Unescape results & don't recrawl 404

crawl.go

@@ -2,7 +2,6 @@ package main
 
 import (
 	"bytes"
-	"fmt"
 	"github.com/terorie/oddb-go/ds/redblackhash"
 	"github.com/terorie/oddb-go/fasturl"
 	"github.com/valyala/fasthttp"
@@ -190,15 +189,7 @@ 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}
 	}
 }

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

@@ -811,3 +811,57 @@ 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
+}

scheduler.go

@@ -5,6 +5,7 @@ import (
 	"encoding/json"
 	"fmt"
 	"github.com/sirupsen/logrus"
+	"github.com/terorie/oddb-go/fasturl"
 	"os"
 	"path"
 	"sync/atomic"
@@ -91,6 +92,8 @@ func (t *Task) collect(results chan File) error {
 	defer f.Close()
 
 	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)

worker.go

@@ -2,6 +2,7 @@ package main
 
 import (
 	"github.com/sirupsen/logrus"
+	"github.com/valyala/fasthttp"
 	"math"
 	"sort"
 	"strings"
@@ -39,9 +40,16 @@ func (w WorkerContext) step(results chan<- File, 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.StatusUnauthorized,
+				fasthttp.StatusForbidden,
+				fasthttp.StatusNotFound:
 				return
+			case fasthttp.StatusTooManyRequests:
+				err = ErrRateLimit
+			}
 		}
 
 		if job.Fails > config.Retries {