Merge pull request #4344 from oleibman/parseutf8

Writer/Xls/Parser::advance Should Parse by Character

Author: oleibman

CHANGELOG.md

@@ -46,7 +46,8 @@ and this project adheres to [Semantic Versioning](https://semver.org).
 
 ### Fixed
 
-- Xls Writer Parser Mishandling True/False Argument. [Issue #4331](https://github.com/PHPOffice/PhpSpreadsheet/issues/4331) [PR #4333](https://github.com/PHPOffice/PhpSpreadsheet/pull/4333)
+- Xls writer Parser Mishandling True/False Argument. [Issue #4331](https://github.com/PHPOffice/PhpSpreadsheet/issues/4331) [PR #4333](https://github.com/PHPOffice/PhpSpreadsheet/pull/4333)
+- Xls writer Parser Parse By Character Not Byte. [PR #4344](https://github.com/PHPOffice/PhpSpreadsheet/pull/4344)
 - Minor changes to dynamic array calculations exposed by using explicit array return types in some tests. [PR #4328](https://github.com/PHPOffice/PhpSpreadsheet/pull/4328)
 
 ## 2025-01-26 - 3.9.0

src/PhpSpreadsheet/Writer/Xls/Parser.php

@@ -67,6 +67,8 @@ class Parser
         . '[$]?[A-Ia-i]?[A-Za-z][$]?(\\d+)'
         . '$~u';
 
+    private const UTF8 = 'UTF-8';
+
     /**
      * The index of the character we are currently looking at.
      */
@@ -991,37 +993,38 @@ private function advance(): void
     {
         $token = '';
         $i = $this->currentCharacter;
-        $formula_length = strlen($this->formula);
+        $formula = mb_str_split($this->formula, 1, self::UTF8);
+        $formula_length = count($formula);
         // eat up white spaces
         if ($i < $formula_length) {
-            while ($this->formula[$i] == ' ') {
+            while ($formula[$i] === ' ') {
                 ++$i;
             }
 
             if ($i < ($formula_length - 1)) {
-                $this->lookAhead = $this->formula[$i + 1];
+                $this->lookAhead = $formula[$i + 1];
             }
             $token = '';
         }
 
         while ($i < $formula_length) {
-            $token .= $this->formula[$i];
+            $token .= $formula[$i];
 
             if ($i < ($formula_length - 1)) {
-                $this->lookAhead = $this->formula[$i + 1];
+                $this->lookAhead = $formula[$i + 1];
             } else {
                 $this->lookAhead = '';
             }
 
-            if ($this->match($token) != '') {
+            if ($this->match($token) !== '') {
                 $this->currentCharacter = $i + 1;
                 $this->currentToken = $token;
 
                 return;
             }
 
             if ($i < ($formula_length - 2)) {
-                $this->lookAhead = $this->formula[$i + 2];
+                $this->lookAhead = $formula[$i + 2];
             } else { // if we run out of characters lookAhead becomes empty
                 $this->lookAhead = '';
             }
@@ -1198,8 +1201,8 @@ private function match(string $token): string
     public function parse(string $formula): bool
     {
         $this->currentCharacter = 0;
-        $this->formula = (string) $formula;
-        $this->lookAhead = $formula[1] ?? '';
+        $this->formula = $formula;
+        $this->lookAhead = mb_substr($formula, 1, 1, self::UTF8);
         $this->advance();
         $this->parseTree = $this->condition();
 

tests/PhpSpreadsheetTests/Writer/Xls/NonLatinFormulasTest.php

@@ -0,0 +1,65 @@
+<?php
+
+namespace PhpOffice\PhpSpreadsheetTests\Writer\Xls;
+
+use PhpOffice\PhpSpreadsheet\Cell\DataValidation;
+use PhpOffice\PhpSpreadsheet\Cell\DataValidator;
+use PhpOffice\PhpSpreadsheet\NamedRange;
+use PhpOffice\PhpSpreadsheet\Spreadsheet;
+use PhpOffice\PhpSpreadsheetTests\Functional\AbstractFunctional;
+
+class NonLatinFormulasTest extends AbstractFunctional
+{
+    public function testNonLatin(): void
+    {
+        $spreadsheet = new Spreadsheet();
+        $worksheet = $spreadsheet->getActiveSheet();
+
+        $validation = $worksheet->getCell('B1')->getDataValidation();
+        $validation->setType(DataValidation::TYPE_LIST);
+        $validation->setErrorStyle(DataValidation::STYLE_STOP);
+        $validation->setAllowBlank(false);
+        $validation->setShowInputMessage(true);
+        $validation->setShowErrorMessage(true);
+        $validation->setShowDropDown(true);
+        $validation->setFormula1('"слово, сло"');
+
+        $dataValidator = new DataValidator();
+        $worksheet->getCell('B1')->setValue('слово');
+        self::assertTrue(
+            $dataValidator->isValid($worksheet->getCell('B1'))
+        );
+        $worksheet->getCell('B1')->setValue('слов');
+        self::assertFalse(
+            $dataValidator->isValid($worksheet->getCell('B1'))
+        );
+
+        $worksheet->setTitle('словслов');
+        $worksheet->getCell('A1')->setValue('=словслов!B1');
+        $worksheet->getCell('A2')->setValue("='словслов'!B1");
+        $spreadsheet->addNamedRange(new NamedRange('слсл', $worksheet, '$B$1'));
+        $worksheet->getCell('A3')->setValue('=слсл');
+
+        $robj = $this->writeAndReload($spreadsheet, 'Xls');
+        $spreadsheet->disconnectWorksheets();
+        $sheet0 = $robj->getActiveSheet();
+        self::assertSame('словслов', $sheet0->getTitle());
+        self::assertSame('=словслов!B1', $sheet0->getCell('A1')->getValue());
+        self::assertSame('слов', $sheet0->getCell('A1')->getCalculatedValue());
+        // Quotes around sheet name are stripped off - harmless
+        //self::assertSame("='словслов'!B1", $sheet0->getCell('A2')->getValue());
+        self::assertSame('слов', $sheet0->getCell('A2')->getCalculatedValue());
+        // Formulas with defined names don't work in Xls Writer
+        //self::assertSame('=слсл', $sheet0->getCell('A3')->getValue());
+        // But result should be accurate
+        self::assertSame('слов', $sheet0->getCell('A3')->getCalculatedValue());
+        $names = $robj->getDefinedNames();
+        self::assertCount(1, $names);
+        // name has been uppercased
+        $namedRange = $names['СЛСЛ'] ?? null;
+        self::assertInstanceOf(NamedRange::class, $namedRange);
+        self::assertSame('$B$1', $namedRange->getRange());
+
+        $robj->disconnectWorksheets();
+    }
+}