Backport PR #4189 Csv Method

PHPOffice · Oct 15, 2024 · 2a1ece2 · 2a1ece2
1 parent 4e58d8a
commit 2a1ece2
Show file tree

Hide file tree

Showing 8 changed files with 128 additions and 16 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -5,6 +5,12 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com)
 and this project adheres to [Semantic Versioning](https://semver.org).
 
+## TBD - 2.1.2
+
+### Added
+
+- Method to Test Whether Csv Will Be Affected by Php9 (backport of PR #4189 intended for 3.4.0)
+
 ## 2024-09-29 2.1.1
 
 ### Fixed

diff --git a/src/PhpSpreadsheet/Reader/Csv.php b/src/PhpSpreadsheet/Reader/Csv.php
@@ -10,6 +10,7 @@
 use PhpOffice\PhpSpreadsheet\Shared\StringHelper;
 use PhpOffice\PhpSpreadsheet\Spreadsheet;
 use PhpOffice\PhpSpreadsheet\Worksheet\Worksheet;
+use Throwable;
 
 class Csv extends BaseReader
 {
@@ -74,7 +75,7 @@ class Csv extends BaseReader
      * It is anticipated that it will conditionally be set
      * to null-string for Php9 and above.
      */
-    private static string $defaultEscapeCharacter = '\\';
+    private static string $defaultEscapeCharacter = PHP_VERSION_ID < 90000 ? '\\' : '';
 
     /**
      * Callback for setting defaults in construction.
@@ -286,6 +287,12 @@ private function openFileOrMemory(string $filename): void
         if (!$fhandle) {
             throw new ReaderException($filename . ' is an Invalid Spreadsheet file.');
         }
+        if ($this->inputEncoding === 'UTF-8') {
+            $encoding = self::guessEncodingBom($filename);
+            if ($encoding !== '') {
+                $this->inputEncoding = $encoding;
+            }
+        }
         if ($this->inputEncoding === self::GUESS_ENCODING) {
             $this->inputEncoding = self::guessEncoding($filename, $this->fallbackEncoding);
         }
@@ -313,7 +320,7 @@ public function setTestAutoDetect(bool $value): self
     private function setAutoDetect(?string $value): ?string
     {
         $retVal = null;
-        if ($value !== null && $this->testAutodetect) {
+        if ($value !== null && $this->testAutodetect && PHP_VERSION_ID < 90000) {
             $retVal2 = @ini_set('auto_detect_line_endings', $value);
             if (is_string($retVal2)) {
                 $retVal = $retVal2;
@@ -362,6 +369,20 @@ private function loadStringOrFile(string $filename, Spreadsheet $spreadsheet, bo
         // Deprecated in Php8.1
         $iniset = $this->setAutoDetect('1');
 
+        try {
+            $this->loadStringOrFile2($filename, $spreadsheet, $dataUri);
+            $this->setAutoDetect($iniset);
+        } catch (Throwable $e) {
+            $this->setAutoDetect($iniset);
+
+            throw $e;
+        }
+
+        return $spreadsheet;
+    }
+
+    private function loadStringOrFile2(string $filename, Spreadsheet $spreadsheet, bool $dataUri): void
+    {
         // Open file
         if ($dataUri) {
             $this->openDataUri($filename);
@@ -433,11 +454,6 @@ private function loadStringOrFile(string $filename, Spreadsheet $spreadsheet, bo
 
         // Close file
         fclose($fileHandle);
-
-        $this->setAutoDetect($iniset);
-
-        // Return
-        return $spreadsheet;
     }
 
     /**
@@ -545,6 +561,10 @@ public function getContiguous(): bool
      */
     public function setEscapeCharacter(string $escapeCharacter): self
     {
+        if (PHP_VERSION_ID >= 90000 && $escapeCharacter !== '') {
+            throw new ReaderException('Escape character must be null string for Php9+');
+        }
+
         $this->escapeCharacter = $escapeCharacter;
 
         return $this;
@@ -621,17 +641,15 @@ private static function guessEncodingTestBom(string &$encoding, string $first4,
         }
     }
 
-    private static function guessEncodingBom(string $filename): string
+    public static function guessEncodingBom(string $filename, ?string $convertString = null): string
     {
         $encoding = '';
-        $first4 = file_get_contents($filename, false, null, 0, 4);
-        if ($first4 !== false) {
-            self::guessEncodingTestBom($encoding, $first4, self::UTF8_BOM, 'UTF-8');
-            self::guessEncodingTestBom($encoding, $first4, self::UTF16BE_BOM, 'UTF-16BE');
-            self::guessEncodingTestBom($encoding, $first4, self::UTF32BE_BOM, 'UTF-32BE');
-            self::guessEncodingTestBom($encoding, $first4, self::UTF32LE_BOM, 'UTF-32LE');
-            self::guessEncodingTestBom($encoding, $first4, self::UTF16LE_BOM, 'UTF-16LE');
-        }
+        $first4 = $convertString ?? (string) file_get_contents($filename, false, null, 0, 4);
+        self::guessEncodingTestBom($encoding, $first4, self::UTF8_BOM, 'UTF-8');
+        self::guessEncodingTestBom($encoding, $first4, self::UTF16BE_BOM, 'UTF-16BE');
+        self::guessEncodingTestBom($encoding, $first4, self::UTF32BE_BOM, 'UTF-32BE');
+        self::guessEncodingTestBom($encoding, $first4, self::UTF32LE_BOM, 'UTF-32LE');
+        self::guessEncodingTestBom($encoding, $first4, self::UTF16LE_BOM, 'UTF-16LE');
 
         return $encoding;
     }
@@ -688,4 +706,39 @@ private static function getCsv(
 
         return fgetcsv($stream, $length, $separator, $enclosure, $escape);
     }
+
+    public static function affectedByPhp9(
+        string $filename,
+        string $inputEncoding = 'UTF-8',
+        ?string $delimiter = null,
+        string $enclosure = '"',
+        string $escapeCharacter = '\\'
+    ): bool {
+        if (PHP_VERSION_ID < 70400 || PHP_VERSION_ID >= 90000) {
+            throw new ReaderException('Function valid only for Php7.4 or Php8'); // @codeCoverageIgnore
+        }
+        $reader1 = new self();
+        $reader1->setInputEncoding($inputEncoding)
+            ->setTestAutoDetect(true)
+            ->setEscapeCharacter($escapeCharacter)
+            ->setDelimiter($delimiter)
+            ->setEnclosure($enclosure);
+        $spreadsheet1 = $reader1->load($filename);
+        $sheet1 = $spreadsheet1->getActiveSheet();
+        $array1 = $sheet1->toArray(null, false, false);
+        $spreadsheet1->disconnectWorksheets();
+
+        $reader2 = new self();
+        $reader2->setInputEncoding($inputEncoding)
+            ->setTestAutoDetect(false)
+            ->setEscapeCharacter('')
+            ->setDelimiter($delimiter)
+            ->setEnclosure($enclosure);
+        $spreadsheet2 = $reader2->load($filename);
+        $sheet2 = $spreadsheet2->getActiveSheet();
+        $array2 = $sheet2->toArray(null, false, false);
+        $spreadsheet2->disconnectWorksheets();
+
+        return $array1 !== $array2;
+    }
 }
diff --git a/tests/PhpSpreadsheetTests/Reader/Csv/CsvLineEndingTest.php b/tests/PhpSpreadsheetTests/Reader/Csv/CsvLineEndingTest.php
@@ -25,6 +25,9 @@ protected function tearDown(): void
      */
     public function testEndings(string $ending): void
     {
+        if ($ending === "\r" && PHP_VERSION_ID >= 90000) {
+            self::markTestSkipped('Mac line endings not supported for Php9+');
+        }
         $this->tempFile = $filename = File::temporaryFilename();
         $data = ['123', '456', '789'];
         file_put_contents($filename, implode($ending, $data));

diff --git a/tests/PhpSpreadsheetTests/Reader/Csv/CsvTest.php b/tests/PhpSpreadsheetTests/Reader/Csv/CsvTest.php
@@ -114,6 +114,10 @@ public static function providerCanLoad(): array
 
     public function testEscapeCharacters(): void
     {
+        if (PHP_VERSION_ID >= 90000) {
+            $this->expectException(ReaderException::class);
+            $this->expectExceptionMessage('Escape character must be null string');
+        }
         $reader = (new Csv())->setEscapeCharacter('"');
         $worksheet = $reader->load('tests/data/Reader/CSV/backslash.csv')
             ->getActiveSheet();
@@ -230,6 +234,10 @@ public function testReadNonexistentFileName(): void
      */
     public function testInferSeparator(string $escape, string $delimiter): void
     {
+        if (PHP_VERSION_ID >= 90000 && $escape !== '') {
+            $this->expectException(ReaderException::class);
+            $this->expectExceptionMessage('Escape character must be null string');
+        }
         $reader = new Csv();
         $reader->setEscapeCharacter($escape);
         $filename = 'tests/data/Reader/CSV/escape.csv';

diff --git a/tests/PhpSpreadsheetTests/Reader/Csv/Php9Test.php b/tests/PhpSpreadsheetTests/Reader/Csv/Php9Test.php
@@ -0,0 +1,37 @@
+<?php
+
+declare(strict_types=1);
+
+namespace PhpOffice\PhpSpreadsheetTests\Reader\Csv;
+
+use PhpOffice\PhpSpreadsheet\Reader\Csv;
+use PhpOffice\PhpSpreadsheet\Reader\Exception as ReaderException;
+use PHPUnit\Framework\TestCase;
+
+class Php9Test extends TestCase
+{
+    public function testAffectedByPhp9(): void
+    {
+        if (PHP_VERSION_ID >= 90000) {
+            $this->expectException(ReaderException::class);
+            $this->expectExceptionMessage('Php7.4 or Php8');
+        }
+        $dir = 'tests/data/Reader/CSV';
+        $files = glob("$dir/*");
+        self::assertNotFalse($files);
+        $affected = [];
+        foreach ($files as $file) {
+            $base = basename($file);
+            $encoding = 'UTF-8';
+            if (str_contains($base, 'utf') && !str_contains($base, 'bom')) {
+                $encoding = 'guess';
+            }
+            $result = Csv::affectedByPhp9($file, $encoding);
+            if ($result) {
+                $affected[] = $base;
+            }
+        }
+        $expected = ['backslash.csv', 'escape.csv', 'linend.mac.csv'];
+        self::assertSame($expected, $affected);
+    }
+}
diff --git a/tests/data/Reader/CSV/linend.mac.csv b/tests/data/Reader/CSV/linend.mac.csv
@@ -0,0 +1 @@
+A,12,3

diff --git a/tests/data/Reader/CSV/linend.unix.csv b/tests/data/Reader/CSV/linend.unix.csv
@@ -0,0 +1,2 @@
+A,1
+2,3
diff --git a/tests/data/Reader/CSV/linend.win.csv b/tests/data/Reader/CSV/linend.win.csv
@@ -0,0 +1,2 @@
+A,1
+2,3