Stdlib: strip_tags() VM, 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 | 196 |
+| PHP files on vm.php path | 199 |
 | Source constructs flagged (blockers) | 10 |
-| Source constructs flagged (warnings) | 523 |
+| Source constructs flagged (warnings) | 527 |
 
 ## Compiler CFG gaps (`lib/Compiler.php`)
 
@@ -40,16 +40,18 @@ These `LogicException` messages indicate CFG ops or expressions not yet lowered:
 | `ext/standard/JitHtmlspecialchars.php` | 0 | 1 |
 | `ext/standard/JitImplode.php` | 0 | 1 |
 | `ext/standard/JitNumberFormat.php` | 0 | 1 |
+| `ext/standard/JitParseUrl.php` | 0 | 1 |
 | `ext/standard/JitPath.php` | 0 | 1 |
 | `ext/standard/JitRealpath.php` | 0 | 1 |
 | `ext/standard/JitStrPad.php` | 0 | 1 |
 | `ext/standard/JitStrRepeat.php` | 0 | 1 |
 | `ext/standard/JitStrReplace.php` | 0 | 1 |
 | `ext/standard/JitStringConcat.php` | 0 | 1 |
 | `ext/standard/JitStringIndex.php` | 0 | 1 |
+| `ext/standard/JitStripTags.php` | 0 | 1 |
 | `ext/standard/JitStrpos.php` | 0 | 1 |
 | `ext/standard/JitUrlencode.php` | 0 | 1 |
-| `ext/standard/Module.php` | 0 | 94 |
+| `ext/standard/Module.php` | 0 | 95 |
 | `ext/standard/VmFs.php` | 0 | 3 |
 | `ext/standard/VmNumberFormat.php` | 0 | 1 |
 | `ext/standard/VmString.php` | 0 | 2 |
@@ -135,6 +137,7 @@ These `LogicException` messages indicate CFG ops or expressions not yet lowered:
 | `ext/standard/string_ltrim.php` | 0 | 1 |
 | `ext/standard/string_rtrim.php` | 0 | 1 |
 | `ext/standard/string_trim.php` | 0 | 1 |
+| `ext/standard/strip_tags.php` | 0 | 1 |
 | `ext/standard/strncmp.php` | 0 | 1 |
 | `ext/standard/strpos.php` | 0 | 1 |
 | `ext/standard/strrev.php` | 0 | 1 |
@@ -265,6 +268,11 @@ These `LogicException` messages indicate CFG ops or expressions not yet lowered:
 **Warnings** (review for bootstrap subset):
 - 4 class method(s) — PHPCfg Op\Stmt\ClassMethod not lowered in Compiler
 
+### `ext/standard/JitParseUrl.php`
+
+**Warnings** (review for bootstrap subset):
+- 3 class method(s) — PHPCfg Op\Stmt\ClassMethod not lowered in Compiler
+
 ### `ext/standard/JitPath.php`
 
 **Warnings** (review for bootstrap subset):
@@ -300,6 +308,11 @@ These `LogicException` messages indicate CFG ops or expressions not yet lowered:
 **Warnings** (review for bootstrap subset):
 - 5 class method(s) — PHPCfg Op\Stmt\ClassMethod not lowered in Compiler
 
+### `ext/standard/JitStripTags.php`
+
+**Warnings** (review for bootstrap subset):
+- 3 class method(s) — PHPCfg Op\Stmt\ClassMethod not lowered in Compiler
+
 ### `ext/standard/JitStrpos.php`
 
 **Warnings** (review for bootstrap subset):
@@ -394,18 +407,19 @@ These `LogicException` messages indicate CFG ops or expressions not yet lowered:
 - new str_pad (line 100)
 - new str_split (line 101)
 - new htmlspecialchars (line 102)
-- new header_ (line 103)
-- new http_response_code (line 104)
-- new urlencode (line 105)
-- new rawurlencode (line 106)
-- new parse_url (line 107)
-- new dirname (line 108)
-- new basename (line 109)
-- new realpath (line 110)
-- new getenv_ (line 111)
-- new putenv_ (line 112)
-- new scandir (line 113)
-- new glob_ (line 114)
+- new strip_tags (line 103)
+- new header_ (line 104)
+- new http_response_code (line 105)
+- new urlencode (line 106)
+- new rawurlencode (line 107)
+- new parse_url (line 108)
+- new dirname (line 109)
+- new basename (line 110)
+- new realpath (line 111)
+- new getenv_ (line 112)
+- new putenv_ (line 113)
+- new scandir (line 114)
+- new glob_ (line 115)
 - 2 class method(s) — PHPCfg Op\Stmt\ClassMethod not lowered in Compiler
 
 ### `ext/standard/VmFs.php`
@@ -423,7 +437,7 @@ These `LogicException` messages indicate CFG ops or expressions not yet lowered:
 ### `ext/standard/VmString.php`
 
 **Warnings** (review for bootstrap subset):
-- 40 class method(s) — PHPCfg Op\Stmt\ClassMethod not lowered in Compiler
+- 45 class method(s) — PHPCfg Op\Stmt\ClassMethod not lowered in Compiler
 - 1 closure(s)
 
 ### `ext/standard/abs.php`
@@ -858,6 +872,11 @@ These `LogicException` messages indicate CFG ops or expressions not yet lowered:
 **Warnings** (review for bootstrap subset):
 - 5 class method(s) — PHPCfg Op\Stmt\ClassMethod not lowered in Compiler
 
+### `ext/standard/strip_tags.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

@@ -1,5 +1,6 @@
-Wrote /compiler/docs/capabilities.md (105 builtins).
-t/capability-matrix.php`. Do not edit by hand.
+# Builtin capability matrix
+
+Auto-generated by `script/capability-matrix.php`. Do not edit by hand.
 
 | Function | VM | JIT | AOT | Module | Notes |
 |----------|:--:|:---:|:---:|--------|-------|
@@ -96,6 +97,7 @@ t/capability-matrix.php`. Do not edit by hand.
 | `str_split` | yes | no | no | standard | doc: VM only; not implemented for JIT in this compiler build |
 | `str_starts_with` | yes | yes | yes | standard | AOT PHPT |
 | `strcmp` | yes | yes | yes | standard |  |
+| `strip_tags` | 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/JitStripTags.php

@@ -0,0 +1,78 @@
+<?php
+
+declare(strict_types=1);
+
+namespace PHPCompiler\ext\standard;
+
+use PHPCompiler\JIT\Context;
+use PHPCompiler\JIT\Variable as JITVariable;
+use PHPLLVM\Value;
+
+/**
+ * LLVM JIT/AOT helpers for strip_tags() (mirrors VmString::stripTags).
+ */
+final class JitStripTags
+{
+    public static function stripTags(Context $context, JITVariable $input, ?JITVariable $allowed = null): Value
+    {
+        $inputLiteral = $input->compileTimeString ?? null;
+        if (null !== $inputLiteral) {
+            $allowedLiteral = null;
+            if (null !== $allowed) {
+                if (JITVariable::TYPE_STRING === $allowed->type) {
+                    $allowedLiteral = $allowed->compileTimeString ?? '';
+                } elseif (JITVariable::TYPE_VALUE !== $allowed->type) {
+                    throw new \LogicException(
+                        'strip_tags() allowed_tags must be a string or null in this compiler build'
+                    );
+                }
+            }
+
+            return $context->builder->load(
+                $context->constantStringFromString(VmString::stripTags($inputLiteral, $allowedLiteral))
+            );
+        }
+
+        $inPtr = self::jitStringArg($context, $input);
+        $allowPtr = self::jitAllowedArg($context, $allowed);
+
+        return $context->builder->call(
+            $context->lookupFunction('__compiler_strip_tags'),
+            $inPtr,
+            $allowPtr
+        );
+    }
+
+    private static function jitStringArg(Context $context, JITVariable $arg): Value
+    {
+        if (JITVariable::TYPE_STRING === $arg->type) {
+            return $context->helper->loadValue($arg);
+        }
+        if (JITVariable::TYPE_VALUE === $arg->type) {
+            return $context->builder->call(
+                $context->lookupFunction('__value__readString'),
+                $arg->value
+            );
+        }
+
+        throw new \LogicException('strip_tags() only supports strings in this compiler build');
+    }
+
+    private static function jitAllowedArg(Context $context, ?JITVariable $allowed): Value
+    {
+        if (null === $allowed) {
+            return $context->builder->load($context->constantStringFromString(''));
+        }
+        if (JITVariable::TYPE_STRING === $allowed->type) {
+            return $context->helper->loadValue($allowed);
+        }
+        if (JITVariable::TYPE_VALUE === $allowed->type) {
+            return $context->builder->call(
+                $context->lookupFunction('__value__readString'),
+                $allowed->value
+            );
+        }
+
+        throw new \LogicException('strip_tags() allowed_tags must be a string or null in this compiler build');
+    }
+}

ext/standard/Module.php

@@ -100,6 +100,7 @@ public function getFunctions(): array
             new str_pad(),
             new str_split(),
             new htmlspecialchars(),
+            new strip_tags(),
             new header_(),
             new http_response_code(),
             new urlencode(),

ext/standard/VmString.php

@@ -353,6 +353,134 @@ public static function htmlspecialchars(
         return $out;
     }
 
+    /**
+     * strip_tags() subset: removes HTML/PHP tags; optional allow-list like "<b><p>".
+     * HTML comments and PHP tags remove their inner content; other tags keep inner text.
+     */
+    public static function stripTags(string $string, ?string $allowedTags = null): string
+    {
+        $allowed = null === $allowedTags || '' === $allowedTags
+            ? []
+            : self::parseAllowedTags($allowedTags);
+        $out = '';
+        $len = self::byteLength($string);
+        $i = 0;
+        while ($i < $len) {
+            $ch = $string[$i];
+            if ('<' !== $ch) {
+                $out .= $ch;
+                ++$i;
+                continue;
+            }
+            if ($i + 3 < $len && '<!--' === self::byteSlice($string, $i, 4)) {
+                $end = self::findSubstring($string, '-->', $i + 4);
+                if (false !== $end) {
+                    $i = $end + 3;
+                    continue;
+                }
+            }
+            if ($i + 1 < $len && '<?' === self::byteSlice($string, $i, 2)) {
+                $end = self::findSubstring($string, '?>', $i + 2);
+                if (false !== $end) {
+                    $i = $end + 2;
+                    continue;
+                }
+            }
+            $gt = self::findSubstring($string, '>', $i + 1);
+            if (false === $gt) {
+                $out .= $ch;
+                ++$i;
+                continue;
+            }
+            $tagContent = self::byteSlice($string, $i + 1, $gt - $i - 1);
+            $tagName = self::extractTagName($tagContent);
+            if (null !== $tagName && [] !== $allowed && self::isTagAllowed($tagName, $allowed)) {
+                $out .= self::byteSlice($string, $i, $gt - $i + 1);
+            }
+            $i = $gt + 1;
+        }
+
+        return $out;
+    }
+
+    /**
+     * @return list<string>
+     */
+    private static function parseAllowedTags(string $allowedTags): array
+    {
+        $tags = [];
+        $len = self::byteLength($allowedTags);
+        $i = 0;
+        while ($i < $len) {
+            if ('<' !== $allowedTags[$i]) {
+                ++$i;
+                continue;
+            }
+            $gt = self::findSubstring($allowedTags, '>', $i + 1);
+            if (false === $gt) {
+                break;
+            }
+            $name = self::extractTagName(self::byteSlice($allowedTags, $i + 1, $gt - $i - 1));
+            if (null !== $name && '' !== $name) {
+                $tags[] = $name;
+            }
+            $i = $gt + 1;
+        }
+
+        return $tags;
+    }
+
+    private static function extractTagName(string $tagContent): ?string
+    {
+        $len = self::byteLength($tagContent);
+        $i = 0;
+        while ($i < $len && self::isTagWhitespace($tagContent[$i])) {
+            ++$i;
+        }
+        if ($i < $len && '/' === $tagContent[$i]) {
+            ++$i;
+        }
+        if ($i >= $len) {
+            return null;
+        }
+        $start = $i;
+        while ($i < $len) {
+            $ch = $tagContent[$i];
+            if (self::isTagWhitespace($ch) || '>' === $ch || '/' === $ch) {
+                break;
+            }
+            if (!ctype_alpha($ch) && !ctype_digit($ch)) {
+                return null;
+            }
+            ++$i;
+        }
+        if ($start === $i) {
+            return null;
+        }
+
+        return strtolower(self::byteSlice($tagContent, $start, $i - $start));
+    }
+
+    /**
+     * @param list<string> $allowed
+     */
+    private static function isTagAllowed(string $tagName, array $allowed): bool
+    {
+        $tagName = strtolower($tagName);
+        foreach ($allowed as $name) {
+            if ($tagName === $name) {
+                return true;
+            }
+        }
+
+        return false;
+    }
+
+    private static function isTagWhitespace(string $ch): bool
+    {
+        return str_contains(self::TRIM_DEFAULT, $ch);
+    }
+
     /**
      * @return list<string>
      */