HTML attributes not highlighted if unquoted

[edent]

Tempest version

2.11.4

PHP version

8.4

Operating system

Linux

Description

Consider this HTML:

<img class="a" alt src=example.png>

This library only highlights the attribute "class".

&lt;<span style="color: #0000ff;">img</span> <span style="color: #795E26;">class</span>=&quot;a&quot; alt src=example.png&gt;

HTML only requires attribute values to be quoted if they contain a space, equals sign, or a few other characters. See https://html.spec.whatwg.org/multipage/syntax.html#attributes-2

So, in this case the empty alt and the unquoted src should both be highlighted.

Steps to reproduce

echo $highlighter->parse( '<img class="a" alt src=example.png>', "html" );

[brendt]

Thanks! Are you up for submitting a PR fix? Otherwise I'll try to look into it as soon as possible :)

[edent]

Changing this:

highlight/src/Languages/Xml/Patterns/XmlAttributePattern.php

Line 21 in 5a239a9

return '(?<match>[\w\-]+)="';

To make the quote optional:

return '(?<match>[\w\-]+)="?';

Does work for unquoted but not empty values, but may be invalid for XML.

However, as this isn't a validator, it's probably fine.

If you're happy with that, I'll send a PR and will think on how to tackle the empty value case.

[brendt]

Yes, that should work!

[brendt]

Did some testing, unfortunately it's not as simple as we thought: simply making " optional makes it so that our attribute-name match part can match too many things (anything with \w- character. We'd need to fine-tune to regex to make sure it's only matching words that actually are attributes (so within <> tags)