refactor(web): use parsedown as markdown renderer#152
refactor(web): use parsedown as markdown renderer#152renbaoshuo wants to merge 1 commit intomasterfrom
Conversation
There was a problem hiding this comment.
Pull Request Overview
This PR refactors the web application to replace the V8JS-based markdown renderer with Parsedown as the markdown rendering engine. This change eliminates the dependency on the V8 JavaScript engine and its PHP extension.
Key changes:
- Removes V8JS dependency from installation and Docker setup
- Implements a new UOJMarkdown class extending ParsedownMath with custom features
- Updates blog editor to use Parsedown instead of V8JS with marked.js
Reviewed Changes
Copilot reviewed 5 out of 7 changed files in this pull request and generated 3 comments.
Show a summary per file
| File | Description |
|---|---|
| web/install.sh | Removes V8JS installation and configuration from setup script |
| web/app/models/UOJMarkdown.php | Adds new markdown parser class with user mentions and table styling |
| web/app/models/UOJBlogEditor.php | Replaces V8JS markdown processing with Parsedown implementation |
| web/app/models/HTML.php | Adds factory method for creating UOJMarkdown instances |
| web/Dockerfile | Removes V8JS installation from Docker build process |
Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.
| @@ -0,0 +1,80 @@ | |||
| <?php | |||
|
|
|||
| require_once __DIR__ . '../vendor/parsedown/ParsedownMath.php'; | |||
There was a problem hiding this comment.
The path concatenation is incorrect. There's a missing directory separator between __DIR__ and the relative path. It should be __DIR__ . '/../vendor/parsedown/ParsedownMath.php'.
| require_once __DIR__ . '../vendor/parsedown/ParsedownMath.php'; | |
| require_once __DIR__ . '/../vendor/parsedown/ParsedownMath.php'; |
| $marked = function($md) use ($parsedown, $purifier) { | ||
| $dom = new DOMDocument; | ||
| $dom->loadHTML(mb_convert_encoding($parsedown->text($md), 'HTML-ENTITIES', 'UTF-8')); | ||
| $elements = $dom->getElementsByTagName('li'); | ||
|
|
||
| foreach ($elements as $element) { | ||
| $element->setAttribute( | ||
| 'class', | ||
| $element->getAttribute('class') . ' fragment' | ||
| ); | ||
| } | ||
|
|
||
| return $purifier->purify($dom->saveHTML()); | ||
| }; |
There was a problem hiding this comment.
The loadHTML method may generate warnings for malformed HTML. Additionally, saveHTML() returns the entire HTML document including <html> and <body> tags, which is likely not desired. Consider using libxml_use_internal_errors(true) to suppress warnings and saveHTML($dom->documentElement) or extracting only the body content.
| $element->setAttribute( | ||
| 'class', | ||
| $element->getAttribute('class') . ' fragment' |
There was a problem hiding this comment.
This will create a leading space in the class attribute if the element has no existing class, resulting in a class value like ' fragment'. Consider checking if the existing class is empty or trimming the result.
| $element->setAttribute( | |
| 'class', | |
| $element->getAttribute('class') . ' fragment' | |
| $current_class = $element->getAttribute('class'); | |
| $element->setAttribute( | |
| 'class', | |
| ($current_class === '' ? 'fragment' : $current_class . ' fragment') |
只是移动了代码，尚未测试，暂时置于 Draft 状态。
代码来自 https://github.com/renbaoshuo/S2OJ