Could not open CSV file containing HTML fragment

We now always trust the file extension to avoid false positive of mime
detection for most simple cases. But we still try to guess the mime type
if the file extension does not match or is missing.

Fixes #564

CHANGELOG.md

@@ -15,6 +15,7 @@ and this project adheres to [Semantic Versioning](http://semver.org/).
 
 - Xlsx reader crashed when reading a file with workbook protection - [#553](https://github.com/PHPOffice/PhpSpreadsheet/pull/553)
 - Cell formats with escaped spaces were causing incorrect date formatting - [#557](https://github.com/PHPOffice/PhpSpreadsheet/issues/557)
+- Could not open CSV file containing HTML fragment - [##564](https://github.com/PHPOffice/PhpSpreadsheet/issues/564)
 
 ## [1.3.1] - 2018-06-12
 

src/PhpSpreadsheet/Reader/Csv.php

@@ -507,6 +507,12 @@ class Csv extends BaseReader
 
         fclose($this->fileHandle);
 
+        // Trust file extension if any
+        if (strtolower(pathinfo($pFilename, PATHINFO_EXTENSION)) === 'csv') {
+            return true;
+        }
+
+        // Attempt to guess mimetype
         $type = mime_content_type($pFilename);
         $supportedTypes = [
             'text/csv',

tests/PhpSpreadsheetTests/Reader/CsvTest.php

@@ -83,6 +83,8 @@ class CsvTest extends TestCase
             [false, 'data/Reader/Xml/WithoutStyle.xml'],
             [true, 'data/Reader/CSV/enclosure.csv'],
             [true, 'data/Reader/CSV/semicolon_separated.csv'],
+            [true, 'data/Reader/CSV/contains_html.csv'],
+            [true, 'data/Reader/CSV/csv_without_extension'],
             [true, 'data/Reader/HTML/csv_with_angle_bracket.csv'],
             [true, 'data/Reader/CSV/empty.csv'],
             [true, '../samples/Reader/sampleData/example1.csv'],

tests/data/Reader/CSV/contains_html.csv

@@ -0,0 +1 @@
+"Column A";"<a href=''>Link</a>";"Column C"

tests/data/Reader/CSV/csv_without_extension

@@ -0,0 +1,3 @@
+This;Are;Headers
+Cell A2;Number with comma;25,5
+Two colons and a comma;B|3;:,: