From 671a7e8f2c28e5b851f50ac0dca4de69fe4e1aa1 Mon Sep 17 00:00:00 2001
From: Dominik Pfennig <dominik@suma-ev.de>
Date: Wed, 25 Apr 2018 13:43:43 +0200
Subject: [PATCH] Das Encoding wird nun automatisch entweder aus dem Response
Header, oder aus den Meta Tags der Seite ausgelesen. Fallback ist "UTF-8" das
sollte viele Umlautprobleme beheben.
---
app/HtmlDocument.php | 31 ++++++++++++++++++++++--
app/Http/Controllers/ProxyController.php | 7 ++++--
2 files changed, 34 insertions(+), 4 deletions(-)
diff --git a/app/HtmlDocument.php b/app/HtmlDocument.php
index 09dc186..c880572 100644
--- a/app/HtmlDocument.php
+++ b/app/HtmlDocument.php
@@ -9,10 +9,37 @@ class HtmlDocument extends Document
private $htmlString;
- public function __construct($password, $baseUrl, $htmlString, $encoding = "UTF-8")
+ public function __construct($password, $baseUrl, $htmlString, $encoding)
{
parent::__construct($password, $baseUrl);
- $this->htmlString = mb_convert_encoding($htmlString, 'HTML-ENTITIES', $encoding);
+ $this->htmlString = $this->convertEncoding($htmlString, $encoding);
+ }
+
+ public function convertEncoding($htmlString, $encoding){
+ # If the site sets the document encoding in the "content-type" Header the $encoding variable is not null
+ if($encoding == null){
+ # Otherwise we will try to extract the correct encoding from the meta tag
+ # Let's create a new DOM
+ libxml_use_internal_errors(true);
+ $dom = new DomDocument();
+ $dom->loadHtml($htmlString);
+
+ foreach($dom->getElementsByTagName("meta") as $meta){
+ # If there is a Content-Type Meta Tag
+ if($meta->hasAttribute("http-equiv") && strtolower($meta->getAttribute("http-equiv")) === "content-type" && $meta->hasAttribute("content")){
+ $contentType = $meta->getAttribute("content");
+ $encoding = stripos($contentType, "charset=") !== false ? trim(substr($contentType, stripos($contentType, "charset=")+8)) : null;
+ if($encoding !== null) break;
+ }
+ # If there is a Charset Meta Tag
+ if($meta->hasAttribute("charset")){
+ $encoding = $meta->getAttribute("charset");
+ break;
+ }
+ }
+ if($encoding === null) $encoding = "UTF-8"; # Default Fallback
+ }
+ return mb_convert_encoding($htmlString, 'HTML-ENTITIES', $encoding);
}
public function getResult()
diff --git a/app/Http/Controllers/ProxyController.php b/app/Http/Controllers/ProxyController.php
index 0803fd3..f3faf35 100644
--- a/app/Http/Controllers/ProxyController.php
+++ b/app/Http/Controllers/ProxyController.php
@@ -131,7 +131,7 @@ class ProxyController extends Controller
$supportedContentTypes = [
'text/html',
];
-
+
$targetUrl = str_replace("<<SLASH>>", "/", $url);
$targetUrl = str_rot13(base64_decode($targetUrl));
$this->password = $password;
@@ -159,11 +159,14 @@ class ProxyController extends Controller
# We will parse whether we have a parser for this document type.
# If not, we will not Proxy it:
+ $contentTypeHeader = $result["header"]["content-type"];
$contentType = strpos($result["header"]["content-type"], ";") !== false ? trim(substr($result["header"]["content-type"], 0, strpos($result["header"]["content-type"], ";"))) : trim($result["header"]["content-type"]);
+ $contentEncoding = stripos($contentTypeHeader, "charset=") !== false ? trim(substr($contentTypeHeader, stripos($contentTypeHeader, "charset=")+8)) : null;
switch ($contentType) {
case 'text/html':
# It's a html Document
- $htmlDocument = new HtmlDocument($password, $targetUrl, $result["data"]);
+ $contentEncoding = null;
+ $htmlDocument = new HtmlDocument($password, $targetUrl, $result["data"], $contentEncoding);
$htmlDocument->proxifyContent();
$result["data"] = $htmlDocument->getResult();
break;
--
GitLab