Stdlib: stripos() native LLVM/JIT and AOT

docs/bootstrap-inventory.md

@@ -8,9 +8,9 @@ Regenerate: `php script/bootstrap-inventory.php`
 
 | Metric | Count |
 |--------|------:|
-| PHP files on vm.php path | 238 |
+| PHP files on vm.php path | 239 |
 | Source constructs flagged (blockers) | 10 |
-| Source constructs flagged (warnings) | 626 |
+| Source constructs flagged (warnings) | 628 |
 
 ## Compiler CFG gaps (`lib/Compiler.php`)
 
@@ -58,7 +58,7 @@ These `LogicException` messages indicate CFG ops or expressions not yet lowered:
 | `ext/standard/JitStripTags.php` | 0 | 1 |
 | `ext/standard/JitStrpos.php` | 0 | 1 |
 | `ext/standard/JitUrlencode.php` | 0 | 1 |
-| `ext/standard/Module.php` | 0 | 109 |
+| `ext/standard/Module.php` | 0 | 110 |
 | `ext/standard/VmDate.php` | 0 | 1 |
 | `ext/standard/VmExit.php` | 0 | 2 |
 | `ext/standard/VmFs.php` | 0 | 3 |
@@ -159,6 +159,7 @@ These `LogicException` messages indicate CFG ops or expressions not yet lowered:
 | `ext/standard/string_rtrim.php` | 0 | 1 |
 | `ext/standard/string_trim.php` | 0 | 1 |
 | `ext/standard/strip_tags.php` | 0 | 1 |
+| `ext/standard/stripos.php` | 0 | 1 |
 | `ext/standard/strncmp.php` | 0 | 1 |
 | `ext/standard/strpos.php` | 0 | 1 |
 | `ext/standard/strrev.php` | 0 | 1 |
@@ -449,64 +450,65 @@ These `LogicException` messages indicate CFG ops or expressions not yet lowered:
 - new substr (line 69)
 - new strrev (line 70)
 - new strpos (line 71)
-- new str_contains (line 72)
-- new str_starts_with (line 73)
-- new str_ends_with (line 74)
-- new strncmp (line 75)
-- new array_count (line 76)
+- new stripos (line 72)
+- new str_contains (line 73)
+- new str_starts_with (line 74)
+- new str_ends_with (line 75)
+- new strncmp (line 76)
 - new array_count (line 77)
-- new array_key_exists (line 78)
-- new in_array (line 79)
-- new array_push (line 80)
-- new array_pop (line 81)
-- new array_shift (line 82)
-- new sort_ (line 83)
-- new array_values (line 84)
-- new array_keys (line 85)
-- new array_merge (line 86)
-- new array_slice (line 87)
-- new explode (line 88)
-- new implode (line 89)
-- new str_replace (line 90)
-- new nl2br (line 91)
-- new array_reverse (line 92)
-- new array_search (line 93)
-- new array_sum (line 94)
-- new array_flip (line 95)
-- new array_unique (line 96)
-- new array_fill (line 97)
-- new array_combine (line 98)
-- new range (line 99)
-- new bin2hex (line 100)
-- new hex2bin (line 101)
-- new random_bytes (line 102)
-- new str_pad (line 103)
-- new str_split (line 104)
-- new htmlspecialchars (line 105)
-- new strip_tags (line 106)
-- new header_ (line 107)
-- new header_remove (line 108)
-- new header_list (line 109)
-- new getallheaders_ (line 110)
-- new http_response_code (line 111)
-- new urlencode (line 112)
-- new rawurlencode (line 113)
-- new urldecode (line 114)
-- new rawurldecode (line 115)
-- new parse_url (line 116)
-- new dirname (line 117)
-- new basename (line 118)
-- new realpath (line 119)
-- new file_get_contents (line 120)
-- new getenv_ (line 121)
-- new putenv_ (line 122)
-- new extract_ (line 123)
-- new compact_ (line 124)
-- new scandir (line 125)
-- new glob_ (line 126)
-- new time (line 127)
-- new date (line 128)
-- new gmdate (line 129)
+- new array_count (line 78)
+- new array_key_exists (line 79)
+- new in_array (line 80)
+- new array_push (line 81)
+- new array_pop (line 82)
+- new array_shift (line 83)
+- new sort_ (line 84)
+- new array_values (line 85)
+- new array_keys (line 86)
+- new array_merge (line 87)
+- new array_slice (line 88)
+- new explode (line 89)
+- new implode (line 90)
+- new str_replace (line 91)
+- new nl2br (line 92)
+- new array_reverse (line 93)
+- new array_search (line 94)
+- new array_sum (line 95)
+- new array_flip (line 96)
+- new array_unique (line 97)
+- new array_fill (line 98)
+- new array_combine (line 99)
+- new range (line 100)
+- new bin2hex (line 101)
+- new hex2bin (line 102)
+- new random_bytes (line 103)
+- new str_pad (line 104)
+- new str_split (line 105)
+- new htmlspecialchars (line 106)
+- new strip_tags (line 107)
+- new header_ (line 108)
+- new header_remove (line 109)
+- new header_list (line 110)
+- new getallheaders_ (line 111)
+- new http_response_code (line 112)
+- new urlencode (line 113)
+- new rawurlencode (line 114)
+- new urldecode (line 115)
+- new rawurldecode (line 116)
+- new parse_url (line 117)
+- new dirname (line 118)
+- new basename (line 119)
+- new realpath (line 120)
+- new file_get_contents (line 121)
+- new getenv_ (line 122)
+- new putenv_ (line 123)
+- new extract_ (line 124)
+- new compact_ (line 125)
+- new scandir (line 126)
+- new glob_ (line 127)
+- new time (line 128)
+- new date (line 129)
+- new gmdate (line 130)
 - 2 class method(s) — PHPCfg Op\Stmt\ClassMethod not lowered in Compiler
 
 ### `ext/standard/VmDate.php`
@@ -544,7 +546,7 @@ These `LogicException` messages indicate CFG ops or expressions not yet lowered:
 **Warnings** (review for bootstrap subset):
 - new Exception (line 182)
 - new Exception (line 190)
-- 52 class method(s) — PHPCfg Op\Stmt\ClassMethod not lowered in Compiler
+- 56 class method(s) — PHPCfg Op\Stmt\ClassMethod not lowered in Compiler
 - 1 closure(s)
 
 ### `ext/standard/abs.php`
@@ -1041,6 +1043,11 @@ These `LogicException` messages indicate CFG ops or expressions not yet lowered:
 **Warnings** (review for bootstrap subset):
 - 2 class method(s) — PHPCfg Op\Stmt\ClassMethod not lowered in Compiler
 
+### `ext/standard/stripos.php`
+
+**Warnings** (review for bootstrap subset):
+- 2 class method(s) — PHPCfg Op\Stmt\ClassMethod not lowered in Compiler
+
 ### `ext/standard/strncmp.php`
 
 **Warnings** (review for bootstrap subset):

docs/capabilities.md

@@ -109,6 +109,7 @@ Auto-generated by `script/capability-matrix.php`. Do not edit by hand.
 | `str_starts_with` | yes | yes | yes | standard | AOT PHPT |
 | `strcmp` | yes | yes | yes | standard |  |
 | `strip_tags` | yes | yes | yes | standard | JIT PHPT; AOT PHPT |
+| `stripos` | yes | yes | yes | standard | JIT PHPT; AOT PHPT |
 | `strlen` | yes | yes | yes | types | JIT PHPT; AOT PHPT |
 | `strncmp` | yes | yes | yes | standard |  |
 | `strpos` | yes | yes | yes | standard | JIT PHPT; AOT PHPT |

ext/standard/JitStrpos.php

@@ -23,7 +23,8 @@ public static function find(
         Context $context,
         Value $haystack,
         Value $needle,
-        ?Value $offset = null
+        ?Value $offset = null,
+        bool $caseInsensitive = false
     ): Value {
         $map = $context->structFieldMap['__string__'];
         $hayLen = $context->builder->load(
@@ -39,8 +40,9 @@ public static function find(
             $searchPtr = $context->builder->inBoundsGEP($hayPtr, $clamped);
         }
 
+        $searchFn = $caseInsensitive ? 'strcasestr' : 'strstr';
         $found = $context->builder->call(
-            $context->lookupFunction('strstr'),
+            $context->lookupFunction($searchFn),
             $searchPtr,
             $needlePtr
         );

ext/standard/Module.php

@@ -69,6 +69,7 @@ public function getFunctions(): array
             new substr(),
             new strrev(),
             new strpos(),
+            new stripos(),
             new str_contains(),
             new str_starts_with(),
             new str_ends_with(),
@@ -159,6 +160,14 @@ public function jitInit(JIT\Context $context): void
             $fn = $context->module->addFunction('strstr', $ft);
             $context->registerFunction('strstr', $fn);
         }
+        try {
+            $context->lookupFunction('strcasestr');
+        } catch (\Throwable $e) {
+            $i8p = $context->getTypeFromString('int8*');
+            $ft = $context->context->functionType($i8p, false, $i8p, $i8p);
+            $fn = $context->module->addFunction('strcasestr', $ft);
+            $context->registerFunction('strcasestr', $fn);
+        }
         try {
             $context->lookupFunction('strtol');
         } catch (\Throwable $e) {

ext/standard/VmString.php

@@ -799,6 +799,22 @@ public static function strpos(string $haystack, string $needle, int $offset = 0)
         return false === $pos ? false : $pos;
     }
 
+    /**
+     * @return int|false
+     */
+    public static function stripos(string $haystack, string $needle, int $offset = 0)
+    {
+        if ('' === $needle) {
+            throw new \LogicException('stripos(): Argument #2 ($needle) cannot be empty');
+        }
+        if ($offset < 0) {
+            $offset = 0;
+        }
+        $pos = self::findSubstringCaseInsensitive($haystack, $needle, $offset);
+
+        return false === $pos ? false : $pos;
+    }
+
     public static function startsWith(string $haystack, string $needle): bool
     {
         $nlen = self::byteLength($needle);
@@ -850,6 +866,27 @@ private static function compareBytes(string $haystack, string $needle, int $leng
         return true;
     }
 
+    private static function compareBytesCaseInsensitive(string $haystack, string $needle, int $length, int $hayOffset = 0): bool
+    {
+        for ($i = 0; $i < $length; ++$i) {
+            if (self::asciiLowerByte($haystack[$hayOffset + $i]) !== self::asciiLowerByte($needle[$i])) {
+                return false;
+            }
+        }
+
+        return true;
+    }
+
+    private static function asciiLowerByte(string $byte): string
+    {
+        $ord = self::byteOrd($byte);
+        if ($ord >= 65 && $ord <= 90) {
+            return self::byteChr($ord + 32);
+        }
+
+        return $byte;
+    }
+
     /**
      * @return int|false
      */
@@ -873,6 +910,29 @@ private static function findSubstring(string $haystack, string $needle, int $off
         return false;
     }
 
+    /**
+     * @return int|false
+     */
+    private static function findSubstringCaseInsensitive(string $haystack, string $needle, int $offset)
+    {
+        $hayLen = self::byteLength($haystack);
+        $needleLen = self::byteLength($needle);
+        if (0 === $needleLen) {
+            return false;
+        }
+        if ($offset >= $hayLen) {
+            return false;
+        }
+        $limit = $hayLen - $needleLen;
+        for ($i = $offset; $i <= $limit; ++$i) {
+            if (self::compareBytesCaseInsensitive($haystack, $needle, $needleLen, $i)) {
+                return $i;
+            }
+        }
+
+        return false;
+    }
+
     private static function asciiCaseTransform(string $string, bool $toLower): string
     {
         $out = '';

ext/standard/stripos.php

@@ -0,0 +1,73 @@
+<?php
+
+declare(strict_types=1);
+
+namespace PHPCompiler\ext\standard;
+
+use PHPCompiler\Frame;
+use PHPCompiler\Func\Internal;
+use PHPCompiler\JIT\Context;
+use PHPCompiler\JIT\Variable as JITVariable;
+use PHPCompiler\VM\Variable;
+use PHPLLVM\Value;
+
+/**
+ * stripos() for two strings (subset of PHP; ASCII case fold, non-empty needle, no offset in JIT).
+ */
+final class stripos extends Internal
+{
+    public function execute(Frame $frame): void
+    {
+        $argc = count($frame->calledArgs);
+        if ($argc < 2 || $argc > 3) {
+            throw new \LogicException('stripos() requires two or three arguments');
+        }
+        $haystack = $frame->calledArgs[0]->resolveIndirect();
+        $needle = $frame->calledArgs[1]->resolveIndirect();
+        if (null === $frame->returnVar) {
+            return;
+        }
+        if (Variable::TYPE_STRING !== $haystack->type || Variable::TYPE_STRING !== $needle->type) {
+            throw new \LogicException('stripos() only supports strings in this compiler build');
+        }
+        $offset = 0;
+        if (3 === $argc) {
+            $offVar = $frame->calledArgs[2]->resolveIndirect();
+            if (Variable::TYPE_INTEGER !== $offVar->type) {
+                throw new \LogicException('stripos() offset must be an integer in this compiler build');
+            }
+            $offset = $offVar->toInt();
+        }
+        $result = VmString::stripos($haystack->toString(), $needle->toString(), $offset);
+        if (false === $result) {
+            $frame->returnVar->bool(false);
+        } else {
+            $frame->returnVar->int($result);
+        }
+    }
+
+    public Context $context;
+
+    public function call(Context $context, JITVariable ...$args): Value
+    {
+        $this->context = $context;
+        $argc = count($args);
+        if ($argc < 2 || $argc > 3) {
+            throw new \LogicException('stripos() requires two or three arguments');
+        }
+        if (JITVariable::TYPE_STRING !== $args[0]->type || JITVariable::TYPE_STRING !== $args[1]->type) {
+            throw new \LogicException('stripos() only supports strings in this compiler build');
+        }
+        if (3 === $argc && JITVariable::TYPE_NATIVE_LONG !== $args[2]->type) {
+            throw new \LogicException('stripos() offset must be an integer in this compiler build');
+        }
+
+        $hay = $context->helper->loadValue($args[0]);
+        $needle = $context->helper->loadValue($args[1]);
+        $offset = 3 === $argc
+            ? $context->builder->truncOrBitCast($context->helper->loadValue($args[2]), $context->getTypeFromString('int64'))
+            : null;
+
+        return JitStrpos::find($context, $hay, $needle, $offset, true);
+    }
+}

test/compliance/cases/stdlib/stripos.phpt

@@ -0,0 +1,11 @@
+--TEST--
+stdlib stripos()
+--FILE--
+<?php
+echo stripos('Hello', 'LL'), "\n";
+echo stripos('Hello', 'x') == false ? 'y' : 'n', "\n";
+echo stripos('Hello', 'L', 3), "\n";
+--EXPECT--
+2
+y
+3