Add matchStrictGroups and other strict group operations to avoid nullable matches (#14)

Author: Seldaek

README.md

@@ -80,6 +80,17 @@ if (Preg::isMatch('{fo+}', $string, $matches)) // bool
 if (Preg::isMatchAll('{fo+}', $string, $matches)) // bool
 ```
 
+Finally the `Preg` class provides a few `*StrictGroups` method variants that ensure match groups
+are always present and thus non-nullable, making it easier to write type-safe code:
+
+```php
+use Composer\Pcre\Preg;
+
+// $matches is guaranteed to be an array of strings, if a subpattern does not match and produces a null it will throw
+if (Preg::matchStrictGroups('{fo+}', $string, $matches))
+if (Preg::matchAllStrictGroups('{fo+}', $string, $matches))
+```
+
 If you would prefer a slightly more verbose usage, replacing by-ref arguments by result objects, you can use the `Regex` class:
 
 ```php

src/MatchAllResult.php

@@ -35,9 +35,9 @@ final class MatchAllResult
 
     /**
      * @param 0|positive-int $count
-     * @param array<array<string|null>> $matches
+     * @param array<int|string, array<string|null>> $matches
      */
-    public function __construct($count, array $matches)
+    public function __construct(int $count, array $matches)
     {
         $this->matches = $matches;
         $this->matched = (bool) $count;

src/MatchAllStrictGroupsResult.php

@@ -0,0 +1,46 @@
+<?php
+
+/*
+ * This file is part of composer/pcre.
+ *
+ * (c) Composer <https://github.com/composer>
+ *
+ * For the full copyright and license information, please view
+ * the LICENSE file that was distributed with this source code.
+ */
+
+namespace Composer\Pcre;
+
+final class MatchAllStrictGroupsResult
+{
+    /**
+     * An array of match group => list of matched strings
+     *
+     * @readonly
+     * @var array<int|string, list<string>>
+     */
+    public $matches;
+
+    /**
+     * @readonly
+     * @var 0|positive-int
+     */
+    public $count;
+
+    /**
+     * @readonly
+     * @var bool
+     */
+    public $matched;
+
+    /**
+     * @param 0|positive-int $count
+     * @param array<array<string>> $matches
+     */
+    public function __construct(int $count, array $matches)
+    {
+        $this->matches = $matches;
+        $this->matched = (bool) $count;
+        $this->count = $count;
+    }
+}

src/MatchAllWithOffsetsResult.php

@@ -39,7 +39,7 @@ final class MatchAllWithOffsetsResult
      * @param array<int|string, list<array{string|null, int}>> $matches
      * @phpstan-param array<int|string, list<array{string|null, int<-1, max>}>> $matches
      */
-    public function __construct($count, array $matches)
+    public function __construct(int $count, array $matches)
     {
         $this->matches = $matches;
         $this->matched = (bool) $count;

src/MatchResult.php

@@ -31,7 +31,7 @@ final class MatchResult
      * @param 0|positive-int $count
      * @param array<string|null> $matches
      */
-    public function __construct($count, array $matches)
+    public function __construct(int $count, array $matches)
     {
         $this->matches = $matches;
         $this->matched = (bool) $count;

src/MatchStrictGroupsResult.php

@@ -0,0 +1,39 @@
+<?php
+
+/*
+ * This file is part of composer/pcre.
+ *
+ * (c) Composer <https://github.com/composer>
+ *
+ * For the full copyright and license information, please view
+ * the LICENSE file that was distributed with this source code.
+ */
+
+namespace Composer\Pcre;
+
+final class MatchStrictGroupsResult
+{
+    /**
+     * An array of match group => string matched
+     *
+     * @readonly
+     * @var array<int|string, string>
+     */
+    public $matches;
+
+    /**
+     * @readonly
+     * @var bool
+     */
+    public $matched;
+
+    /**
+     * @param 0|positive-int $count
+     * @param array<string> $matches
+     */
+    public function __construct(int $count, array $matches)
+    {
+        $this->matches = $matches;
+        $this->matched = (bool) $count;
+    }
+}

src/MatchWithOffsetsResult.php

@@ -33,7 +33,7 @@ final class MatchWithOffsetsResult
      * @param array<array{string|null, int}> $matches
      * @phpstan-param array<int|string, array{string|null, int<-1, max>}> $matches
      */
-    public function __construct($count, array $matches)
+    public function __construct(int $count, array $matches)
     {
         $this->matches = $matches;
         $this->matched = (bool) $count;

src/Preg.php

@@ -38,6 +38,25 @@ public static function match(string $pattern, string $subject, ?array &$matches
         return $result;
     }
 
+    /**
+     * Variant of `match()` which outputs non-null matches (or throws)
+     *
+     * @param non-empty-string $pattern
+     * @param array<string> $matches Set by method
+     * @param int-mask<PREG_UNMATCHED_AS_NULL> $flags PREG_UNMATCHED_AS_NULL is always set, no other flags are supported
+     * @return 0|1
+     * @throws UnexpectedNullMatchException
+     *
+     * @param-out array<int|string, string> $matches
+     */
+    public static function matchStrictGroups(string $pattern, string $subject, ?array &$matches = null, int $flags = 0, int $offset = 0): int
+    {
+        $result = self::match($pattern, $subject, $matchesInternal, $flags, $offset);
+        $matches = self::enforceNonNullMatches($pattern, $matchesInternal, 'match');
+
+        return $result;
+    }
+
     /**
      * Runs preg_match with PREG_OFFSET_CAPTURE
      *
@@ -61,18 +80,15 @@ public static function matchWithOffsets(string $pattern, string $subject, ?array
     /**
      * @param non-empty-string   $pattern
      * @param array<int|string, list<string|null>> $matches Set by method
-     * @param int-mask<PREG_UNMATCHED_AS_NULL|PREG_SET_ORDER> $flags PREG_UNMATCHED_AS_NULL is always set, no other flags are supported
+     * @param int-mask<PREG_UNMATCHED_AS_NULL> $flags PREG_UNMATCHED_AS_NULL is always set, no other flags are supported
      * @return 0|positive-int
      *
      * @param-out array<int|string, list<string|null>> $matches
      */
     public static function matchAll(string $pattern, string $subject, ?array &$matches = null, int $flags = 0, int $offset = 0): int
     {
         self::checkOffsetCapture($flags, 'matchAllWithOffsets');
-
-        if (($flags & PREG_SET_ORDER) !== 0) {
-            throw new \InvalidArgumentException('PREG_SET_ORDER is not supported as it changes the type of $matches');
-        }
+        self::checkSetOrder($flags);
 
         $result = preg_match_all($pattern, $subject, $matches, $flags | PREG_UNMATCHED_AS_NULL, $offset);
         if (!is_int($result)) { // PHP < 8 may return null, 8+ returns int|false
@@ -82,6 +98,25 @@ public static function matchAll(string $pattern, string $subject, ?array &$match
         return $result;
     }
 
+    /**
+     * Variant of `match()` which outputs non-null matches (or throws)
+     *
+     * @param non-empty-string   $pattern
+     * @param array<int|string, list<string|null>> $matches Set by method
+     * @param int-mask<PREG_UNMATCHED_AS_NULL> $flags PREG_UNMATCHED_AS_NULL is always set, no other flags are supported
+     * @return 0|positive-int
+     * @throws UnexpectedNullMatchException
+     *
+     * @param-out array<int|string, list<string>> $matches
+     */
+    public static function matchAllStrictGroups(string $pattern, string $subject, ?array &$matches = null, int $flags = 0, int $offset = 0): int
+    {
+        $result = self::matchAll($pattern, $subject, $matchesInternal, $flags, $offset);
+        $matches = self::enforceNonNullMatchAll($pattern, $matchesInternal, 'matchAll');
+
+        return $result;
+    }
+
     /**
      * Runs preg_match_all with PREG_OFFSET_CAPTURE
      *
@@ -94,6 +129,8 @@ public static function matchAll(string $pattern, string $subject, ?array &$match
      */
     public static function matchAllWithOffsets(string $pattern, string $subject, ?array &$matches, int $flags = 0, int $offset = 0): int
     {
+        self::checkSetOrder($flags);
+
         $result = preg_match_all($pattern, $subject, $matches, $flags | PREG_UNMATCHED_AS_NULL | PREG_OFFSET_CAPTURE, $offset);
         if (!is_int($result)) { // PHP < 8 may return null, 8+ returns int|false
             throw PcreException::fromFunction('preg_match_all', $pattern);
@@ -241,6 +278,8 @@ public static function grep(string $pattern, array $array, int $flags = 0): arra
     }
 
     /**
+     * Variant of match() which returns a bool instead of int
+     *
      * @param non-empty-string   $pattern
      * @param array<string|null> $matches Set by method
      * @param int-mask<PREG_UNMATCHED_AS_NULL> $flags PREG_UNMATCHED_AS_NULL is always set, no other flags are supported
@@ -253,6 +292,23 @@ public static function isMatch(string $pattern, string $subject, ?array &$matche
     }
 
     /**
+     * Variant of `isMatch()` which outputs non-null matches (or throws)
+     *
+     * @param non-empty-string $pattern
+     * @param array<string> $matches Set by method
+     * @param int-mask<PREG_UNMATCHED_AS_NULL> $flags PREG_UNMATCHED_AS_NULL is always set, no other flags are supported
+     * @throws UnexpectedNullMatchException
+     *
+     * @param-out array<int|string, string> $matches
+     */
+    public static function isMatchStrictGroups(string $pattern, string $subject, ?array &$matches = null, int $flags = 0, int $offset = 0): bool
+    {
+        return (bool) self::matchStrictGroups($pattern, $subject, $matches, $flags, $offset);
+    }
+
+    /**
+     * Variant of matchAll() which returns a bool instead of int
+     *
      * @param non-empty-string   $pattern
      * @param array<int|string, list<string|null>> $matches Set by method
      * @param int-mask<PREG_UNMATCHED_AS_NULL> $flags PREG_UNMATCHED_AS_NULL is always set, no other flags are supported
@@ -265,6 +321,22 @@ public static function isMatchAll(string $pattern, string $subject, ?array &$mat
     }
 
     /**
+     * Variant of `isMatchAll()` which outputs non-null matches (or throws)
+     *
+     * @param non-empty-string $pattern
+     * @param array<int|string, list<string>> $matches Set by method
+     * @param int-mask<PREG_UNMATCHED_AS_NULL> $flags PREG_UNMATCHED_AS_NULL is always set, no other flags are supported
+     *
+     * @param-out array<int|string, list<string>> $matches
+     */
+    public static function isMatchAllStrictGroups(string $pattern, string $subject, ?array &$matches = null, int $flags = 0, int $offset = 0): bool
+    {
+        return (bool) self::matchAllStrictGroups($pattern, $subject, $matches, $flags, $offset);
+    }
+
+    /**
+     * Variant of matchWithOffsets() which returns a bool instead of int
+     *
      * Runs preg_match with PREG_OFFSET_CAPTURE
      *
      * @param non-empty-string   $pattern
@@ -279,6 +351,8 @@ public static function isMatchWithOffsets(string $pattern, string $subject, ?arr
     }
 
     /**
+     * Variant of matchAllWithOffsets() which returns a bool instead of int
+     *
      * Runs preg_match_all with PREG_OFFSET_CAPTURE
      *
      * @param non-empty-string   $pattern
@@ -298,4 +372,47 @@ private static function checkOffsetCapture(int $flags, string $useFunctionName):
             throw new \InvalidArgumentException('PREG_OFFSET_CAPTURE is not supported as it changes the type of $matches, use ' . $useFunctionName . '() instead');
         }
     }
+
+    private static function checkSetOrder(int $flags): void
+    {
+        if (($flags & PREG_SET_ORDER) !== 0) {
+            throw new \InvalidArgumentException('PREG_SET_ORDER is not supported as it changes the type of $matches');
+        }
+    }
+
+    /**
+     * @param array<int|string, string|null> $matches
+     * @return array<int|string, string>
+     * @throws UnexpectedNullMatchException
+     */
+    private static function enforceNonNullMatches(string $pattern, array $matches, string $variantMethod)
+    {
+        foreach ($matches as $group => $match) {
+            if (null === $match) {
+                throw new UnexpectedNullMatchException('Pattern "'.$pattern.'" had an unexpected unmatched group "'.$group.'", make sure the pattern always matches or use '.$variantMethod.'() instead.');
+            }
+        }
+
+        /** @var array<string> */
+        return $matches;
+    }
+
+    /**
+     * @param array<int|string, list<string|null>> $matches
+     * @return array<int|string, list<string>>
+     * @throws UnexpectedNullMatchException
+     */
+    private static function enforceNonNullMatchAll(string $pattern, array $matches, string $variantMethod)
+    {
+        foreach ($matches as $group => $groupMatches) {
+            foreach ($groupMatches as $match) {
+                if (null === $match) {
+                    throw new UnexpectedNullMatchException('Pattern "'.$pattern.'" had an unexpected unmatched group "'.$group.'", make sure the pattern always matches or use '.$variantMethod.'() instead.');
+                }
+            }
+        }
+
+        /** @var array<int|string, list<string>> */
+        return $matches;
+    }
 }