Add context options to handle BOM

malarzm · malarzm · commit 35d8e61ad4f1 · 2019-10-09T21:40:15.000+02:00
diff --git a/src/Symfony/Component/Serializer/CHANGELOG.md b/src/Symfony/Component/Serializer/CHANGELOG.md
@@ -5,6 +5,7 @@ CHANGELOG
 -----
 
  * deprecated the `XmlEncoder::TYPE_CASE_ATTRIBUTES` constant, use `XmlEncoder::TYPE_CAST_ATTRIBUTES` instead
+ * added option to output an UTF-8 BOM in CSV encoder via `CsvEncoder::OUTPUT_UTF8_BOM_KEY` context option
 
 4.3.0
 -----
diff --git a/src/Symfony/Component/Serializer/Encoder/CsvEncoder.php b/src/Symfony/Component/Serializer/Encoder/CsvEncoder.php
@@ -12,6 +12,7 @@
 namespace Symfony\Component\Serializer\Encoder;
 
 use Symfony\Component\Serializer\Exception\InvalidArgumentException;
+use Symfony\Component\Serializer\Exception\UnexpectedValueException;
 
 /**
  * Encodes CSV data.
@@ -30,6 +31,9 @@ class CsvEncoder implements EncoderInterface, DecoderInterface
     const ESCAPE_FORMULAS_KEY = 'csv_escape_formulas';
     const AS_COLLECTION_KEY = 'as_collection';
     const NO_HEADERS_KEY = 'no_headers';
+    const OUTPUT_UTF8_BOM_KEY = 'output_utf8_bom';
+
+    private const UTF8_BOM = "\xEF\xBB\xBF";
 
     private $formulasStartCharacters = ['=', '-', '+', '@'];
     private $defaultContext = [
@@ -40,6 +44,7 @@ class CsvEncoder implements EncoderInterface, DecoderInterface
         self::HEADERS_KEY => [],
         self::KEY_SEPARATOR_KEY => '.',
         self::NO_HEADERS_KEY => false,
+        self::OUTPUT_UTF8_BOM_KEY => false,
     ];
 
     /**
@@ -90,7 +95,7 @@ public function encode($data, $format, array $context = [])
             }
         }
 
-        list($delimiter, $enclosure, $escapeChar, $keySeparator, $headers, $escapeFormulas) = $this->getCsvOptions($context);
+        list($delimiter, $enclosure, $escapeChar, $keySeparator, $headers, $escapeFormulas, $outputBom) = $this->getCsvOptions($context);
 
         foreach ($data as &$value) {
             $flattened = [];
@@ -114,6 +119,14 @@ public function encode($data, $format, array $context = [])
         $value = stream_get_contents($handle);
         fclose($handle);
 
+        if ($outputBom) {
+            if (!preg_match('//u', $value)) {
+                throw new UnexpectedValueException('You are trying to add an UTF-8 BOM to a non UTF-8 text.');
+            }
+
+            $value = self::UTF8_BOM.$value;
+        }
+
         return $value;
     }
 
@@ -134,6 +147,10 @@ public function decode($data, $format, array $context = [])
         fwrite($handle, $data);
         rewind($handle);
 
+        if (0 === strpos($data, self::UTF8_BOM)) {
+            fseek($handle, \strlen(self::UTF8_BOM));
+        }
+
         $headers = null;
         $nbHeaders = 0;
         $headerCount = [];
@@ -238,12 +255,13 @@ private function getCsvOptions(array $context): array
         $keySeparator = $context[self::KEY_SEPARATOR_KEY] ?? $this->defaultContext[self::KEY_SEPARATOR_KEY];
         $headers = $context[self::HEADERS_KEY] ?? $this->defaultContext[self::HEADERS_KEY];
         $escapeFormulas = $context[self::ESCAPE_FORMULAS_KEY] ?? $this->defaultContext[self::ESCAPE_FORMULAS_KEY];
+        $outputBom = $context[self::OUTPUT_UTF8_BOM_KEY] ?? $this->defaultContext[self::OUTPUT_UTF8_BOM_KEY];
 
         if (!\is_array($headers)) {
             throw new InvalidArgumentException(sprintf('The "%s" context variable must be an array or null, given "%s".', self::HEADERS_KEY, \gettype($headers)));
         }
 
-        return [$delimiter, $enclosure, $escapeChar, $keySeparator, $headers, $escapeFormulas];
+        return [$delimiter, $enclosure, $escapeChar, $keySeparator, $headers, $escapeFormulas, $outputBom];
     }
 
     /**
diff --git a/src/Symfony/Component/Serializer/Tests/Encoder/CsvEncoderTest.php b/src/Symfony/Component/Serializer/Tests/Encoder/CsvEncoderTest.php
@@ -13,6 +13,7 @@
 
 use PHPUnit\Framework\TestCase;
 use Symfony\Component\Serializer\Encoder\CsvEncoder;
+use Symfony\Component\Serializer\Exception\UnexpectedValueException;
 
 /**
  * @author Kévin Dunglas <dunglas@gmail.com>
@@ -595,4 +596,38 @@ public function testDecodeWithoutHeader()
             CsvEncoder::NO_HEADERS_KEY => true,
         ]));
     }
+
+    public function testBOMIsAddedOnDemand()
+    {
+        $value = ['foo' => 'hello', 'bar' => 'hey ho'];
+
+        $this->assertEquals("\xEF\xBB\xBF".<<<'CSV'
+foo,bar
+hello,"hey ho"
+
+CSV
+            , $this->encoder->encode($value, 'csv', [CsvEncoder::OUTPUT_UTF8_BOM_KEY => true]));
+    }
+
+    public function testBOMCanNotBeAddedToNonUtf8Csv()
+    {
+        $value = [mb_convert_encoding('ÄÖÜ', 'ISO-8859-1', 'UTF-8')];
+
+        $this->expectException(UnexpectedValueException::class);
+        $this->expectExceptionMessage('You are trying to add an UTF-8 BOM to a non UTF-8 text.');
+        $this->encoder->encode($value, 'csv', [CsvEncoder::OUTPUT_UTF8_BOM_KEY => true]);
+    }
+
+    public function testBOMIsStripped()
+    {
+        $csv = "\xEF\xBB\xBF".<<<'CSV'
+foo,bar
+hello,"hey ho"
+
+CSV;
+        $this->assertEquals(
+            ['foo' => 'hello', 'bar' => 'hey ho'],
+            $this->encoder->decode($csv, 'csv', [CsvEncoder::AS_COLLECTION_KEY => false])
+        );
+    }
 }
