Merge pull request #10 from chillu/pulls/truncate-db-cache

Truncate FileContentCache by default to avoid SQL query errors

README.md

@@ -64,6 +64,10 @@ File:
 ```
 
 By default any extracted content will be cached against the database row.
+In order to stay within common size constraints for SQL queries required in this operation,
+the cache sets a maximum character length after which content gets truncated (default: 500000).
+You can configure this value through `FileTextCache_Database.max_content_length` in your yaml configuration.
+
 
 Alternatively, extracted content can be cached using SS_Cache to prevent excessive database growth.
 In order to swap out the cache backend you can use the following yaml configuration.

_config/config.yml

@@ -6,3 +6,6 @@ Injector:
 
 SolrCellTextExtractor:
 #  base_url: 'http://localhost:8983/solr/update/extract'
+
+FileTextCache_Database:
+  max_content_length: 500000

code/extensions/FileTextCache.php

@@ -19,7 +19,8 @@ interface FileTextCache {
 }
 
 /**
- * Caches the extracted content on the record for the file
+ * Caches the extracted content on the record for the file.
+ * Limits the stored file content by default to avoid hitting query size limits.
  */
 class FileTextCache_Database implements FileTextCache {
 	
@@ -28,7 +29,8 @@ class FileTextCache_Database implements FileTextCache {
 	}
 
 	public function save(File $file, $content) {
-		$file->FileContentCache = $content;
+		$maxLength = Config::inst()->get('FileTextCache_Database', 'max_content_length');
+		$file->FileContentCache = ($maxLength) ? substr($content, 0, $maxLength) : $content;
 		$file->write();
 	}
 

tests/FileTextCacheDatabaseTest.php

@@ -0,0 +1,17 @@
+<?php
+class FileTextCacheDatabaseTest extends SapphireTest {
+	
+	public function testTruncatesByMaxLength() {
+		Config::nest();
+		
+		Config::inst()->update('FileTextCache_Database', 'max_content_length', 5);
+		$cache = new FileTextCache_Database();
+		$file = $this->getMock('File', array('write'));
+		$content = '0123456789';
+		$cache->save($file, $content);
+		$this->assertEquals($cache->load($file), '01234');
+
+		Config::unnest();
+	}
+
+}