Infinite loop when parsing invalid HTML

[PWalkow]

I found a piece of invalid html that is causing infinite loop (and maximum execution time issues).
similiar issue was #179 that was closed already but this example shows that similiar cases exist, how we could fix it without having access to change the html?

<img src="www.example.com/image.jpg" data-caption="How did <i>

I understand it is an invalid html but the parser should not stuck when trying to parse it and fail gracefully rather than causing infinite execution time

[paquettg]

Thanks for the report.

I have added this to the release in 2.1.1 and, once I confirm the issue with a test, I'll see to a fix.

If you can provide a PR with a fix and test I would also appreciate that.

Thank you.

[EnriqueGF]

I'm having the same issue with a broken html.

[zrlhk]

I update to 2.1.1 and it's still Infinite loop

[tfedor]

@paquettg The issue is caused by parsing attribute where you don't check length of file; here: https://github.com/paquettg/php-html-parser/blob/master/src/PHPHtmlParser/Dom.php#L760 and here: https://github.com/paquettg/php-html-parser/blob/master/src/PHPHtmlParser/Dom.php#L772

Fix should be easy, something like:

} while (!empty($moreString) && $this->content->getPosition() < $this->content->getSize());

I'd push a PR, but adding this fix caused this strspn to return false and I'm not sure what's that supposed to do:
https://github.com/paquettg/php-html-parser/blob/master/src/PHPHtmlParser/Content.php#L237

I'd just fix that by throwing LogicalException, but gonna leave that up to you.

[PWalkow]

@paquettg
I understand that some of the things were fixed to stop infinite loops but an example that I had in my integration tests with this library is still failing for:
'<the thing broke itV'

[paquettg]

I'll try this out on the 3.0.0 dev branch later tonight