support ETag and Last-Modified headers

habr_twister.php

@@ -4,9 +4,10 @@ mb_internal_encoding('UTF-8');
 setlocale(LC_CTYPE, 'en_US.UTF-8');
 
 // Load RSS feed
+require_once 'rssreader.php';
 $feed_uri = 'http://habrahabr.ru/rss/hubs/new/';
-$rss = simplexml_load_file($feed_uri);
+$rss = getRssFeed($feed_uri);
-if (!$rss) die('Cannot read rss');
+if (!$rss) die('Cannot read rss or it is up to date');
 
 // Initialise TwisterPost
 require_once 'twisterpost.php';

rssreader.php

@@ -0,0 +1,53 @@
+<?php
+
+function getRssFeed($feed_uri, $reload = false)
+{
+	$meta_file = md5($feed_uri) . '.json';
+	$meta = @json_decode(@file_get_contents($meta_file), true);
+
+	$ch = curl_init();
+	curl_setopt($ch, CURLOPT_URL, $feed_uri);
+	curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, 20);
+	curl_setopt($ch, CURLOPT_HEADER, true);
+	curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
+	curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
+	if (!$reload) {
+		if (isset($meta['etag']))
+			curl_setopt($ch, CURLOPT_HTTPHEADER, array('If-None-Match: ' . $meta['etag']));
+		if (isset($meta['lastModified']))
+			curl_setopt($ch, CURLOPT_HTTPHEADER, array('If-Modified-Since: ' . $meta['lastModified']));
+	}
+
+	$response = curl_exec($ch);
+	$info = curl_getinfo($ch);
+
+	$http_code = $info['http_code'];
+	if ($http_code !== 200)
+		return false;
+
+	$headers = substr($response, 0, $info['header_size']);
+	$headers_arr = array();
+	foreach (explode("\n", $headers) as $h) {
+		$h = explode(':', $h, 2);
+		if (!isset($h[1])) continue;
+		$key = $h[0];
+		$value = trim($h[1]);
+		if (!isset($headers_arr[$key]))
+			$headers_arr[$key] = $value;
+		elseif (is_array($headers[$key]))
+			$headers_arr[$key][] = $value;
+		else
+			$headers_arr[$key] = array($headers_arr[$key], $value);
+	}
+
+	$meta = array();
+	if (isset($headers_arr['Etag']))
+		$meta['etag'] = $headers_arr['Etag'];
+	if (isset($headers_arr['Last-Modified']))
+		$meta['lastModified'] = $headers_arr['Last-Modified'];
+	file_put_contents($meta_file, json_encode($meta));
+
+	$body = substr($response, $info['header_size']);
+
+	return @simplexml_load_string($body);
+}