WIP Xlsx Reader Vml File with Unclosed br Tags

src/PhpSpreadsheet/Reader/Xlsx.php

@@ -121,9 +121,12 @@ public static function falseToArray($value): array
         return is_array($value) ? $value : [];
     }
 
-    private function loadZip(string $filename, string $ns = ''): SimpleXMLElement
+    private function loadZip(string $filename, string $ns = '', bool $replaceUnclosedBr = false): SimpleXMLElement
     {
         $contents = $this->getFromZipArchive($this->zip, $filename);
+        if ($replaceUnclosedBr) {
+            $contents = str_replace('<br>', '<br/>', $contents);
+        }
         $rels = simplexml_load_string(
             $this->securityScanner->scan($contents),
             'SimpleXMLElement',
@@ -1030,7 +1033,7 @@ protected function loadSpreadsheetFromFile(string $filename): Spreadsheet
 
                                     try {
                                         // no namespace okay - processed with Xpath
-                                        $vmlCommentsFile = $this->loadZip($relPath, '');
+                                        $vmlCommentsFile = $this->loadZip($relPath, '', true);
                                         $vmlCommentsFile->registerXPathNamespace('v', Namespaces::URN_VML);
                                     } catch (Throwable $ex) {
                                         //Ignore unparsable vmlDrawings. Later they will be moved from $unparsedVmlDrawings to $unparsedLoadedData

tests/PhpSpreadsheetTests/Reader/Xlsx/ColorTabTest.php

@@ -12,6 +12,30 @@ public function testIssue2316(): void
         $filename = 'tests/data/Reader/XLSX/colortabs.xlsx';
         $reader = new Xlsx();
         $spreadsheet = $reader->load($filename);
+        $unparsed = $spreadsheet->getUnparsedLoadedData();
+        self::assertArrayHasKey('sheets', $unparsed);
+        self::assertArrayHasKey('Worksheet', $unparsed['sheets']);
+        self::assertArrayNotHasKey('vmlDrawings', $unparsed['sheets']['Worksheet']);
+
+        // theme color
+        self::assertSame('FF548135', $spreadsheet->getSheet(0)->getTabColor()->getArgb());
+        // rgb color
+        self::assertSame('FFFFC000', $spreadsheet->getSheet(1)->getTabColor()->getArgb());
+        $spreadsheet->disconnectWorksheets();
+    }
+
+    public function testIssue3125(): void
+    {
+        // Very artificial - no real sample file to go with issue.
+        $filename = 'tests/data/Reader/XLSX/colortabs.badbr.xlsx';
+        $reader = new Xlsx();
+        $spreadsheet = $reader->load($filename);
+        $unparsed = $spreadsheet->getUnparsedLoadedData();
+        self::assertArrayHasKey('sheets', $unparsed);
+        self::assertArrayHasKey('Worksheet', $unparsed['sheets']);
+        // Before fix, vml in sample file was unparseable as xml (unclosed br tag),
+        // so condition below would fail.
+        self::assertArrayNotHasKey('vmlDrawings', $unparsed['sheets']['Worksheet']);
 
         // theme color
         self::assertSame('FF548135', $spreadsheet->getSheet(0)->getTabColor()->getArgb());