Added possibility to limit the amount of data retrieved

README.md

@@ -72,6 +72,9 @@ Httpful highly encourages sending in pull requests.  When submitting a pull requ
 
 # Changelog
 
+## 0.2.18
+- FEATURE [Allow to limit the amount of data retrieved](https://github.com/nategood/httpful/issues/137)
+
 ## 0.2.17
 
  - FEATURE [PR #144](https://github.com/nategood/httpful/pull/144) Adds additional parameter to the Response class to specify additional meta data about the request/response (e.g. number of redirect).
@@ -201,4 +204,3 @@ Httpful highly encourages sending in pull requests.  When submitting a pull requ
   - Created AbstractMimeHandler type that all Mime Handlers must extend
   - Pulled out the parsing/serializing logic from the Request/Response classes into their own MimeHandler classes
   - Added ability to register new mime handlers for mime types
-

composer.json

@@ -4,12 +4,17 @@
     "homepage": "http://github.com/nategood/httpful",
     "license": "MIT",
     "keywords": ["http", "curl", "rest", "restful", "api", "requests"],
-    "version": "0.2.17",
+    "version": "0.2.18",
     "authors": [
         {
             "name": "Nate Good",
             "email": "[email protected]",
             "homepage": "http://nategood.com"
+        },
+        {
+            "name": "Diego Zanella",
+            "email": "[email protected]",
+            "homepage": "http://aelia.co"
         }
     ],
     "require": {

src/Httpful/Request.php

@@ -61,6 +61,11 @@ class Request
     // Template Request object
     private static $_template;
 
+    // @var int The maximum amount of data to retrieve.
+    protected $download_limit;
+    // @var string The data retrieved by the CURL request. Used only a download limit is set.
+    protected $retrieved_data;
+
     /**
      * We made the constructor private to force the factory style.  This was
      * done to keep the syntax cleaner and better the support the idea of
@@ -201,14 +206,43 @@ public function send()
         $result = curl_exec($this->_ch);
 
         if ($result === false) {
-            if ($curlErrorNumber = curl_errno($this->_ch)) {
-                $curlErrorString = curl_error($this->_ch);
-                $this->_error($curlErrorString);
-                throw new ConnectionErrorException('Unable to connect: ' . $curlErrorNumber . ' ' . $curlErrorString);
+            $curlErrorNumber = curl_errno($this->_ch);
+
+            // If error number is CURLE_WRITE_ERROR, it may just be that we hit
+            // the download limit. In such case, we can get the data retrieved so
+            // far and carry on. No need to throw an exception, as we are not
+            // dealing with an actual error.
+            if(($curlErrorNumber == CURLE_WRITE_ERROR) &&
+               ($this->download_limit > 0)) {
+              $result = true;
+            }
+            else {
+              // Any other error number represents an actual error
+              if ($curlErrorNumber) {
+                  $curlErrorString = curl_error($this->_ch);
+                  $this->_error($curlErrorString);
+                  throw new ConnectionErrorException('Unable to connect: ' . $curlErrorNumber . ' ' . $curlErrorString);
+              }
+
+              $this->_error('Unable to connect.');
+              throw new ConnectionErrorException('Unable to connect.');
             }
+        }
 
-            $this->_error('Unable to connect.');
-            throw new ConnectionErrorException('Unable to connect.');
+        /* Result can be "true" in two cases:
+         * - When download limit is greater than zero, and the limit set was
+         *   larger than the page size (i.e. the whole page was fetched, despite
+         *   the limit).
+         * - When download limit is greater than zero and error CURLE_WRITE_ERROR
+         *   was raised (i.e. the transfer was interrupted because the limit was
+         *   reached).
+         *
+         * In both cases, the data is actually stored in $this->retrieved_data,
+         * therefore it must be put back inside $result, where the library
+         * expects to find it.
+         */
+        if($result === true) {
+          $result = $this->retrieved_data;
         }
 
         $info = curl_getinfo($this->_ch);
@@ -1135,4 +1169,51 @@ public static function options($uri)
     {
         return self::init(Http::OPTIONS)->uri($uri);
     }
+
+    /**
+     * Allows to limit the size of retrieved data. Useful when you only need to
+     * get the headers of a page, as remote servers usually don't honour the
+     * "range" header in HTTP requests.
+     *
+     * IMPORTANT: setting the limit too low will cause the request to fail, because
+     * the response will not contain the headers and body expected by the parser.
+     * The minimum value should be at least 1000, to ensure that all headers are
+     * retrieved and the parsing can succeed.
+     *
+     * @param int size The amount of data to retrieve, in bytes.
+     * @return Httpful\Request
+     */
+    public function limit($size) {
+      if((int)$size > 0) {
+        $this->download_limit = $size;
+        $this->retrieved_data = '';
+        $this->addOnCurlOption(CURLOPT_BINARYTRANSFER, 1);
+        $this->addOnCurlOption(CURLOPT_WRITEFUNCTION, array($this, 'downloadLimiter'));
+      }
+
+      return $this;
+    }
+
+    /**
+     * Callback for the Request::limit() method. This method keeps track of the
+     * data retrieved, and interrupts the transfer once the limit has been
+     * reached.
+     *
+     * @param object ch The CURL handle.
+     * @param string chunk A chunk of the data retrieved by the CURL request.
+     * @return int The length of the retrieved chunk, or -1 if the limit has
+     * been reached.
+     */
+    public function downloadLimiter($ch, $chunk) {
+      $len = strlen($this->retrieved_data) + strlen($chunk);
+      if($len >= $this->download_limit) {
+        $this->retrieved_data .= substr($chunk, 0, $this->download_limit - strlen($this->retrieved_data));
+        //echo strlen($this->retrieved_data) , ' ', $this->retrieved_data;
+        return -1;
+      }
+
+      $this->retrieved_data .= $chunk;
+
+      return strlen($chunk);
+    }
 }