Handle errors and warnings from loadHTML()

[tweeper.git] / tweeper.php

diff --git a/tweeper.php b/tweeper.php
index ef55022..e1483f1 100644 (file)
--- a/tweeper.php
+++ b/tweeper.php
@@ -47,20 +47,21 @@ class Tweeper {
     return Tweeper::epoch_to_gmdate($timestamp);
   }
 
-  private static function get_contents($uri) {
-    # https://www.wjsams.com/c/docs/Wiki/Php-HowToSetUserAgentOnFileGetContents
-    $opts = array(
-      'http' => array(
-        'method' => "GET",
-        'header' => join(array(
-          "Accept-language: en\r\n",
-          "User-Agent: {" + Tweeper::$USER_AGENT + "}\r\n"
-        ))
-      )
-    );
+  private static function get_contents($url)
+  {
+    $ch = curl_init($url);
+    curl_setopt_array($ch, array(
+      CURLOPT_HEADER => FALSE,
+      CURLOPT_FOLLOWLOCATION => TRUE,  // follow http redirects to get the real URL
+      CURLOPT_RETURNTRANSFER => TRUE,
+      CURLOPT_SSL_VERIFYHOST => FALSE,
+      CURLOPT_SSL_VERIFYPEER => FALSE,
+      CURLOPT_HTTPHEADER => array('Accept-language: en'),
+      CURLOPT_USERAGENT => Tweeper::$USER_AGENT,
+    ));
+    $contents = curl_exec($ch);
+    curl_close($ch);
 
-    $context = stream_context_create($opts);
-    $contents = file_get_contents($uri, false, $context);
     return $contents;
   }
 
@@ -68,16 +69,16 @@ class Tweeper {
   {
     $ch = curl_init($url);
     curl_setopt_array($ch, array(
-        CURLOPT_HEADER => TRUE,
-        CURLOPT_NOBODY => TRUE,
-        CURLOPT_FOLLOWLOCATION => TRUE,  // follow http redirects to get the real URL
-        CURLOPT_RETURNTRANSFER => TRUE,
-        CURLOPT_SSL_VERIFYHOST => FALSE,
-        CURLOPT_SSL_VERIFYPEER => FALSE,
-        CURLOPT_USERAGENT => Tweeper::$USER_AGENT,
+      CURLOPT_HEADER => TRUE,
+      CURLOPT_NOBODY => TRUE,
+      CURLOPT_FOLLOWLOCATION => TRUE,  // follow http redirects to get the real URL
+      CURLOPT_RETURNTRANSFER => TRUE,
+      CURLOPT_SSL_VERIFYHOST => FALSE,
+      CURLOPT_SSL_VERIFYPEER => FALSE,
+      CURLOPT_USERAGENT => Tweeper::$USER_AGENT,
     ));
     curl_exec($ch);
-    $url_info = curl_getinfo($ch);;
+    $url_info = curl_getinfo($ch);
     curl_close($ch);
 
     return $url_info;
@@ -121,11 +122,48 @@ class Tweeper {
     return $dom->saveXML($enc);
   }
 
+  /* Mimic the message from libxml.c::php_libxml_ctx_error_level() */
+  private function log_xml_error($error) {
+    $output = "";
+
+    switch ($error->level) {
+    case LIBXML_ERR_WARNING:
+      $output .= "Warning $error->code: ";
+      break;
+    case LIBXML_ERR_ERROR:
+      $output .= "Error $error->code: ";
+      break;
+    case LIBXML_ERR_FATAL:
+      $output .= "Fatal Error $error->code: ";
+      break;
+    }
+
+    $output .= trim($error->message);
+
+    if ($error->file) {
+      $output .= " in $error->file";
+    } else {
+      $output .= " in Entity,";
+    }
+
+    $output .=" line $error->line";
+
+    error_log($output);
+  }
+
   public function tweep($uri) {
     $html = Tweeper::get_contents($uri);
 
     $xmlDoc = new DOMDocument();
+
+    // Handle warnings and errors when loading invalid HTML.
+    $xml_errors_value = libxml_use_internal_errors(true);
     $xmlDoc->loadHTML($html);
+    foreach (libxml_get_errors() as $xml_error) {
+      $this->log_xml_error($xml_error);
+    }
+    libxml_clear_errors();
+    libxml_use_internal_errors($xml_errors_value);
 
     $output = $this->xsltProcessor->transformToXML($xmlDoc);
 
@@ -204,7 +242,7 @@ $url = parse_url($options['src_url']);
 if (FALSE === $url || empty($url["host"]))
   die("Invalid url: ${options['src_url']}\n");
 
-$stylesheet = __DIR__ . "/rss_converter_" . $url["host"] . ".xsl";
+$stylesheet = "file://" . __DIR__ . "/rss_converter_" . $url["host"] . ".xsl";
 if (FALSE === file_exists($stylesheet))
   die("Conversion to RSS not supported: {$url["host"]}\n");