.Since(w.lastRateLimit) > 5 * time.Second {
@@ -125,14 +153,21 @@ func (w WorkerContext) queueJob(job Job) {
 		} else {
 			time.Sleep(time.Duration(math.Sqrt(float64(50 * w.numRateLimits))) *
 				100 * time.Millisecond)
-			w.in <- job
 		}
-	} else {
-		w.in <- job
 	}
+
+	w.in <- job
 }
 
 func (w WorkerContext) finishJob(job *Job) {
 	job.OD.Wait.Done()
-	globalWait.Done()
+}
+
+func isErrSilent(err error) bool {
+	if !config.PrintHTTP {
+		if _, ok := err.(*HttpError); ok {
+			return true
+		}
+	}
+	return false
 }