Smart Punctuation converts double dash in url (e.g. youtube video id) to en-dash

[svenignor]

Version(s) affected

2.7.0

Description

We have an md article with multiple youtube videos embedded. one of the videos has a double dash in its id.
The video link is broken due to this.

How to reproduce

convert a markdown with

the link is for example (there may be many more): https://www.youtube.com/watch?v=--khbXchTeE

in it

Possible solution

exclude double dash conversion inside of urls

Additional context

No response

Did this project help you today? Did it make you happy in any way?

I love it 💯 👍 It made my life as developer and TYPO3 Admin a lot easier. We are writing our contents in a slightly modified version of your commonmark and added some custom extensions

[colinodell]

Thanks for reporting this!

Unfortunately I'm having trouble replicating this issue. Could you please share the exact Markdown you are using? And are you using the EmbedExtension too, or directly embedding some HTML?

[svenignor]

I use the embed extension also and the YouTube link is on one line without surrounding text or < >.

the used markdown and the regex from the DashParser (https://github.com/thephpleague/commonmark/blob/2.7/src/Extension/SmartPunct/DashParser.php#L31) in action can be seen here: https://regex101.com/r/VWATVi/1

[colinodell]

Hmm 🤔 I tried recreating this bug with the following program:

<?php

use League\CommonMark\Environment\Environment;
use League\CommonMark\Extension\CommonMark\CommonMarkCoreExtension;
use League\CommonMark\Extension\Embed\Bridge\OscaroteroEmbedAdapter;
use League\CommonMark\Extension\Embed\EmbedExtension;
use League\CommonMark\Extension\SmartPunct\SmartPunctExtension;
use League\CommonMark\MarkdownConverter;

require_once __DIR__ . '/vendor/autoload.php';

$config = [
    'embed' => [
        'adapter' => new OscaroteroEmbedAdapter(),
        'allowed_domains' => ['youtube.com'],
    ],
];

$environment = new Environment($config);
$environment->addExtension(new CommonMarkCoreExtension());
$environment->addExtension(new EmbedExtension());
$environment->addExtension(new SmartPunctExtension());

$converter = new MarkdownConverter($environment);

echo $converter->convert('https://www.youtube.com/watch?v=--khbXchTeE')->getContent();

And got the following output with the expected double dash:

<p><a href="https://www.youtube.com/watch?v=--khbXchTeE">https://www.youtube.com/watch?v=--khbXchTeE</a></p>

(Note that the video's owner has disabled embedding, which is why this creates a link instead of an <iframe>)

Do you get the same results when running my test script? Can you show me a modified version which demonstrates the en-dash problem you're seeeing?

[svenignor]

Oh, my bad. Sorry for wasting this unnecessary time. I just realized I commented out the embed extension a while ago. It works with the embed extension. However, the "error" occurs without the embed extension (I used your script with the embed extension enabled/disabled to make sure it wasn't something else in my project causing the error).

[colinodell]

Okay, I think I see what's happening:

• If you use just CommonMarkCoreExtension ("vanilla CommonMark") then https://www.youtube.com/watch?v=--khbXchTeE is treated as plain text and not a link
• If you add SmartPunctExtension then any -- sequence within plain text will be converted to –

<?php

use League\CommonMark\Environment\Environment;
use League\CommonMark\Extension\Autolink\AutolinkExtension;
use League\CommonMark\Extension\CommonMark\CommonMarkCoreExtension;
use League\CommonMark\Extension\SmartPunct\SmartPunctExtension;
use League\CommonMark\MarkdownConverter;

require_once __DIR__ . '/vendor/autoload.php';

$config = [];

$environment = new Environment($config);
$environment->addExtension(new CommonMarkCoreExtension());
// TODO: Try un-commenting these together or separately
// $environment->addExtension(new SmartPunctExtension());
// $environment->addExtension(new AutolinkExtension());

$converter = new MarkdownConverter($environment);

echo $converter->convert('https://www.youtube.com/watch?v=--khbXchTeE')->getContent();

To create links you either need to:

• Use the [link](/uri "title") syntax; or,
• Use the <link> syntax; or,
• Use the AutolinkExtension or GithubFlavoredMarkdownExtension which allow plain-text, unformatted URLs to be converted to links

I'm guessing you probably don't have AutolinkExtension or GithubFlavoredMarkdownExtension enabled? Which would explain why the SmartPunctExtension is acting on that text.