Changeset 44 in code for trunk/morty.go

Timestamp:

Nov 28, 2016, 9:59:14 AM (9 years ago)

Author:

alex

Message:

[fix] fix #14. Fetched URI matched the mortyurl.
Query parameters in requested URI were parsed and set again. Unfortunately a query ?a&b were changed into ?a=b= which can lead to 404 errors.

File:

• trunk/morty.go (modified) (13 diffs)

trunk/morty.go

@@ -18 +18 @@
         "github.com/valyala/fasthttp"
         "golang.org/x/net/html"
-        "golang.org/x/net/html/charset"
-        "golang.org/x/text/encoding"
+        "golang.org/x/text/encoding/charmap"
 )
 
@@ -124 +123 @@
 `
 
-var HTML_META_CONTENT_TYPE string = "<meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\">"
-
 func (p *Proxy) RequestHandler(ctx *fasthttp.RequestCtx) {
 
@@ -167 +164 @@
         req.SetConnectionClose()
 
-        reqQuery := parsedURI.Query()
-        ctx.QueryArgs().VisitAll(func(key, value []byte) {
-                reqQuery.Add(string(key), string(value))
-        })
-
-        parsedURI.RawQuery = reqQuery.Encode()
-
-        uriStr := parsedURI.String()
-
-        log.Println("getting", uriStr)
-
-        req.SetRequestURI(uriStr)
+        requestURIStr := string(requestURI)
+
+        log.Println("getting", requestURIStr)
+
+        req.SetRequestURI(requestURIStr)
         req.Header.SetUserAgentBytes([]byte("Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/53.0.2785.143 Safari/537.36"))
 
@@ -217 +207 @@
                         }
                 }
-                error_message := fmt.Sprintf("invalid response: %d", resp.StatusCode())
+                error_message := fmt.Sprintf("invalid response: %d (%s)", resp.StatusCode(), requestURIStr)
                 p.serveMainPage(ctx, resp.StatusCode(), errors.New(error_message))
                 return
@@ -240 +230 @@
         var responseBody []byte
 
-        if len(contentInfo) == 2 && bytes.Contains(contentInfo[0], []byte("text")) {
-                e, ename, _ := charset.DetermineEncoding(resp.Body(), string(contentType))
-                if (e != encoding.Nop) && (!strings.EqualFold("utf-8", ename)) {
-                        responseBody, err = e.NewDecoder().Bytes(resp.Body())
-                        if err != nil {
-                                // HTTP status code 503 : Service Unavailable
-                                p.serveMainPage(ctx, 503, err)
-                                return
-                        }
-                } else {
-                        responseBody = resp.Body()
+        if len(contentInfo) == 2 && bytes.Contains(contentInfo[1], []byte("ISO-8859-2")) && bytes.Contains(contentInfo[0], []byte("text")) {
+                var err error
+                responseBody, err = charmap.ISO8859_2.NewDecoder().Bytes(resp.Body())
+                if err != nil {
+                        // HTTP status code 503 : Service Unavailable
+                        p.serveMainPage(ctx, 503, err)
+                        return
                 }
         } else {
@@ -333 +319 @@
         unsafeElements := make([][]byte, 0, 8)
         state := STATE_DEFAULT
+
         for {
                 token := decoder.Next()
@@ -360 +347 @@
                                         for {
                                                 attrName, attrValue, moreAttr := decoder.TagAttr()
-                                                if bytes.Equal(attrName, []byte("href")) {
-                                                        parsedURI, err := url.Parse(string(attrValue))
-                                                        if err == nil {
-                                                                rc.BaseURL = parsedURI
-                                                        }
+                                                if !bytes.Equal(attrName, []byte("href")) {
+                                                        continue
+                                                }
+                                                parsedURI, err := url.Parse(string(attrValue))
+                                                if err == nil {
+                                                        rc.BaseURL = parsedURI
                                                 }
                                                 if !moreAttr {
@@ -395 +383 @@
                                 }
 
-                                if bytes.Equal(tag, []byte("meta")) {
-                                        sanitizeMetaTag(rc, out, attrs)
-                                        break
-                                }
-
                                 fmt.Fprintf(out, "<%s", tag)
 
                                 if hasAttrs {
-                                        sanitizeAttrs(rc, out, attrs)
+                                        if bytes.Equal(tag, []byte("meta")) {
+                                                sanitizeMetaAttrs(rc, out, attrs)
+                                        } else {
+                                                sanitizeAttrs(rc, out, attrs)
+                                        }
                                 }
 
@@ -413 +400 @@
                                                 state = STATE_IN_STYLE
                                         }
-                                }
-
-                                if bytes.Equal(tag, []byte("head")) {
-                                        fmt.Fprintf(out, HTML_META_CONTENT_TYPE)
                                 }
 
@@ -516 +499 @@
 }
 
-func sanitizeMetaTag(rc *RequestConfig, out io.Writer, attrs [][][]byte) {
+func sanitizeMetaAttrs(rc *RequestConfig, out io.Writer, attrs [][][]byte) {
         var http_equiv []byte
         var content []byte
@@ -529 +512 @@
                         content = attrValue
                 }
-                if bytes.Equal(attrName, []byte("charset")) {
-                        // exclude <meta charset="...">
-                        return
-                }
-        }
-
-        if bytes.Equal(http_equiv, []byte("content-type")) {
-                return
-        }
-
-        out.Write([]byte("<meta"))
+        }
+
         urlIndex := bytes.Index(bytes.ToLower(content), []byte("url="))
         if bytes.Equal(http_equiv, []byte("refresh")) && urlIndex != -1 {
@@ -556 +530 @@
                 sanitizeAttrs(rc, out, attrs)
         }
-        out.Write([]byte(">"))
+
 }
 
@@ -613 +587 @@
                 return fmt.Sprintf("./?mortyurl=%s", url.QueryEscape(uri)), nil
         }
+
         return fmt.Sprintf("./?mortyhash=%s&mortyurl=%s", hash(uri, rc.Key), url.QueryEscape(uri)), nil
 }