From 813855b2b2a4f39c73abc72a757730d33a7d9381 Mon Sep 17 00:00:00 2001
From: Paul Barton <info@thisispaul.co.uk>
Date: Wed, 10 Oct 2018 15:27:14 +0100
Subject: [PATCH] Fix CSV delimiter detection on line breaks

The CSV Reader can now correctly ignore line breaks inside
enclosures which allows it to determine the delimiter
correctly.

Fixes #716
Fixes #717
---
 CHANGELOG.md                                  |  1 +
 src/PhpSpreadsheet/Reader/Csv.php             | 42 ++++++++++++++++---
 tests/PhpSpreadsheetTests/Reader/CsvTest.php  |  6 +++
 .../Reader/CSV/line_break_in_enclosure.csv    | 18 ++++++++
 4 files changed, 62 insertions(+), 5 deletions(-)
 create mode 100644 tests/data/Reader/CSV/line_break_in_enclosure.csv

diff --git a/CHANGELOG.md b/CHANGELOG.md
index cac43c9d..31048b70 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -17,6 +17,7 @@ and this project adheres to [Semantic Versioning](http://semver.org/).
 - Xls file cause the exception during open by Xls reader - [#402](https://github.com/PHPOffice/PhpSpreadsheet/issues/402)
 - Skip non numeric value in SUMIF - [#618](https://github.com/PHPOffice/PhpSpreadsheet/pull/618)
 - OFFSET should allow omitted height and width - [#561](https://github.com/PHPOffice/PhpSpreadsheet/issues/561)
+- Correctly determine delimiter when CSV contains line breaks inside enclosures - [#716](https://github.com/PHPOffice/PhpSpreadsheet/issues/716)
 
 ## [1.4.1] - 2018-09-30
 
diff --git a/src/PhpSpreadsheet/Reader/Csv.php b/src/PhpSpreadsheet/Reader/Csv.php
index 6899773a..0e18bcc3 100644
--- a/src/PhpSpreadsheet/Reader/Csv.php
+++ b/src/PhpSpreadsheet/Reader/Csv.php
@@ -163,11 +163,7 @@ class Csv extends BaseReader
 
         // Count how many times each of the potential delimiters appears in each line
         $numberLines = 0;
-        while (($line = fgets($this->fileHandle)) !== false && (++$numberLines < 1000)) {
-            // Drop everything that is enclosed to avoid counting false positives in enclosures
-            $enclosure = preg_quote($this->enclosure, '/');
-            $line = preg_replace('/(' . $enclosure . '.*' . $enclosure . ')/U', '', $line);
-
+        while (($line = $this->getNextLine()) !== false && (++$numberLines < 1000)) {
             $countLine = [];
             for ($i = strlen($line) - 1; $i >= 0; --$i) {
                 $char = $line[$i];
@@ -230,6 +226,42 @@ class Csv extends BaseReader
         return $this->skipBOM();
     }
 
+    /**
+     * Get the next full line from the file.
+     *
+     * @param string $line
+     *
+     * @return bool|string
+     */
+    private function getNextLine($line = '')
+    {
+        // Get the next line in the file
+        $newLine = fgets($this->fileHandle);
+
+        // Return false if there is no next line
+        if ($newLine === false) {
+            return false;
+        }
+
+        // Add the new line to the line passed in
+        $line = $line . $newLine;
+
+        // Drop everything that is enclosed to avoid counting false positives in enclosures
+        $enclosure = preg_quote($this->enclosure, '/');
+        $line = preg_replace('/(' . $enclosure . '.*' . $enclosure . ')/U', '', $line);
+
+        // See if we have any enclosures left in the line
+        $matches = [];
+        preg_match('/(' . $enclosure . ')/', $line, $matches);
+
+        // if we still have an enclosure then we need to read the next line aswell
+        if (count($matches) > 0) {
+            $line = $this->getNextLine($line);
+        }
+
+        return $line;
+    }
+
     /**
      * Return worksheet info (Name, Last Column Letter, Last Column Index, Total Rows, Total Columns).
      *
diff --git a/tests/PhpSpreadsheetTests/Reader/CsvTest.php b/tests/PhpSpreadsheetTests/Reader/CsvTest.php
index eeddbb08..e748cd1b 100644
--- a/tests/PhpSpreadsheetTests/Reader/CsvTest.php
+++ b/tests/PhpSpreadsheetTests/Reader/CsvTest.php
@@ -43,6 +43,12 @@ class CsvTest extends TestCase
                 'C2',
                 '25,5',
             ],
+            [
+                __DIR__ . '/../../data/Reader/CSV/line_break_in_enclosure.csv',
+                ',',
+                'A3',
+                'Test',
+            ],
             [
                 __DIR__ . '/../../data/Reader/HTML/csv_with_angle_bracket.csv',
                 ',',
diff --git a/tests/data/Reader/CSV/line_break_in_enclosure.csv b/tests/data/Reader/CSV/line_break_in_enclosure.csv
new file mode 100644
index 00000000..70e84bcd
--- /dev/null
+++ b/tests/data/Reader/CSV/line_break_in_enclosure.csv
@@ -0,0 +1,18 @@
+Name,Copy,URL
+Test,"This is a test
+with line breaks
+that breaks the
+delimiters",http://google.com
+Test,"This is a test
+with line breaks
+that breaks the
+delimiters",http://google.com
+Test,"This is a test
+with line breaks
+that breaks the
+delimiters",http://google.com
+Test,"This is a test
+with line breaks
+that breaks the
+delimiters",http://google.com
+Test,"This is a test",http://google.com