From 1fcf0bf1da9090d61f12242fa7521b13b7d41dab Mon Sep 17 00:00:00 2001
From: Florian Schmitt <mrflos@gmail.com>
Date: Thu, 27 Mar 2025 14:50:01 +0300
Subject: [PATCH] feat(updater): optional httpCheck + IA optimisations

---
 utils.inc.php       | 106 +++++++++++++++++++++++++++++++++-----------
 yeswiki-updater.php |   9 +++-
 2 files changed, 87 insertions(+), 28 deletions(-)

diff --git a/utils.inc.php b/utils.inc.php
index 3daf0ad..126c1cf 100644
--- a/utils.inc.php
+++ b/utils.inc.php
@@ -2,7 +2,6 @@
 
 require 'vendor/autoload.php';
 
-use Amp\Future;
 use Amp\Http\Client\HttpClientBuilder;
 use Amp\Http\Client\Request;
 
@@ -484,41 +483,94 @@ function addStatistics()
 
 }
 
-function searchWikis($path, $pattern, $depth = 1)
+function searchWikis($path, $pattern, $depth = 1, $httpCheck = false)
 {
-    $it = new RecursiveDirectoryIterator($path);
-    $list = array();
-    $httpClient = HttpClientBuilder::buildDefault();
-    $nb = 0;
-    $files = new RecursiveIteratorIterator($it);
-    $files->setMaxDepth($depth);
-    foreach ($files as $file) {
-        if (preg_match('/' . preg_quote($pattern) . '$/i', $file)) {
-            $nb++;
-            $wakkaConfig = [];
-            include_once($file);
-            $list[$nb] = [
-                'PATH' => dirname($file),
+    // Use a more efficient iterator that filters files by name
+    $dirIterator = new RecursiveDirectoryIterator($path, RecursiveDirectoryIterator::SKIP_DOTS);
+    $iterator = new RecursiveIteratorIterator($dirIterator);
+    $iterator->setMaxDepth($depth);
+
+    $list = [];
+    $count = 0;
+
+    // Process filesystem operations first
+    foreach ($iterator as $file) {
+        // Direct basename comparison instead of regex
+        if ($file->isFile() && basename($file) === $pattern) {
+            $count++;
+            $filePath = $file->getPathname();
+
+            // Extract config without using include_once which is slow
+            $wakkaConfig = extractWakkaConfig($filePath);
+
+            $list[$count] = [
+                'PATH' => dirname($filePath),
                 'URL' => $wakkaConfig['base_url'] ?? 'KO',
                 'VERSION' => $wakkaConfig['yeswiki_version'] ?? 'KO',
                 'RELEASE' => $wakkaConfig['yeswiki_release'] ?? 'KO',
             ];
         }
     }
-    try {
-        $responses = Future\awaitAll(array_map(function ($l) use ($httpClient) {
-            return Amp\async(fn () => $httpClient->request(new Request($l['URL'], 'HEAD')));
-        }, $list));
-        foreach ($responses[0] as $key => $response) {
-            $list[$key]['STATUS'] = 'ERROR';
+
+    // Process HTTP requests in smaller batches to avoid overwhelming resources
+    if (!empty($list) && $httpCheck) {
+        try {
+            // Process in batches of 10 for better performance
+            $batchSize = 10;
+            $batches = array_chunk($list, $batchSize, true);
+
+            foreach ($batches as $batch) {
+                $httpClient = HttpClientBuilder::buildDefault();
+                $futures = [];
+
+                foreach ($batch as $key => $item) {
+                    if ($item['URL'] !== 'KO') {
+                        $futures[$key] = Amp\async(fn () => $httpClient->request(new Request($item['URL'], 'HEAD')));
+                    }
+                }
+
+                $responses = Amp\Future\awaitAll($futures);
+
+                // Update statuses from responses
+                foreach ($responses[0] as $key => $error) {
+                    $list[$key]['STATUS'] = 'ERROR';
+                }
+
+                foreach ($responses[1] as $key => $response) {
+                    $list[$key]['STATUS'] = $response->getStatus() . ' ' . $response->getReason();
+                }
+            }
+        } catch (Exception $e) {
+            echo $e->getMessage(), "\n";
         }
-        foreach ($responses[1] as $key => $response) {
-            $list[$key]['STATUS'] = $response->getStatus() . ' ' . $response->getReason();
-        }
-    } catch (Exception $e) {
-        // If any one of the requests fails the combo will fail
-        echo $e->getMessage(), "\n";
     }
 
     return $list;
 }
+
+/**
+ * Extract configuration from a wakka.config.php file without using include
+ * which is much faster for many files
+ */
+function extractWakkaConfig($filePath)
+{
+    $config = [];
+    $content = file_get_contents($filePath);
+
+    // Extract base_url
+    if (preg_match('/[\'"]base_url[\'"]\s*=>\s*[\'"]([^\'"]+)[\'"]/i', $content, $matches)) {
+        $config['base_url'] = $matches[1];
+    }
+
+    // Extract yeswiki_version
+    if (preg_match('/[\'"]yeswiki_version[\'"]\s*=>\s*[\'"]([^\'"]+)[\'"]/i', $content, $matches)) {
+        $config['yeswiki_version'] = $matches[1];
+    }
+
+    // Extract yeswiki_release
+    if (preg_match('/[\'"]yeswiki_release[\'"]\s*=>\s*[\'"]([^\'"]+)[\'"]/i', $content, $matches)) {
+        $config['yeswiki_release'] = $matches[1];
+    }
+
+    return $config;
+}
diff --git a/yeswiki-updater.php b/yeswiki-updater.php
index b9df9d7..4c75242 100755
--- a/yeswiki-updater.php
+++ b/yeswiki-updater.php
@@ -51,6 +51,12 @@ if ($isRoot) {
         'description'  => 'Depth to scan folders for wikis',
         'defaultValue' => 1
       ],
+      'httpcheck' => [
+        'prefix'       => 'http',
+        'longPrefix'   => 'httpcheck',
+        'description'  => 'Perform an http request to check if the yeswikis are working',
+        'noValue' => true
+      ],
     ]);
     try {
         $climate->arguments->parse();
@@ -68,7 +74,8 @@ if ($isRoot) {
             $output = $climate->arguments->get('output');
             $nobackup = $climate->arguments->get('nobackup');
             $depth = $climate->arguments->get('depth');
-            $matches = searchWikis($path, 'wakka.config.php', $depth);
+            $httpCheck = $climate->arguments->get('httpcheck');
+            $matches = searchWikis($path, 'wakka.config.php', $depth, $httpCheck);
             if (count($matches) == 0) {
                 $climate->info('No yeswiki found on path ' . $path . ' with depth ' . $depth);
                 exit;