Fix problem with Document class and html() function

[dustingrofthenumber]

Resolves #18

[Log1x]

Appreciate the PR! Seeing an initial issue with the Regex for comments. Can we double check these handle every condition?

[davideprevosto]

I encountered the same problem; the code in the PR seems to solve my problem. Have you got any update on this?
Thank you!

[QWp6t]

I believe you should be able to resolve this without regex just by using LIBXML_NOXMLDECL.

You should also be able to suppress errors using a flag as well.

I haven't tested this, but it should theoretically work as a drop-in replacement for this class cc @Log1x @retlehs

class Document
{
    /**
     * The DOMDocument instance.
     */
    protected DOMDocument $document;

    /**
     * Initialize the Document instance.
     */
    public function __construct(string $html)
    {
        $this->document = new DOMDocument(encoding: 'UTF-8');

        $this->document->loadHTML($html,
            LIBXML_HTML_NOIMPLIED | LIBXML_HTML_NODEFDTD | LIBXML_NOXMLDECL | LIBXML_NOWARNING | LIBXML_NOERROR
        );
    }

    /**
     * Make a new instance of the Document.
     */
    public static function make(string $html): self
    {
        return new static($html);
    }

    /**
     * Loop through each node in the document and execute the provided callback.
     */
    public function each(callable $callback): self
    {
        foreach ($this->xpath('//*') as $node) {
            $callback($node);
        }

        return $this;
    }

    /**
     * Evaluate the given XPath expression.
     */
    public function xpath(string $expression): DOMNodeList
    {
        return (new DOMXPath($this->document))->query($expression);
    }

    /**
     * Get the saved document HTML.
     */
    public function html(): string
    {
        return trim($this->document->saveHTML());
    }

    /**
     * Call the given method on the root document.
     */
    public function __call(string $name, array $arguments): mixed
    {
        return $this->document->{$name}(...$arguments);
    }

    /**
     * Get the given property from the root document.
     */
    public function __get(string $name): mixed
    {
        return $this->document->{$name};
    }
}

[davideprevosto]

I believe you should be able to resolve this without regex just by using LIBXML_NOXMLDECL.

@QWp6t I've tried your code and it's fixing the issue for me. I've also removed suppress() and clear() using the suggested flags.

[KristoferN]

Same problem on macOS with Herd PHP 8.3. libxml2 Version => 2.15.1

[dustingrofthenumber]

The method @QWp6t mentioned does work for me also. @Log1x should I switch approaches and commit it instead?

[Log1x]

@Log1x should I switch approaches and commit it instead?

Sounds good to me! Appreciate it.

[dustingrofthenumber]

Done! @Log1x I hope this does it :)

[davideprevosto]

Not sure @dustingrofthenumber, but I think you should remove:

use Illuminate\Support\Str;

[dustingrofthenumber]

Yep, you are right, it is unused, removed!

[QWp6t]

LGTM!