[dx] Add HTMLAverseRectorInterface to teach rules to skip HTML + PHP …

…mix that would go wrong (#6445)

* skip html + php mix

* [dx] Add HTMLAverseRectorInterface

rules-tests/CodingStyle/Rector/Stmt/NewlineAfterStatementRector/Fixture/skip_html_php_mix.php.inc

@@ -0,0 +1,17 @@
+<?php
+
+namespace Rector\Tests\CodingStyle\Rector\Stmt\NewlineAfterStatementRector\Fixture;
+
+if (mt_rand(0, 1)) { ?>
+    <div>
+        <h1>Hi</h1>
+    </div>
+
+<?php } else { ?>
+    <div>
+        <h1>Bye</h1>
+    </div>
+
+<?php } ?>
+
+End of file

rules/CodingStyle/Rector/ClassMethod/NewlineBeforeNewAssignSetRector.php

@@ -15,14 +15,15 @@
 use PhpParser\Node\Stmt\Expression;
 use PhpParser\Node\Stmt\Function_;
 use PhpParser\Node\Stmt\Nop;
+use Rector\Contract\Rector\HTMLAverseRectorInterface;
 use Rector\Rector\AbstractRector;
 use Symplify\RuleDocGenerator\ValueObject\CodeSample\CodeSample;
 use Symplify\RuleDocGenerator\ValueObject\RuleDefinition;
 
 /**
  * @see \Rector\Tests\CodingStyle\Rector\ClassMethod\NewlineBeforeNewAssignSetRector\NewlineBeforeNewAssignSetRectorTest
  */
-final class NewlineBeforeNewAssignSetRector extends AbstractRector
+final class NewlineBeforeNewAssignSetRector extends AbstractRector implements HTMLAverseRectorInterface
 {
     private ?string $previousStmtVariableName = null;
 

rules/CodingStyle/Rector/Stmt/NewlineAfterStatementRector.php

@@ -24,6 +24,7 @@
 use PhpParser\Node\Stmt\TryCatch;
 use PhpParser\Node\Stmt\While_;
 use Rector\Contract\PhpParser\Node\StmtsAwareInterface;
+use Rector\Contract\Rector\HTMLAverseRectorInterface;
 use Rector\NodeTypeResolver\Node\AttributeKey;
 use Rector\Rector\AbstractRector;
 use Symplify\RuleDocGenerator\ValueObject\CodeSample\CodeSample;
@@ -32,7 +33,7 @@
 /**
  * @see \Rector\Tests\CodingStyle\Rector\Stmt\NewlineAfterStatementRector\NewlineAfterStatementRectorTest
  */
-final class NewlineAfterStatementRector extends AbstractRector
+final class NewlineAfterStatementRector extends AbstractRector implements HTMLAverseRectorInterface
 {
     /**
      * @var array<class-string<Node>>

rules/TypeDeclaration/Rector/StmtsAwareInterface/DeclareStrictTypesRector.php

@@ -10,20 +10,21 @@
 use PhpParser\Node\Stmt;
 use PhpParser\Node\Stmt\Declare_;
 use PhpParser\Node\Stmt\DeclareDeclare;
-use PhpParser\Node\Stmt\InlineHTML;
 use PhpParser\Node\Stmt\Nop;
 use Rector\ChangesReporting\ValueObject\RectorWithLineChange;
 use Rector\Contract\PhpParser\Node\StmtsAwareInterface;
+use Rector\Contract\Rector\HTMLAverseRectorInterface;
 use Rector\PhpParser\Node\CustomNode\FileWithoutNamespace;
 use Rector\Rector\AbstractRector;
 use Rector\TypeDeclaration\NodeAnalyzer\DeclareStrictTypeFinder;
+use Rector\ValueObject\Application\File;
 use Symplify\RuleDocGenerator\ValueObject\CodeSample\CodeSample;
 use Symplify\RuleDocGenerator\ValueObject\RuleDefinition;
 
 /**
  * @see \Rector\Tests\TypeDeclaration\Rector\StmtsAwareInterface\DeclareStrictTypesRector\DeclareStrictTypesRectorTest
  */
-final class DeclareStrictTypesRector extends AbstractRector
+final class DeclareStrictTypesRector extends AbstractRector implements HTMLAverseRectorInterface
 {
     public function __construct(
         private readonly DeclareStrictTypeFinder $declareStrictTypeFinder
@@ -65,6 +66,10 @@ public function beforeTraverse(array $nodes): ?array
             return null;
         }
 
+        if ($this->startWithShebang($this->file)) {
+            return null;
+        }
+
         if ($nodes === []) {
             return null;
         }
@@ -79,10 +84,6 @@ public function beforeTraverse(array $nodes): ?array
                 return null;
             }
 
-            if ($currentStmt instanceof InlineHTML) {
-                return null;
-            }
-
             $nodes = $rootStmt->stmts;
             $stmt = $currentStmt;
         }
@@ -124,4 +125,9 @@ public function refactor(Node $node): ?Node
         // workaroudn, as Rector now only hooks to specific nodes, not arrays
         return null;
     }
+
+    private function startWithShebang(File $file): bool
+    {
+        return str_starts_with($file->getFileContent(), '#!');
+    }
 }

src/Contract/Rector/HTMLAverseRectorInterface.php

@@ -0,0 +1,13 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Rector\Contract\Rector;
+
+/**
+ * Rector rule with this marker interface will skip all files
+ * with any HTML node. This is practical to avoid malformed PHP+HTML files
+ */
+interface HTMLAverseRectorInterface
+{
+}

src/Rector/AbstractRector.php

@@ -23,6 +23,7 @@
 use Rector\Application\Provider\CurrentFileProvider;
 use Rector\BetterPhpDocParser\PhpDocInfo\PhpDocInfo;
 use Rector\ChangesReporting\ValueObject\RectorWithLineChange;
+use Rector\Contract\Rector\HTMLAverseRectorInterface;
 use Rector\Contract\Rector\RectorInterface;
 use Rector\Exception\ShouldNotHappenException;
 use Rector\NodeDecorator\CreatedByRuleDecorator;
@@ -125,6 +126,10 @@ final public function enterNode(Node $node): int|Node|null
             return null;
         }
 
+        if (is_a($this, HTMLAverseRectorInterface::class, true) && $this->file->containsHTML()) {
+            return null;
+        }
+
         $filePath = $this->file->getFilePath();
         if ($this->skipper->shouldSkipCurrentNode($this, $filePath, static::class, $node)) {
             return null;
@@ -134,6 +139,7 @@ final public function enterNode(Node $node): int|Node|null
 
         // ensure origNode pulled before refactor to avoid changed during refactor, ref https://3v4l.org/YMEGN
         $originalNode = $node->getAttribute(AttributeKey::ORIGINAL_NODE) ?? $node;
+
         $refactoredNode = $this->refactor($node);
 
         // @see NodeTraverser::* codes, e.g. removal of node of stopping the traversing

src/ValueObject/Application/File.php

@@ -6,6 +6,8 @@
 
 use PhpParser\Node;
 use PhpParser\Node\Stmt;
+use PhpParser\Node\Stmt\InlineHTML;
+use PhpParser\NodeFinder;
 use Rector\ChangesReporting\ValueObject\RectorWithLineChange;
 use Rector\Exception\ShouldNotHappenException;
 use Rector\ValueObject\Reporting\FileDiff;
@@ -38,6 +40,11 @@ final class File
      */
     private array $rectorWithLineChanges = [];
 
+    /**
+     * Cached result per file
+     */
+    private ?bool $containsHtml = null;
+
     public function __construct(
         private readonly string $filePath,
         private string $fileContent
@@ -150,4 +157,16 @@ public function getRectorWithLineChanges(): array
     {
         return $this->rectorWithLineChanges;
     }
+
+    public function containsHTML(): bool
+    {
+        if ($this->containsHtml !== null) {
+            return $this->containsHtml;
+        }
+
+        $nodeFinder = new NodeFinder();
+
+        $this->containsHtml = (bool) $nodeFinder->findFirstInstanceOf($this->oldStmts, InlineHTML::class);
+        return $this->containsHtml;
+    }
 }