fix: independant HTML implementation

src/Languages/Html/HtmlLanguage.php

@@ -5,14 +5,18 @@
 namespace Tempest\Highlight\Languages\Html;
 
 use Override;
+use Tempest\Highlight\Languages\Base\BaseLanguage;
 use Tempest\Highlight\Languages\Html\Injections\CssAttributeInHtmlInjection;
 use Tempest\Highlight\Languages\Html\Injections\CssInHtmlInjection;
 use Tempest\Highlight\Languages\Html\Injections\JavaScriptInHtmlInjection;
 use Tempest\Highlight\Languages\Html\Injections\PhpInHtmlInjection;
 use Tempest\Highlight\Languages\Html\Injections\PhpShortEchoInHtmlInjection;
-use Tempest\Highlight\Languages\Xml\XmlLanguage;
+use Tempest\Highlight\Languages\Html\Patterns\HtmlAttributePattern;
+use Tempest\Highlight\Languages\Html\Patterns\HtmlCloseTagPattern;
+use Tempest\Highlight\Languages\Html\Patterns\HtmlOpenTagPattern;
+use Tempest\Highlight\Languages\Xml\Patterns\XmlCommentPattern;
 
-class HtmlLanguage extends XmlLanguage
+class HtmlLanguage extends BaseLanguage
 {
     #[Override]
     public function getName(): string
@@ -38,6 +42,10 @@ public function getPatterns(): array
     {
         return [
             ...parent::getPatterns(),
+            new HtmlOpenTagPattern(),
+            new HtmlCloseTagPattern(),
+            new HtmlAttributePattern(),
+            new XmlCommentPattern(),
         ];
     }
 }

src/Languages/Html/Patterns/HtmlAttributePattern.php

@@ -0,0 +1,47 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Html\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: '<x-hello attr="">', output: 'attr')]
+#[PatternTest(input: '<a href="">', output: 'href')]
+#[PatternTest(input: '<a data-type="">', output: 'data-type')]
+#[PatternTest(input: '<x-post :foreach="$this->posts as $post">', output: ':foreach')]
+#[PatternTest(input: '<xsl xmlns:xsl="http">', output: 'xmlns:xsl')]
+#[PatternTest(input: "<item attr='value'>", output: 'attr')]
+#[PatternTest(input: "<item simple=value>", output: 'simple')]
+#[PatternTest(input: "<item with-hyphen=simple-with-hyphen>", output: 'with-hyphen')]
+#[PatternTest(input: "<div class=''><span style=''></span></div>", output: ['class', 'style'])]
+#[PatternTest(input: '<item
+    id =
+    "multiline-attr">', output: 'id')]
+#[PatternTest(input: '<item
+    type
+    =
+    "multiline-attr">', output: 'type')]
+#[PatternTest(input: '<item
+    a
+    ="multiline-attr">', output: 'a')]
+#[PatternTest(input: '<p></p>', output: null)]
+# Not yet implemented
+# #[PatternTest(input: "<script defer>", output: 'defer')]
+final readonly class HtmlAttributePattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return '(?<match>[\w\-:]+)\s*=\s*(?:["\']|[\w-]+)';
+    }
+
+    public function getTokenType(): TokenTypeEnum
+    {
+        return TokenTypeEnum::PROPERTY;
+    }
+}

src/Languages/Html/Patterns/HtmlCloseTagPattern.php

@@ -0,0 +1,37 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Html\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: '</simple>', output: 'simple')]
+#[PatternTest(input: '</UPPERCASE>', output: 'UPPERCASE')]
+#[PatternTest(input: '</CamelCase>', output: 'CamelCase')]
+#[PatternTest(input: '</x-hello>', output: 'x-hello')]
+#[PatternTest(input: '</a>', output: 'a')]
+# The following are not HTML valid
+#[PatternTest(input: '</ns:tag>', output: null)]
+#[PatternTest(input: '</point.x>', output: null)]
+#[PatternTest(input: '</point_y>', output: null)]
+#[PatternTest(input: '</_private>', output: null)]
+#[PatternTest(input: '</1tag>', output: null)]
+#[PatternTest(input: '</-tag>', output: null)]
+final readonly class HtmlCloseTagPattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return '<\/(?<match>[a-zA-Z][a-zA-Z0-9\-]*(?![:_\.\w]))';
+    }
+
+    public function getTokenType(): TokenTypeEnum
+    {
+        return TokenTypeEnum::KEYWORD;
+    }
+}

src/Languages/Html/Patterns/HtmlOpenTagPattern.php

@@ -0,0 +1,37 @@
+<?php
+
+declare(strict_types=1);
+
+namespace Tempest\Highlight\Languages\Html\Patterns;
+
+use Tempest\Highlight\IsPattern;
+use Tempest\Highlight\Pattern;
+use Tempest\Highlight\PatternTest;
+use Tempest\Highlight\Tokens\TokenTypeEnum;
+
+#[PatternTest(input: '<div>', output: 'div')]
+#[PatternTest(input: '<UPPERCASE>', output: 'UPPERCASE')]
+#[PatternTest(input: '<CamelCase>', output: 'CamelCase')]
+#[PatternTest(input: '<a-component>', output: 'a-component')]
+#[PatternTest(input: '<a href="">', output: 'a')]
+# The following are not HTML valid
+#[PatternTest(input: '<_private>', output: null)]
+#[PatternTest(input: '<ns:tag>', output: null)]
+#[PatternTest(input: '<point.x>', output: null)]
+#[PatternTest(input: '<point_y>', output: null)]
+#[PatternTest(input: '<1tag>', output: null)]
+#[PatternTest(input: '<-tag>', output: null)]
+final readonly class HtmlOpenTagPattern implements Pattern
+{
+    use IsPattern;
+
+    public function getPattern(): string
+    {
+        return '<(?<match>[a-zA-Z][a-zA-Z0-9\-]*(?![:_\.\w]))';
+    }
+
+    public function getTokenType(): TokenTypeEnum
+    {
+        return TokenTypeEnum::KEYWORD;
+    }
+}

src/Languages/Xml/Patterns/XmlAttributePattern.php

@@ -13,13 +13,24 @@
 #[PatternTest(input: '<a href="">', output: 'href')]
 #[PatternTest(input: '<a data-type="">', output: 'data-type')]
 #[PatternTest(input: '<xsl xmlns:xsl="http">', output: 'xmlns:xsl')]
+#[PatternTest(input: "<item attr='value'>", output: 'attr')]
+#[PatternTest(input: '<item
+    id =
+    "multiline-attr">', output: 'id')]
+#[PatternTest(input: '<item
+    type
+    =
+    "multiline-attr">', output: 'type')]
+#[PatternTest(input: '<item
+    a
+    ="multiline-attr">', output: 'a')]
 final readonly class XmlAttributePattern implements Pattern
 {
     use IsPattern;
 
     public function getPattern(): string
     {
-        return '(?<match>[\w\-\:]+)="';
+        return '(?<match>[\w\-\:]+)\s*=\s*["\']';
     }
 
     public function getTokenType(): TokenTypeEnum

src/Languages/Xml/Patterns/XmlCloseTagPattern.php

@@ -9,16 +9,25 @@
 use Tempest\Highlight\PatternTest;
 use Tempest\Highlight\Tokens\TokenTypeEnum;
 
+#[PatternTest(input: '</simple>', output: 'simple')]
+#[PatternTest(input: '</UPPERCASE>', output: 'UPPERCASE')]
+#[PatternTest(input: '</CamelCase>', output: 'CamelCase')]
 #[PatternTest(input: '</x-hello>', output: 'x-hello')]
 #[PatternTest(input: '</a>', output: 'a')]
 #[PatternTest(input: '</ns:tag>', output: 'ns:tag')]
+#[PatternTest(input: '</point.x>', output: 'point.x')]
+#[PatternTest(input: '</point_y>', output: 'point_y')]
+#[PatternTest(input: '</_private>', output: '_private')]
+# The following are not valid XML tags
+#[PatternTest(input: '</1tag>', output: null)]
+#[PatternTest(input: '</-tag>', output: null)]
 final readonly class XmlCloseTagPattern implements Pattern
 {
     use IsPattern;
 
     public function getPattern(): string
     {
-        return '<\/(?<match>[\w\-\:]+)';
+        return '<\/(?<match>[a-zA-Z_][\w\-\:\.]*)';
     }
 
     public function getTokenType(): TokenTypeEnum

src/Languages/Xml/Patterns/XmlOpenTagPattern.php

@@ -9,17 +9,26 @@
 use Tempest\Highlight\PatternTest;
 use Tempest\Highlight\Tokens\TokenTypeEnum;
 
+#[PatternTest(input: '<simple>', output: 'simple')]
+#[PatternTest(input: '<UPPERCASE>', output: 'UPPERCASE')]
+#[PatternTest(input: '<CamelCase>', output: 'CamelCase')]
 #[PatternTest(input: '<x-hello attr="">', output: 'x-hello')]
 #[PatternTest(input: '<a href="">', output: 'a')]
 #[PatternTest(input: '<br/>', output: 'br')]
 #[PatternTest(input: '<ns:tag>', output: 'ns:tag')]
+#[PatternTest(input: '<point.x>', output: 'point.x')]
+#[PatternTest(input: '<point_y>', output: 'point_y')]
+#[PatternTest(input: '<_private>', output: '_private')]
+# The following are not valid XML tags
+#[PatternTest(input: '<1tag>', output: null)]
+#[PatternTest(input: '<-tag>', output: null)]
 final readonly class XmlOpenTagPattern implements Pattern
 {
     use IsPattern;
 
     public function getPattern(): string
     {
-        return '<(?<match>[\w\-\:]+)';
+        return '<(?<match>[a-zA-Z_][\w\-\:\.]*)';
     }
 
     public function getTokenType(): TokenTypeEnum

tests/targets/blade.md

@@ -0,0 +1,18 @@
+```blade
+<!-- resources/views/layouts/app.blade.php -->
+
+<html>
+    <head>
+        <title>App Name - @yield('title')</title>
+    </head>
+    <body>
+        @section('sidebar')
+            This is the master sidebar.
+        @show
+
+        <div class="container">
+            @yield('content')
+        </div>
+    </body>
+</html>
+```

tests/targets/tempest.md

@@ -0,0 +1,25 @@
+```blade
+<x-base title="Home">
+    <x-post :foreach="$this->posts as $post">
+        {{-- a comment which won't be rendered to HTML --}}
+
+        {!! $post->title !!}
+
+        <span :if="$this->showDate($post)">
+            {{ $post->date }}
+        </span>
+        <span :else>
+            -
+        </span>
+    </x-post>
+    <div :forelse>
+        <form action="…">
+            <x-csrf-token />
+            <math-α></math-α>
+            <emotion-😍></emotion-😍>
+        </form>
+    </div>
+
+    <x-footer />
+</x-base>
+```

tests/targets/twig.md

@@ -0,0 +1,18 @@
+```twig
+<!DOCTYPE html>
+<html>
+    <head>
+        <title>My Webpage</title>
+    </head>
+    <body>
+        <ul id="navigation">
+        {% for item in navigation %}
+            <li><a href="{{ item.href }}">{{ item.caption }}</a></li>
+        {% endfor %}
+        </ul>
+
+        <h1>My Webpage</h1>
+        {{ a_variable }}
+    </body>
+</html>
+```

tests/targets/xml.md

@@ -51,6 +51,12 @@
     </b:nested>
   </a:element>
 
+  <!-- Default namespace override and undeclaration -->
+  <coordinates>
+      <point.x>123.45</point.x>
+      <point.y>42,15</point.y>
+  </coordinates>
+
   <!-- Default namespace override and undeclaration -->
   <container xmlns="http://example.com/override">
     <item xmlns="">Back to no namespace</item>