kill perf

.gitignore

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

README.md

@@ -1,7 +1,2 @@
-# od-database Go crawler 🚀
+# oddb 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

@@ -1,64 +1,42 @@
 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() {
@@ -71,24 +49,14 @@ 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 {
@@ -105,10 +73,6 @@ 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)
@@ -117,19 +81,6 @@ 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) {
@@ -137,7 +88,7 @@ func configMissing(key string) {
 	os.Exit(1)
 }
 
-func configOOB(key string, v interface{}) {
+func configOOB(key string, v int) {
-	fmt.Fprintf(os.Stderr, "config: illegal value %v for key %s!\n", v, key)
+	fmt.Fprintf(os.Stderr, "config: illegal value %d for %key!\n", v, key)
 	os.Exit(1)
 }

config.yml

@@ -1,60 +1,26 @@
 # OD-Database server settings
 server:
   # Connection URL
-  url: http://od-db.mine.terorie.com/api
+  url: localhost:6969
-
   # 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: 10s
+  resource_stats: 1s
-
   # 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 processed at once
+  # Number of sites that can be
-  tasks: 100
+  # processed at once
-
+  tasks: 3
   # Number of connections per site
-  # Please be careful with this setting!
+  connections: 2
-  # 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

@@ -2,114 +2,116 @@ package main
 
 import (
 	"bytes"
-	"crypto/tls"
+	"fmt"
-	"github.com/terorie/od-database-crawler/ds/redblackhash"
+	"github.com/sirupsen/logrus"
-	"github.com/terorie/od-database-crawler/fasturl"
+	"github.com/terorie/oddb-go/ds/redblackhash"
+	"github.com/terorie/oddb-go/fasturl"
+	"github.com/terorie/oddb-go/runes"
+	"github.com/terorie/oddb-go/runespath"
 	"github.com/valyala/fasthttp"
 	"golang.org/x/crypto/blake2b"
 	"golang.org/x/net/html"
+	"golang.org/x/net/html/atom"
 	"path"
 	"strconv"
 	"strings"
 	"time"
 )
 
-var client = fasthttp.Client {
+var client fasthttp.Client
-	TLSConfig: &tls.Config{
-		InsecureSkipVerify: true,
-	},
-}
 
 func GetDir(j *Job, f *File) (links []fasturl.URL, err error) {
 	f.IsDir = true
-	f.Name = path.Base(j.Uri.Path)
+	f.Name = runespath.Base(j.Uri.Path)
 
 	req := fasthttp.AcquireRequest()
-	if config.UserAgent != "" {
-		req.Header.SetUserAgent(config.UserAgent)
-	}
 	req.SetRequestURI(j.UriStr)
 
 	res := fasthttp.AcquireResponse()
 	defer fasthttp.ReleaseResponse(res)
 
-	err = client.DoTimeout(req, res, config.Timeout)
+	err = client.Do(req, res)
 	fasthttp.ReleaseRequest(req)
 
 	if err != nil {
+		logrus.Error(err)
 		return
 	}
 
 	err = checkStatusCode(res.StatusCode())
-	if err != nil {
+	if err != nil { return }
-		return
-	}
 
 	body := res.Body()
 	doc := html.NewTokenizer(bytes.NewReader(body))
 
 	var linkHref string
+	var linkTexts []string
 	for {
-		err = nil
-
 		tokenType := doc.Next()
+		token := doc.Token()
 		if tokenType == html.ErrorToken {
 			break
 		}
 
 		switch tokenType {
 		case html.StartTagToken:
-			name, hasAttr := doc.TagName()
+			if token.DataAtom == atom.A {
-			if len(name) == 1 && name[0] == 'a' {
+				for _, attr := range token.Attr {
-				for hasAttr {
+					if attr.Key == "href" {
-					var ks, vs []byte
+						linkHref = attr.Val
-					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:
-			name, _ := doc.TagName()
+			if linkHref != "" && token.DataAtom == atom.A {
-			if len(name) == 1 && name[0] == 'a' {
 				// Copy params
 				href := linkHref
+				linkText := strings.Join(linkTexts, " ")
 
 				// Reset params
 				linkHref = ""
+				linkTexts = nil
 
-				if strings.LastIndexByte(href, '?') != -1 {
+				// TODO Optimized decision tree
-					continue
+				for _, entry := range urlBlackList {
+					if href == entry {
+						goto nextToken
 					}
-
-				switch href {
-				case "", " ", ".", "..", "/":
-					continue
 				}
-
+				for _, entry := range urlPartBlackList {
-				if strings.Contains(href, "../") {
+					if strings.Contains(href, entry) {
-					continue
+						goto nextToken
+					}
+				}
+				for _, entry := range fileNameBlackList {
+					if strings.Contains(linkText, entry) {
+						goto nextToken
+					}
 				}
 
 				var link fasturl.URL
-				err = j.Uri.ParseRel(&link, href)
+				err = j.Uri.ParseRel(&link, []rune(href))
-				if err != nil {
+				if err != nil { continue }
-					continue
-				}
 
-				if link.Scheme != j.Uri.Scheme ||
+				if !runes.Equals(link.Scheme, j.Uri.Scheme) ||
-					link.Host != j.Uri.Host ||
+					!runes.Equals(link.Host, j.Uri.Host) ||
-					link.Path == j.Uri.Path ||
+					 runes.Equals(link.Path, j.Uri.Path) ||
-					!strings.HasPrefix(link.Path, j.Uri.Path) {
+					!runes.HasPrefix(link.Path, j.Uri.Path) {
 					continue
 				}
 
 				links = append(links, link)
 			}
 		}
+
+		nextToken:
 	}
 
 	return
@@ -117,87 +119,102 @@ func GetDir(j *Job, f *File) (links []fasturl.URL, err error) {
 
 func GetFile(u fasturl.URL, f *File) (err error) {
 	f.IsDir = false
-	u.Path = path.Clean(u.Path)
+	u.Path = []rune(path.Clean(string(u.Path)))
-	f.Name = path.Base(u.Path)
+	f.Name = runespath.Base(u.Path)
-	f.Path = strings.Trim(path.Dir(u.Path), "/")
+	f.Path = runes.TrimRune(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.DoTimeout(req, res, config.Timeout)
+	err = client.Do(req, res)
 	fasthttp.ReleaseRequest(req)
 
-	if err != nil {
+	if err != nil { return }
-		return
-	}
 
 	err = checkStatusCode(res.StatusCode())
-	if err != nil {
+	if err != nil { return }
-		return
-	}
 
-	f.applyContentLength(string(res.Header.Peek("content-length")))
+	// TODO Inefficient af
-	f.applyLastModified(string(res.Header.Peek("last-modified")))
+	header := res.Header.Header()
+	f.ParseHeader(header)
 
 	return nil
 }
 
 func (f *File) HashDir(links []fasturl.URL) (o redblackhash.Key) {
 	h, _ := blake2b.New256(nil)
-	h.Write([]byte(f.Name))
+	h.Write([]byte(string(f.Name)))
 	for _, link := range links {
-		fileName := path.Base(link.Path)
+		fileName := runespath.Base(link.Path)
-		h.Write([]byte(fileName))
+		h.Write([]byte(string(fileName)))
 	}
 	sum := h.Sum(nil)
 	copy(o[:redblackhash.KeySize], sum)
 	return
 }
 
-func (f *File) applyContentLength(v string) {
+func (f *File) ParseHeader(h []byte) {
-	if v == "" {
+	var k1, k2 int
-		return
+	var v1, v2 int
-	}
+
-	size, err := strconv.ParseInt(v, 10, 64)
+	// Simple finite state machine
-	if err != nil {
+	state := 0
-		return
+	for i, b := range h {
-	}
+		switch state {
-	if size < 0 {
+		case 0:
-		return
+			if b == byte(':') {
-	}
+				state = 1
-	f.Size = size
+				k2 = i
 			}
 
-// TODO Cleanup
+		case 1:
-func (f *File) applyLastModified(v string) {
+			state = 2
-	if v == "" {
+
-		return
+		case 2:
+			state = 3
+			v1 = i
+
+		case 3:
+			if b == byte('\r') {
+				state = 4
 			}
-	var t time.Time
+
+		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 size < 0 { break }
+		f.Size = size
+
+	case "last-modified":
 		var err error
-	t, err = time.Parse(time.RFC1123, v)
+		f.MTime, err = time.Parse(time.RFC1123, v)
-	if err == nil {
+		if err == nil { break }
-		f.MTime = t.Unix()
+		f.MTime, err = time.Parse(time.RFC850, v)
-		return
+		if err == nil { break }
-	}
-	t, err = time.Parse(time.RFC850, v)
-	if err == nil {
-		f.MTime = t.Unix()
-		return
-	}
 		// TODO Parse asctime
-	t, err = time.Parse("2006-01-02", v[:10])
+		f.MTime, err = time.Parse("2006-01-02", v[:10])
-	if err == nil {
+		if err == nil { break }
-		f.MTime = t.Unix()
-		return
 	}
 }
 
@@ -205,7 +222,53 @@ 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 &HttpError{status}
+		return fmt.Errorf("got HTTP status %d", 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

@@ -14,9 +14,7 @@
 package redblackhash
 
 import (
-	"bytes"
 	"fmt"
-	"sync"
 )
 
 const (
@@ -29,7 +27,6 @@ type Key [KeySize]byte
 
 // Tree holds elements of the red-black tree
 type Tree struct {
-	sync.Mutex
 	Root       *Node
 	size       int
 }
@@ -44,7 +41,42 @@ type Node struct {
 }
 
 func (k *Key) Compare(o *Key) int {
-	return bytes.Compare(k[:], o[:])
+	// TODO Assembly
+	/*for i := 0; i < KeySize / 8; i++ {
+		a := uint64(k[i+0]      ) |
+			 uint64(k[i+1] >>  8) |
+			 uint64(k[i+2] >> 16) |
+			 uint64(k[i+3] >> 24) |
+			 uint64(k[i+4] >> 32) |
+			 uint64(k[i+5] >> 40) |
+			 uint64(k[i+6] >> 48) |
+			 uint64(k[i+7] >> 56)
+
+		b := uint64(o[i+0]      ) |
+			 uint64(o[i+1] >>  8) |
+			 uint64(o[i+2] >> 16) |
+			 uint64(o[i+3] >> 24) |
+			 uint64(o[i+4] >> 32) |
+			 uint64(o[i+5] >> 40) |
+			 uint64(o[i+6] >> 48) |
+			 uint64(o[i+7] >> 56)
+
+		switch {
+		case a < b:
+			return -1
+		case a > b:
+			return 1
+		}
+	}*/
+	for i := 0; i < KeySize; i++ {
+		switch {
+		case k[i] < o[i]:
+			return -1
+		case k[i] > o[i]:
+			return 1
+		}
+	}
+	return 0
 }
 
 // Put inserts node into the tree.

errors.go

@@ -1,17 +1,8 @@
 package main
 
-import (
+import "errors"
-	"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

@@ -15,34 +15,19 @@ package fasturl
 import (
 	"errors"
 	"fmt"
+	"github.com/terorie/oddb-go/runes"
 	"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
+	URL []rune
 	Err error
 }
 
-func (e *Error) Error() string { return e.Op + " " + e.URL + ": " + e.Err.Error() }
+func (e *Error) Error() string { return e.Op + " " + string(e.URL) + ": " + e.Err.Error() }
 
 type timeout interface {
 	Timeout() bool
@@ -115,7 +100,7 @@ func (e InvalidHostError) Error() string {
 //
 // 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 {
+func shouldEscape(c rune, mode encoding) bool {
 	// §2.3 Unreserved characters (alphanum)
 	if 'A' <= c && c <= 'Z' || 'a' <= c && c <= 'z' || '0' <= c && c <= '9' {
 		return false
@@ -192,9 +177,29 @@ func shouldEscape(c byte, mode encoding) bool {
 	return true
 }
 
+// QueryUnescape does the inverse transformation of QueryEscape,
+// converting each 3-byte encoded substring of the form "%AB" into the
+// hex-decoded byte 0xAB.
+// It returns an error if any % is not followed by two hexadecimal
+// digits.
+func QueryUnescape(s []rune) ([]rune, error) {
+	return unescape(s, encodeQueryComponent)
+}
+
+// PathUnescape does the inverse transformation of PathEscape,
+// converting each 3-byte encoded substring of the form "%AB" into the
+// hex-decoded byte 0xAB. It returns an error if any % is not followed
+// by two hexadecimal digits.
+//
+// PathUnescape is identical to QueryUnescape except that it does not
+// unescape '+' to ' ' (space).
+func PathUnescape(s []rune) ([]rune, error) {
+	return unescape(s, encodePathSegment)
+}
+
 // unescape unescapes a string; the mode specifies
 // which section of the URL string is being unescaped.
-func unescape(s string, mode encoding) (string, error) {
+func unescape(s []rune, mode encoding) ([]rune, error) {
 	// Count %, check that they're well-formed.
 	n := 0
 	hasPlus := false
@@ -202,12 +207,12 @@ func unescape(s string, mode encoding) (string, error) {
 		switch s[i] {
 		case '%':
 			n++
-			if i+2 >= len(s) || !ishex(s[i+1]) || !ishex(s[i+2]) {
+			if i+2 >= len(s) || !ishex(byte(s[i+1])) || !ishex(byte(s[i+2])) {
 				s = s[i:]
 				if len(s) > 3 {
 					s = s[:3]
 				}
-				return "", EscapeError(s)
+				return nil, EscapeError(s)
 			}
 			// Per https://tools.ietf.org/html/rfc3986#page-21
 			// in the host component %-encoding can only be used
@@ -215,8 +220,8 @@ func unescape(s string, mode encoding) (string, error) {
 			// 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" {
+			if mode == encodeHost && unhex(byte(s[i+1])) < 8 && !runes.Equals(s[i:i+3], []rune("%25")) {
-				return "", EscapeError(s[i : i+3])
+				return nil, EscapeError(s[i : i+3])
 			}
 			if mode == encodeZone {
 				// RFC 6874 says basically "anything goes" for zone identifiers
@@ -226,9 +231,9 @@ func unescape(s string, mode encoding) (string, error) {
 				// 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])
+				v := unhex(byte(s[i+1]))<<4 | unhex(byte(s[i+2]))
-				if s[i:i+3] != "%25" && v != ' ' && shouldEscape(v, encodeHost) {
+				if !runes.Equals(s[i:i+3], []rune("%25")) && v != ' ' && shouldEscape(rune(v), encodeHost) {
-					return "", EscapeError(s[i : i+3])
+					return nil, EscapeError(s[i : i+3])
 				}
 			}
 			i += 3
@@ -237,7 +242,7 @@ func unescape(s string, mode encoding) (string, error) {
 			i++
 		default:
 			if (mode == encodeHost || mode == encodeZone) && s[i] < 0x80 && shouldEscape(s[i], mode) {
-				return "", InvalidHostError(s[i : i+1])
+				return nil, InvalidHostError(s[i : i+1])
 			}
 			i++
 		}
@@ -252,7 +257,7 @@ func unescape(s string, mode encoding) (string, error) {
 	for i := 0; i < len(s); {
 		switch s[i] {
 		case '%':
-			t[j] = unhex(s[i+1])<<4 | unhex(s[i+2])
+			t[j] = unhex(byte(s[i+1]))<<4 | unhex(byte(s[i+2]))
 			j++
 			i += 3
 		case '+':
@@ -264,15 +269,27 @@ func unescape(s string, mode encoding) (string, error) {
 			j++
 			i++
 		default:
-			t[j] = s[i]
+			t[j] = byte(s[i])
 			j++
 			i++
 		}
 	}
-	return string(t), nil
+	return []rune(string(t)), nil
 }
 
-func escape(s string, mode encoding) string {
+// QueryEscape escapes the string so it can be safely placed
+// inside a URL query.
+func QueryEscape(s []rune) []rune {
+	return escape(s, encodeQueryComponent)
+}
+
+// PathEscape escapes the string so it can be safely placed
+// inside a URL path segment.
+func PathEscape(s []rune) []rune {
+	return escape(s, encodePathSegment)
+}
+
+func escape(s []rune, mode encoding) []rune {
 	spaceCount, hexCount := 0, 0
 	for i := 0; i < len(s); i++ {
 		c := s[i]
@@ -302,11 +319,11 @@ func escape(s string, mode encoding) string {
 			t[j+2] = "0123456789ABCDEF"[c&15]
 			j += 3
 		default:
-			t[j] = s[i]
+			t[j] = byte(s[i])
 			j++
 		}
 	}
-	return string(t)
+	return []rune(string(t))
 }
 
 // A URL represents a parsed URL (technically, a URI reference).
@@ -327,15 +344,19 @@ func escape(s string, mode encoding) string {
 // 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
+	Scheme     []rune
-	Host       string    // host or host:port
+	Opaque     []rune    // encoded opaque data
-	Path       string    // path (relative paths may omit leading slash)
+	Host       []rune    // host or host:port
+	Path       []rune    // path (relative paths may omit leading slash)
+	RawPath    []rune    // encoded path hint (see EscapedPath method)
+	ForceQuery bool      // append a query ('?') even if RawQuery is empty
+	RawQuery   []rune    // encoded query values, without '?'
 }
 
 // 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) {
+func getscheme(rawurl []rune) (scheme []rune, path []rune, err error) {
 	for i := 0; i < len(rawurl); i++ {
 		c := rawurl[i]
 		switch {
@@ -343,42 +364,34 @@ func getscheme(rawurl string) (scheme Scheme, path string, err error) {
 			// do nothing
 		case '0' <= c && c <= '9' || c == '+' || c == '-' || c == '.':
 			if i == 0 {
-				return SchemeInvalid, rawurl, nil
+				return nil, rawurl, nil
 			}
 		case c == ':':
 			if i == 0 {
-				return SchemeInvalid, "", errors.New("missing protocol scheme")
+				return nil, nil, errors.New("missing protocol scheme")
 			}
-			switch rawurl[:i] {
+			scheme = 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 nil, rawurl, nil
 		}
 	}
-	return SchemeInvalid, rawurl, nil
+	return nil, 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) {
+func split(s []rune, c rune, cutc bool) ([]rune, []rune) {
-	i := strings.Index(s, c)
+	i := strings.Index(string(s), string(c)) // TODO Optimize
 	if i < 0 {
-		return s, ""
+		return s, nil
 	}
 	if cutc {
-		return s[:i], s[i+len(c):]
+		return s[:i], s[i+1:]
 	}
 	return s[:i], s[i:]
 }
@@ -389,14 +402,14 @@ func split(s string, c string, cutc bool) (string, string) {
 // (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 {
+func (u *URL) Parse(rawurl []rune) error {
 	// Cut off #frag
-	s, frag := split(rawurl, "#", true)
+	s, frag := split(rawurl, '#', true)
 	err := u.parse(s, false)
 	if err != nil {
 		return &Error{"parse", s, err}
 	}
-	if frag == "" {
+	if len(frag) == 0 {
 		return nil
 	}
 	return nil
@@ -407,7 +420,7 @@ func (u *URL) Parse(rawurl string) error {
 // 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 {
+func (u *URL) ParseRequestURI(rawurl []rune) error {
 	err := u.parse(rawurl, true)
 	if err != nil {
 		return &Error{"parse", rawurl, err}
@@ -419,16 +432,16 @@ func (u *URL) ParseRequestURI(rawurl string) error {
 // 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 {
+func (u *URL) parse(rawurl []rune, viaRequest bool) error {
-	var rest string
+	var rest []rune
 	var err error
 
-	if rawurl == "" && viaRequest {
+	if len(rawurl) == 0 && viaRequest {
 		return errors.New("empty url")
 	}
 
-	if rawurl == "*" {
+	if runes.Equals(rawurl, []rune("*")) {
-		u.Path = "*"
+		u.Path = []rune("*")
 		return nil
 	}
 
@@ -438,15 +451,17 @@ func (u *URL) parse(rawurl string, viaRequest bool) error {
 		return err
 	}
 
-	if strings.HasSuffix(rest, "?") && strings.Count(rest, "?") == 1 {
+	if runes.HasSuffix(rest, []rune("?")) && runes.Count(rest, '?') == 1 {
+		u.ForceQuery = true
 		rest = rest[:len(rest)-1]
 	} else {
-		rest, _ = split(rest, "?", true)
+		rest, u.RawQuery = split(rest, '?', true)
 	}
 
-	if !strings.HasPrefix(rest, "/") {
+	if !runes.HasPrefix(rest, []rune("/")) {
-		if u.Scheme != SchemeInvalid {
+		if len(u.Scheme) != 0 {
 			// We consider rootless paths per RFC 3986 as opaque.
+			u.Opaque = rest
 			return nil
 		}
 		if viaRequest {
@@ -459,59 +474,65 @@ func (u *URL) parse(rawurl string, viaRequest bool) error {
 		// 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, ":")
+		colon := runes.IndexRune(rest, ':')
-		slash := strings.Index(rest, "/")
+		slash := runes.IndexRune(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, "//") {
+	if (len(u.Scheme) != 0 || !viaRequest && !runes.HasPrefix(rest, []rune("///"))) && runes.HasPrefix(rest, []rune("//")) {
-		var authority string
+		var authority []rune
-		authority, rest = split(rest[2:], "/", false)
+		authority, rest = split(rest[2:], '/', false)
 		u.Host, err = parseAuthority(authority)
 		if err != nil {
 			return err
 		}
 	}
-	u.Path = rest
+	// Set Path and, optionally, RawPath.
+	// RawPath is a hint of the encoding of Path. We don't want to set it if
+	// the default escaping of Path is equivalent, to help make sure that people
+	// don't rely on it in general.
+	if err := u.setPath(rest); err != nil {
+		return err
+	}
 	return nil
 }
 
-func parseAuthority(authority string) (host string, err error) {
+func parseAuthority(authority []rune) (host []rune, err error) {
-	i := strings.LastIndex(authority, "@")
+	i := runes.LastIndexRune(authority, '@')
 	if i < 0 {
 		host, err = parseHost(authority)
 	} else {
 		host, err = parseHost(authority[i+1:])
 	}
 	if err != nil {
-		return "", err
+		return nil, err
 	}
 	if i < 0 {
 		return host, nil
 	}
 	userinfo := authority[:i]
 	if !validUserinfo(userinfo) {
-		return "", errors.New("fasturl: invalid userinfo")
+		return nil, 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) {
+func parseHost(host []rune) ([]rune, error) {
-	if strings.HasPrefix(host, "[") {
+	if runes.HasPrefix(host, []rune("[")) {
 		// Parse an IP-Literal in RFC 3986 and RFC 6874.
 		// E.g., "[fe80::1]", "[fe80::1%25en0]", "[fe80::1]:80".
-		i := strings.LastIndex(host, "]")
+		i := runes.LastIndexRune(host, ']')
 		if i < 0 {
-			return "", errors.New("missing ']' in host")
+			return nil, errors.New("missing ']' in host")
 		}
 		colonPort := host[i+1:]
 		if !validOptionalPort(colonPort) {
-			return "", fmt.Errorf("invalid port %q after host", colonPort)
+			return nil, fmt.Errorf("invalid port %q after host", colonPort)
 		}
 
 		// RFC 6874 defines that %25 (%-encoded percent) introduces
@@ -520,35 +541,106 @@ func parseHost(host string) (string, error) {
 		// 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")
+		zone := strings.Index(string(host[:i]), "%25")
 		if zone >= 0 {
 			host1, err := unescape(host[:zone], encodeHost)
 			if err != nil {
-				return "", err
+				return nil, err
 			}
 			host2, err := unescape(host[zone:i], encodeZone)
 			if err != nil {
-				return "", err
+				return nil, err
 			}
 			host3, err := unescape(host[i:], encodeHost)
 			if err != nil {
-				return "", err
+				return nil, err
 			}
-			return host1 + host2 + host3, nil
+			// TODO Optimize
+			return runes.Create(host1, host2, host3), nil
 		}
 	}
 
 	var err error
 	if host, err = unescape(host, encodeHost); err != nil {
-		return "", err
+		return nil, err
 	}
 	return host, nil
 }
 
+// setPath sets the Path and RawPath fields of the URL based on the provided
+// escaped path p. It maintains the invariant that RawPath is only specified
+// when it differs from the default encoding of the path.
+// For example:
+// - setPath("/foo/bar")   will set Path="/foo/bar" and RawPath=""
+// - setPath("/foo%2fbar") will set Path="/foo/bar" and RawPath="/foo%2fbar"
+// setPath will return an error only if the provided path contains an invalid
+// escaping.
+func (u *URL) setPath(p []rune) error {
+	path, err := unescape(p, encodePath)
+	if err != nil {
+		return err
+	}
+	u.Path = path
+	if escp := escape(path, encodePath); runes.Equals(p, escp) {
+		// Default encoding is fine.
+		u.RawPath = nil
+	} else {
+		u.RawPath = p
+	}
+	return nil
+}
+
+// EscapedPath returns the escaped form of u.Path.
+// In general there are multiple possible escaped forms of any path.
+// EscapedPath returns u.RawPath when it is a valid escaping of u.Path.
+// Otherwise EscapedPath ignores u.RawPath and computes an escaped
+// form on its own.
+// The String and RequestURI methods use EscapedPath to construct
+// their results.
+// In general, code should call EscapedPath instead of
+// reading u.RawPath directly.
+func (u *URL) EscapedPath() []rune {
+	if len(u.RawPath) != 0 && validEncodedPath(u.RawPath) {
+		p, err := unescape(u.RawPath, encodePath)
+		if err == nil && runes.Equals(p, u.Path) {
+			return u.RawPath
+		}
+	}
+	if runes.Equals(u.Path, []rune("*")) {
+		return []rune("*") // don't escape (Issue 11202)
+	}
+	return escape(u.Path, encodePath)
+}
+
+// validEncodedPath reports whether s is a valid encoded path.
+// It must not contain any bytes that require escaping during path encoding.
+func validEncodedPath(s []rune) bool {
+	for i := 0; i < len(s); i++ {
+		// RFC 3986, Appendix A.
+		// pchar = unreserved / pct-encoded / sub-delims / ":" / "@".
+		// shouldEscape is not quite compliant with the RFC,
+		// so we check the sub-delims ourselves and let
+		// shouldEscape handle the others.
+		switch s[i] {
+		case '!', '$', '&', '\'', '(', ')', '*', '+', ',', ';', '=', ':', '@':
+			// ok
+		case '[', ']':
+			// ok - not specified in RFC 3986 but left alone by modern browsers
+		case '%':
+			// ok - percent encoded, will decode
+		default:
+			if shouldEscape(s[i], encodePath) {
+				return false
+			}
+		}
+	}
+	return true
+}
+
 // validOptionalPort reports whether port is either an empty string
 // or matches /^:\d*$/
-func validOptionalPort(port string) bool {
+func validOptionalPort(port []rune) bool {
-	if port == "" {
+	if len(port) == 0 {
 		return true
 	}
 	if port[0] != ':' {
@@ -562,6 +654,46 @@ func validOptionalPort(port string) bool {
 	return true
 }
 
+func (u *URL) Runes() (buf []rune) {
+	if len(u.Scheme) != 0 {
+		buf = append(buf, u.Scheme...)
+		buf = append(buf, ':')
+	}
+	if len(u.Opaque) != 0 {
+		buf = append(buf, u.Opaque...)
+	} else {
+		if len(u.Scheme) != 0 || len(u.Host) != 0 {
+			if len(u.Host) != 0 || len(u.Path) != 0 {
+				buf = append(buf, '/', '/')
+			}
+			if h := u.Host; len(h) != 0 {
+				buf = append(buf, escape(h, encodeHost)...)
+			}
+		}
+		path := u.EscapedPath()
+		if len(path) != 0 && path[0] != '/' && len(u.Host) != 0 {
+			buf = append(buf, '/')
+		}
+		if len(buf) == 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 := runes.IndexRune(path, ':'); i > -1 && runes.IndexRune(path[:i], '/') == -1 {
+				buf = append(buf, '.', '/')
+			}
+		}
+		buf = append(buf, path...)
+	}
+	if u.ForceQuery || len(u.RawQuery) != 0 {
+		buf = append(buf, '?')
+		buf = append(buf, u.RawQuery...)
+	}
+	return
+}
+
 // String reassembles the URL into a valid URL string.
 // The general form of the result is one of:
 //
@@ -583,126 +715,28 @@ func validOptionalPort(port string) bool {
 //	- 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
+	return string(u.Runes())
-	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 {
+func resolvePath(base, ref []rune) []rune {
-	var full string
+	var full []rune
-	if ref == "" {
+	if len(ref) == 0 {
 		full = base
 	} else if ref[0] != '/' {
-		i := strings.LastIndex(base, "/")
+		// TODO Optimize
-		full = base[:i+1] + ref
+		i := strings.LastIndex(string(base), "/")
+		full = runes.Create(base[:i+1], ref)
 	} else {
 		full = ref
 	}
-	if full == "" {
+	if len(full) == 0 {
-		return ""
+		return nil
-	} else if full == "/" {
-		return "/"
 	}
-
+	var dst []string
-	dst := make([]rune, len(full))
+	// TODO Optimize
-	dst = dst[0:0]
+	src := strings.Split(string(full), "/")
-
-	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 ".":
@@ -717,21 +751,22 @@ func resolvePath(base, ref string) string {
 	}
 	if last := src[len(src)-1]; last == "." || last == ".." {
 		// Add final slash to the joined path.
-		dst = append(dst, "")
+		dst = append(dst, "") // TODO Wtf?
 	}
-	return "/" + strings.TrimPrefix(strings.Join(dst, "/"), "/")*/
+	// TODO Optimize
+	return []rune("/" + 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
+	return len(u.Scheme) != 0
 }
 
 // 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 {
+func (u *URL) ParseRel(out *URL, ref []rune) error {
 	var refurl URL
 
 	err := refurl.Parse(ref)
@@ -751,22 +786,92 @@ func (u *URL) ParseRel(out *URL, ref string) error {
 // ignores base and returns a copy of ref.
 func (u *URL) ResolveReference(url *URL, ref *URL) {
 	*url = *ref
-	if ref.Scheme == SchemeInvalid {
+	if len(ref.Scheme) == 0 {
 		url.Scheme = u.Scheme
 	}
-	if ref.Scheme != SchemeInvalid || ref.Host != "" {
+	if len(ref.Scheme) != 0 || len(ref.Host) != 0 {
 		// 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, "")
+		url.setPath(resolvePath(ref.EscapedPath(), nil))
 		return
 	}
+	if len(ref.Opaque) != 0 {
+		url.Host = nil
+		url.Path = nil
+		return
+	}
+	if len(ref.Path) == 0 && len(ref.RawQuery) == 0 {
+		url.RawQuery = u.RawQuery
+	}
 	// The "abs_path" or "rel_path" cases.
 	url.Host = u.Host
-	url.Path = resolvePath(u.Path, ref.Path)
+	url.setPath(resolvePath(u.EscapedPath(), ref.EscapedPath()))
 	return
 }
 
+// RequestURI returns the encoded path?query or opaque?query
+// string that would be used in an HTTP request for u.
+func (u *URL) RequestURI() []rune {
+	result := u.Opaque
+	if len(result) == 0 {
+		result = u.EscapedPath()
+		if len(result) == 0 {
+			result = []rune("/")
+		}
+	} else {
+		if runes.HasPrefix(result, []rune("//")) {
+			result = runes.Create(u.Scheme, []rune(":"), result)
+		}
+	}
+	if u.ForceQuery || len(u.RawQuery) != 0 {
+		result = append(result, '?')
+		result = append(result, u.RawQuery...)
+	}
+	return result
+}
+
+// Hostname returns u.Host, without any port number.
+//
+// If Host is an IPv6 literal with a port number, Hostname returns the
+// IPv6 literal without the square brackets. IPv6 literals may include
+// a zone identifier.
+func (u *URL) Hostname() []rune {
+	return stripPort(u.Host)
+}
+
+// Port returns the port part of u.Host, without the leading colon.
+// If u.Host doesn't contain a port, Port returns an empty string.
+func (u *URL) Port() []rune {
+	return portOnly(u.Host)
+}
+
+func stripPort(hostport []rune) []rune {
+	colon := runes.IndexRune(hostport, ':')
+	if colon == -1 {
+		return hostport
+	}
+	if i := runes.IndexRune(hostport, ']'); i != -1 {
+		return runes.TrimPrefix(hostport[:i], []rune("["))
+	}
+	return hostport[:colon]
+}
+
+func portOnly(hostport []rune) []rune {
+	colon := runes.IndexRune(hostport, ':')
+	if colon == -1 {
+		return nil
+	}
+	// TODO Optimize
+	if i := strings.Index(string(hostport), "]:"); i != -1 {
+		return hostport[i+len("]:"):]
+	}
+	if strings.Contains(string(hostport), "]") {
+		return nil
+	}
+	return hostport[colon+len(":"):]
+}
+
 // Marshaling interface implementations.
 // Would like to implement MarshalText/UnmarshalText but that will change the JSON representation of URLs.
 
@@ -776,7 +881,7 @@ func (u *URL) MarshalBinary() (text []byte, err error) {
 
 func (u *URL) UnmarshalBinary(text []byte) error {
 	var u1 URL
-	err := u1.Parse(string(text))
+	err := u1.Parse([]rune(string(text)))
 	if err != nil {
 		return err
 	}
@@ -792,7 +897,7 @@ func (u *URL) UnmarshalBinary(text []byte) error {
 //                   / "*" / "+" / "," / ";" / "="
 //
 // It doesn't validate pct-encoded. The caller does that via func unescape.
-func validUserinfo(s string) bool {
+func validUserinfo(s []rune) bool {
 	for _, r := range s {
 		if 'A' <= r && r <= 'Z' {
 			continue
@@ -813,57 +918,3 @@ func validUserinfo(s string) bool {
 	}
 	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

@@ -1,897 +0,0 @@
-// 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

@@ -3,141 +3,90 @@ package main
 import (
 	"context"
 	"github.com/sirupsen/logrus"
-	"github.com/terorie/od-database-crawler/fasturl"
+	"github.com/terorie/oddb-go/fasturl"
+	"github.com/terorie/oddb-go/runes"
 	"github.com/urfave/cli"
+	"log"
+	"net/http"
+	_ "net/http/pprof"
 	"os"
 	"strings"
-	"sync/atomic"
 	"time"
 )
 
 var app = cli.App {
-	Name:         "od-database-crawler",
+	Name:         "oddb-go",
 	Usage:        "OD-Database Go crawler",
-	Version:      "1.0.2",
+	Version:      "0.1",
 	BashComplete: cli.DefaultAppComplete,
 	Writer:       os.Stdout,
-	Action:       cmdBase,
+	Compiled:     buildDate,
 	Commands:     []cli.Command{
 		{
 			Name:      "crawl",
 			Usage:     "Crawl a list of URLs",
-			ArgsUsage: "<site>",
+			ArgsUsage: "[site, site, ...]",
 			Action:    cmdCrawler,
 		},
 	},
-	After: func(i *cli.Context) error {
-		exitHooks.Execute()
-		return nil
-	},
 }
 
-var exitHooks Hooks
-
 func init() {
 	prepareConfig()
 }
 
 func main() {
-	err := os.MkdirAll("crawled", 0755)
+	go func() {
-	if err != nil {
+		log.Println(http.ListenAndServe("localhost:42069", nil))
-		panic(err)
+	}()
-	}
-
-	readConfig()
 	app.Run(os.Args)
 }
 
-func cmdBase(_ *cli.Context) error {
-	// TODO Graceful shutdown
-	appCtx := context.Background()
-	forceCtx := context.Background()
-
-	inRemotes := make(chan *OD)
-	go Schedule(forceCtx, inRemotes)
-
-	ticker := time.NewTicker(config.Recheck)
-	defer ticker.Stop()
-	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
-			}
-
-			var baseUri fasturl.URL
-			err = baseUri.Parse(t.Url)
-			if urlErr, ok := err.(*fasturl.Error); ok && urlErr.Err == fasturl.ErrUnknownScheme {
-				// Not an error
-				err = nil
-
-				// 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
-}
-
 func cmdCrawler(clic *cli.Context) error {
-	if clic.NArg() != 1 {
+	readConfig()
+
+	if clic.NArg() == 0 {
 		cli.ShowCommandHelpAndExit(clic, "crawl", 1)
 	}
 
-	arg := clic.Args()[0]
+	args := clic.Args()
+	remotes := make([]*OD, len(args))
+	for i, argStr := range args {
 		// https://github.com/golang/go/issues/19779
-	if !strings.Contains(arg, "://") {
+		if !strings.Contains(argStr, "://") {
-		arg = "http://" + arg
+			argStr = "http://" + argStr
 		}
+
+		arg := []rune(argStr)
+
 		var u fasturl.URL
 		err := u.Parse(arg)
-	if !strings.HasSuffix(u.Path, "/") {
+		if !runes.HasSuffix(u.Path, []rune("/")) {
-		u.Path += "/"
+			u.Path = append(u.Path, '/')
 		}
 		if err != nil { return err }
+		remotes[i] = &OD{ BaseUri: u }
+	}
 
-	// TODO Graceful shutdown
+	c := context.Background()
-	forceCtx := context.Background()
 
 	inRemotes := make(chan *OD)
-	go Schedule(forceCtx, inRemotes)
+	go Schedule(c, inRemotes)
 
-	ticker := time.NewTicker(3 * time.Second)
+	for _, remote := range remotes {
-	defer ticker.Stop()
+		globalWait.Add(1)
-
+		inRemotes <- remote
-	task := Task {
-		WebsiteId: 0,
-		Url: u.String(),
 	}
-	ScheduleTask(inRemotes, &task, &u)
 
 	// Wait for all jobs to finish
 	globalWait.Wait()
 
+	logrus.Info("All dirs processed!")
+
 	return nil
 }
+
+var buildDate = time.Date(
+	2018, 10, 28,
+	17, 10, 0, 0,
+	time.UTC)

model.go

@@ -1,27 +1,12 @@
 package main
 
 import (
-	"github.com/terorie/od-database-crawler/ds/redblackhash"
+	"github.com/terorie/oddb-go/ds/redblackhash"
-	"github.com/terorie/od-database-crawler/fasturl"
+	"github.com/terorie/oddb-go/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       fasturl.URL
@@ -31,25 +16,26 @@ type Job struct {
 }
 
 type OD struct {
-	Task    Task
-	Result  TaskResult
 	Wait    sync.WaitGroup
 	BaseUri fasturl.URL
+	Files   []File
 	WCtx    WorkerContext
 	Scanned redblackhash.Tree
+
+	lock    sync.Mutex
 }
 
 type File struct {
-	Name  string `json:"name"`
+	Name  []rune    `json:"name"`
 	Size  int64     `json:"size"`
-	MTime int64  `json:"mtime"`
+	MTime time.Time `json:"mtime"`
-	Path  string `json:"path"`
+	Path  []rune    `json:"path"`
 	IsDir bool      `json:"-"`
 }
 
 func (o *OD) LoadOrStoreKey(k *redblackhash.Key) (exists bool) {
-	o.Scanned.Lock()
+	o.lock.Lock()
-	defer o.Scanned.Unlock()
+	defer o.lock.Unlock()
 
 	exists = o.Scanned.Get(k)
 	if exists { return true }

release.sh

@@ -1,25 +0,0 @@
-#!/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

runes/runes.go

@@ -0,0 +1,98 @@
+package runes
+
+func Create(rs ...[]rune) (x []rune) {
+	for _, r := range rs {
+		x = append(x, r...)
+	}
+	return x
+}
+
+func IndexRune(s []rune, r rune) int {
+	for i, sr := range s {
+		if r == sr {
+			return i
+		}
+	}
+	return -1
+}
+
+func LastIndexRune(s []rune, r rune) int {
+	for i := len(s)-1; i >= 0; i-- {
+		sr := s[i]
+		if r == sr {
+			return i
+		}
+	}
+	return -1
+}
+
+func Equals(a, b []rune) bool {
+	if len(a) != len(b) {
+		return false
+	}
+	for i := 0; i < len(a); i++ {
+		if a[i] != b[i] {
+			return false
+		}
+	}
+	return true
+}
+
+func Count(s []rune, r rune) (n int) {
+	for _, sr := range s {
+		if sr == r {
+			n++
+		}
+	}
+	return
+}
+
+func HasPrefix(s, prefix []rune) bool {
+	return len(s) >= len(prefix) && Equals(s[0:len(prefix)], prefix)
+}
+
+func HasSuffix(s, suffix []rune) bool {
+	return len(s) >= len(suffix) && Equals(s[len(s)-len(suffix):], suffix)
+}
+
+// TrimPrefix returns s without the provided leading prefix string.
+// If s doesn't start with prefix, s is returned unchanged.
+func TrimPrefix(s, prefix []rune) []rune {
+	if HasPrefix(s, prefix) {
+		return s[len(prefix):]
+	}
+	return s
+}
+
+func TrimRune(s []rune, r rune) (out []rune) {
+	if len(s) == 0 {
+		return nil
+	}
+
+	var i = 0
+	var sr rune
+	// Trim prefix
+	for i, sr = range s {
+		if sr == r {
+			out = s[i:]
+		} else {
+			break
+		}
+	}
+	s = out
+
+	if len(s) == 0 {
+		return nil
+	}
+
+	// Trim suffix
+	for i := len(s)-1; i >= 0; i++ {
+		if s[i] == r {
+			out = s[:i]
+		} else {
+			break
+		}
+	}
+
+	return out
+}

runespath/path.go

@@ -0,0 +1,44 @@
+// 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 path implements utility routines for manipulating slash-separated
+// paths.
+//
+// The path package should only be used for paths separated by forward
+// slashes, such as the paths in URLs. This package does not deal with
+// Windows paths with drive letters or backslashes; to manipulate
+// operating system paths, use the path/filepath package.
+package runespath
+
+import (
+	"github.com/terorie/oddb-go/runes"
+)
+
+// Base returns the last element of path.
+// Trailing slashes are removed before extracting the last element.
+// If the path is empty, Base returns ".".
+// If the path consists entirely of slashes, Base returns "/".
+func Base(path []rune) []rune {
+	if len(path) == 0 {
+		return []rune(".")
+	}
+	// Strip trailing slashes.
+	for len(path) > 0 && path[len(path)-1] == '/' {
+		path = path[0 : len(path)-1]
+	}
+	// Find the last element
+	if i := runes.LastIndexRune(path, '/'); i >= 0 {
+		path = path[i+1:]
+	}
+	// If empty now, it had only slashes.
+	if len(path) == 0 {
+		return []rune("/")
+	}
+	return path
+}
+
+// IsAbs reports whether the path is absolute.
+func IsAbs(path string) bool {
+	return len(path) > 0 && path[0] == '/'
+}

scheduler.go

@@ -2,201 +2,55 @@ 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 activeTasksLock sync.Mutex
+var activeTasks int32
-var activeTasks = make(map[uint64]bool)
-var numActiveTasks int32
 var totalBuffered int64
 
 func Schedule(c context.Context, remotes <-chan *OD) {
 	go Stats(c)
 
-	for remote := range remotes {
+	for {
+		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(results)
+				go remote.WCtx.Worker()
 			}
 
 			// Enqueue initial job
-		atomic.AddInt32(&numActiveTasks, 1)
+			atomic.AddInt32(&activeTasks, 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(results)
+			go remote.Watch()
-
-		// 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 ScheduleTask(remotes chan<- *OD, t *Task, u *fasturl.URL) {
+func (r *OD) Watch() {
-	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
-	o.Wait.Wait()
+	r.Wait.Wait()
-	close(o.WCtx.in)
+	close(r.WCtx.in)
-	atomic.AddInt32(&numActiveTasks, -1)
+	atomic.AddInt32(&activeTasks, -1)
 
-	// Log finish
+	logrus.WithField("url", r.BaseUri.String()).
-
+		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

@@ -10,28 +10,34 @@ import (
 	"net/http"
 	"net/url"
 	"os"
+	"path/filepath"
 	"strconv"
-	"time"
+	"strings"
 )
 
-var serverClient = http.Client {
+const (
-	Timeout: config.ServerTimeout,
+	fileListChunkSize int64 = 5000000 // 5 mb
-}
+)
+
+var serverClient = http.DefaultClient
 
 func FetchTask() (t *Task, err error) {
-	res, err := serverClient.PostForm(
+	escToken, _ := json.Marshal(config.Token)
+	payload := `{"token":` + string(escToken) + `}`
+
+	req, err := http.NewRequest(
+		http.MethodPost,
 		config.ServerUrl + "/task/get",
-		url.Values{ "token": {config.Token} })
+		strings.NewReader(payload))
+	if err != nil { return }
+
+	res, err := serverClient.Do(req)
 	if err != nil { return }
 	defer res.Body.Close()
 
-	switch res.StatusCode {
+	if res.StatusCode != 200 {
-	case 200:
+		err = fmt.Errorf("http %s", res.Status)
-		break
+		return
-	case 404, 500:
-		return nil, nil
-	default:
-		return nil, fmt.Errorf("http %s", res.Status)
 	}
 
 	t = new(Task)
@@ -41,17 +47,21 @@ func FetchTask() (t *Task, err error) {
 	return
 }
 
-func PushResult(result *TaskResult, f *os.File) (err error) {
+func PushResult(result *TaskResult) (err error) {
-	if result.WebsiteId == 0 {
+	filePath := filepath.Join(
-		// Not a real result, don't push
+		".", "crawled",
-		return nil
+		fmt.Sprintf("%d.json", result.WebsiteId))
-	}
 
-	// Rewind to the beginning of the file
+	defer os.Remove(filePath)
-	_, err = f.Seek(0, 0)
+
-	if err != nil {
+	f, err := os.Open(filePath)
+	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 {
@@ -63,110 +73,104 @@ func PushResult(result *TaskResult, f *os.File) (err error) {
 		return
 	}
 
-	// Upload result ignoring errors
+	err = uploadResult(result)
-	uploadResult(result)
+	if err != nil {
+		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) error {
+func uploadChunks(websiteId uint64, f *os.File) (err error) {
-	eof := false
+	for iter := 1; iter > 0; iter++ {
-	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 err }
+		if err != nil { return }
 		err = multi.WriteField("website_id", fmt.Sprintf("%d", websiteId))
-		if err != nil { return err }
+		if err != nil { return }
 
 		// Copy chunk to file_list
 		formFile, err := multi.CreateFormFile("file_list", "file_list")
-		var n int64
+		_, err = io.CopyN(formFile, f, fileListChunkSize)
-		n, err = io.CopyN(formFile, f, config.ChunkSize)
+		if err == io.EOF {
-		if err != io.EOF && err != nil {
+			break
-			return err
+		} else if err == io.ErrUnexpectedEOF {
-		}
-		if n == 0 {
-			// Don't upload, no content
-			return nil
-		} else if n < config.ChunkSize {
 			err = nil
 			// Break at end of iteration
-			eof = true
+			iter = -420
-		}
-
-		multi.Close()
-
-		for retried := false; true; retried = true {
-			err = nil
-			if retried {
-				// Error occurred, retry upload
-				time.Sleep(5 * time.Second)
 		}
 
 		req, err := http.NewRequest(
 			http.MethodPost,
 			config.ServerUrl + "/task/upload",
 			&b)
-			req.Header.Set("content-type", multi.FormDataContentType())
+		if err != nil { return err }
-			if err != nil { continue }
 
 		res, err := serverClient.Do(req)
-			if err != nil { continue }
+		if err != nil { return err }
 		res.Body.Close()
 
 		if res.StatusCode != http.StatusOK {
-				logrus.WithField("status", res.Status).
+			return fmt.Errorf("failed to upload list part %d: %s",
-					WithField("part", iter).
+				iter, res.Status)
-					Errorf("Upload failed")
-				continue
 		}
 
-			// Upload successful
+		logrus.Infof("Uploading file list part %d: %s",
-			break
+			iter, res.Status)
 	}
-
+	return
-		logrus.WithField("id", websiteId).
-			WithField("part", iter).
-			Infof("Uploaded files chunk")
-	}
-	return nil
 }
 
 func uploadResult(result *TaskResult) (err error) {
 	resultEnc, err := json.Marshal(result)
 	if err != nil { panic(err) }
 
-	res, err := serverClient.PostForm(
+	payload := url.Values {
-		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.Body.Close()
 
-	if res.StatusCode != http.StatusOK {
+	res, err := serverClient.Do(req)
-		return HttpError{res.StatusCode}
+	if err != nil { return }
-	}
+	res.Body.Close()
-
+
-	return
+	if res.StatusCode != http.StatusOK {
-}
+		return fmt.Errorf("failed to cancel task: %s", res.Status)
-
+	}
-func CancelTask(websiteId uint64) (err error) {
+
-	res, err := serverClient.PostForm(
+	return
-		config.ServerUrl + "/task/cancel",
+}
-		url.Values{
+
-			"token": {config.Token},
+func CancelTask(websiteId uint64) (err error) {
-			"website_id": {strconv.FormatUint(websiteId, 10)},
+	form := 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

@@ -39,10 +39,6 @@ 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),
@@ -57,7 +53,7 @@ func Stats(c context.Context) {
 			runtime.ReadMemStats(&mem)
 
 			logrus.WithFields(logrus.Fields{
-				"queue_count": atomic.LoadInt64(&totalBuffered),
+				"queue_count": totalBuffered,
 				"heap": FormatByteCount(mem.Alloc),
 				"objects": mem.HeapObjects,
 				"num_gc": mem.NumGC,

tasks.go

@@ -0,0 +1,16 @@
+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

@@ -1,9 +1,6 @@
 package main
 
-import (
+import "fmt"
-	"fmt"
-	"sync"
-)
 
 // https://programming.guide/go/formatting-byte-size-to-human-readable-format.html
 func FormatByteCount(b uint64) string {
@@ -19,20 +16,3 @@ 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

@@ -2,10 +2,7 @@ package main
 
 import (
 	"github.com/sirupsen/logrus"
-	"github.com/valyala/fasthttp"
 	"math"
-	"sort"
-	"strings"
 	"sync"
 	"sync/atomic"
 	"time"
@@ -20,13 +17,13 @@ type WorkerContext struct {
 	numRateLimits int
 }
 
-func (w WorkerContext) Worker(results chan<- File) {
+func (w WorkerContext) Worker() {
 	for job := range w.out {
-		w.step(results, job)
+		w.step(job)
 	}
 }
 
-func (w WorkerContext) step(results chan<- File, job Job) {
+func (w WorkerContext) step(job Job) {
 	defer w.finishJob(&job)
 
 	var f File
@@ -40,15 +37,10 @@ func (w WorkerContext) step(results chan<- File, job Job) {
 	if err != nil {
 		job.Fails++
 
-		if httpErr, ok := err.(*HttpError); ok {
+		if err == ErrForbidden {
-			switch httpErr.code {
+			// Don't attempt crawling again
-			case fasthttp.StatusTooManyRequests:
-				err = ErrRateLimit
-			default:
-				// Don't retry HTTP error codes
 			return
 		}
-		}
 
 		if job.Fails > config.Retries {
 			atomic.AddUint64(&totalAborted, 1)
@@ -70,9 +62,7 @@ func (w WorkerContext) step(results chan<- File, job Job) {
 		w.queueJob(job)
 	}
 
-	if !f.IsDir {
+	job.OD.Files = append(job.OD.Files, f)
-		results <- f
-	}
 }
 
 func DoJob(job *Job, f *File) (newJobs []Job, err error) {
@@ -81,11 +71,9 @@ func DoJob(job *Job, f *File) (newJobs []Job, err error) {
 		// Load directory
 		links, err := GetDir(job, f)
 		if err != nil {
-			if !isErrSilent(err) {
 			logrus.WithError(err).
 				WithField("url", job.UriStr).
-					Error("Failed to crawl dir")
+				Error("Failed getting dir")
-			}
 			return nil, err
 		}
 
@@ -97,55 +85,41 @@ func DoJob(job *Job, f *File) (newJobs []Job, err error) {
 			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 {
-			uriStr := link.String()
+			// Skip already queued links
-
+			//if _, old := job.OD.Scanned.LoadOrStore(link, true); old {
-			// Ignore dupes
+			//	continue
-			if uriStr == lastLink {
+			//}
-				continue
+			job.OD.Wait.Add(1)
-			}
-			lastLink = uriStr
-
 			newJobs = append(newJobs, Job{
 				OD:     job.OD,
 				Uri:    link,
-				UriStr: uriStr,
+				UriStr: link.String(),
 				Fails:  0,
 			})
-
-			newJobCount++
 		}
 		if config.Verbose {
 			logrus.WithFields(logrus.Fields{
 				"url":   job.UriStr,
-				"files": newJobCount,
+				"files": len(links),
 			}).Debug("Listed")
 		}
 	} else {
 		// Load file
 		err := GetFile(job.Uri, f)
 		if err != nil {
-			if !isErrSilent(err) {
 			logrus.WithError(err).
-					WithField("url", job.UriStr).
+				WithField("url", job.Uri.String()).
-					Error("Failed to crawl file")
+				Error("Failed getting 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 {
@@ -153,21 +127,14 @@ 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
+	}
+}
 
 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
 }