Notice and error when reading .xls (BIFF8) having invalid xfIndex

[dendemann]

This is:

- [x] a bug report
- [ ] a feature request
- [x] not a usage question

What is the expected behavior?

Reading an Excel 97-2003 document shouldn't throw notices or errors.

What is the current behavior?

Reading an Excel 97-2003 documents isn't possible when the data is corrupted in a certain way.

What are the steps to reproduce?

Legacy .xls documents created with third party libraries might contain corrupt data. In my case xfIndex read from a cell is 0 while there's only one entry within the mapCellXfIndex mapping array at index position 15. Therefore there will be an unchecked array access at pos 0.

Unfortunately I can't provide the file as it contains sensitive customer data. But I found out that, according to the shared strings table (sst), it was originally created by the library "GemBox.Spreadsheet Professional 3.3 for .NET 4.0, Version=33.3.40.1062".

Another guy ran into the same problem but I think no one was looking into it: 432120730

Reading the file with Excel or LibreOffice works without problems.

<?php

use PhpOffice\PhpSpreadsheet\Reader\Xls;

require __DIR__ . '/vendor/autoload.php';

$reader = new Xls();
$spreadsheet = $reader->load(__DIR__ . '/broken.xls');

$cell = $spreadsheet->getActiveSheet()->getCell('A1');
var_dump($cell->getFormattedValue());

PHP Notice:  Undefined offset: 0 in /home/user/test/vendor/phpoffice/phpspreadsheet/src/PhpSpreadsheet/Reader/Xls.php on line 3830
PHP Stack trace:
PHP   1. {main}() /home/user/test/test.php:0
PHP   2. PhpOffice\PhpSpreadsheet\Reader\Xls->load() /home/user/test/test.php:8
PHP   3. PhpOffice\PhpSpreadsheet\Reader\Xls->readLabelSst() /home/user/test/vendor/phpoffice/phpspreadsheet/src/PhpSpreadsheet/Reader/Xls.php:956
PHP Notice:  Undefined index:  in /home/user/test/vendor/phpoffice/phpspreadsheet/src/PhpSpreadsheet/Spreadsheet.php on line 1034
PHP Stack trace:
PHP   1. {main}() /home/user/test/test.php:0
PHP   2. PhpOffice\PhpSpreadsheet\Cell\Cell->getFormattedValue() /home/user/test/test.php:11
PHP   3. PhpOffice\PhpSpreadsheet\Style\NumberFormat->getFormatCode() /home/user/test/vendor/phpoffice/phpspreadsheet/src/PhpSpreadsheet/Cell/Cell.php:167
PHP   4. PhpOffice\PhpSpreadsheet\Style\NumberFormat->getSharedComponent() /home/user/test/vendor/phpoffice/phpspreadsheet/src/PhpSpreadsheet/Style/NumberFormat.php:165
PHP   5. PhpOffice\PhpSpreadsheet\Style\Style->getSharedComponent() /home/user/test/vendor/phpoffice/phpspreadsheet/src/PhpSpreadsheet/Style/NumberFormat.php:112
PHP   6. PhpOffice\PhpSpreadsheet\Spreadsheet->getCellXfByIndex() /home/user/test/vendor/phpoffice/phpspreadsheet/src/PhpSpreadsheet/Style/Style.php:124
PHP Fatal error:  Uncaught Error: Call to a member function getNumberFormat() on null in /home/user/test/vendor/phpoffice/phpspreadsheet/src/PhpSpreadsheet/Style/NumberFormat.php:112
Stack trace:
#0 /home/user/test/vendor/phpoffice/phpspreadsheet/src/PhpSpreadsheet/Style/NumberFormat.php(165): PhpOffice\PhpSpreadsheet\Style\NumberFormat->getSharedComponent()
#1 /home/user/test/vendor/phpoffice/phpspreadsheet/src/PhpSpreadsheet/Cell/Cell.php(167): PhpOffice\PhpSpreadsheet\Style\NumberFormat->getFormatCode()
#2 /home/user/test/test.php(11): PhpOffice\PhpSpreadsheet\Cell\Cell->getFormattedValue()
#3 {main}
  thrown in /home/user/test/vendor/phpoffice/phpspreadsheet/src/PhpSpreadsheet/Style/NumberFormat.php on line 112

Which versions of PhpSpreadsheet and PHP are affected?

At least the following: 1.6.0, 1.9.0 (current)

[PowerKiKi]

Fixed via #1239