Attribute Listener messes up array insert

[Jafa80]

Describe the bug
Wrapping custom Blot using json value to bold renders the json as string. This does not happen when not wrapped in bold.

The delta code which generates the Problem

Example:

{"ops": [{"insert": "test ", "attributes": {"bold": true}}, {"insert": {"mention": {"value": "custom-value"}}, "attributes": {"bold": true}}, {"insert": " test", "attributes": {"bold": true}}]}

The expected html output the delta should produce

Example:

<p><strong>test </strong><strong><mention>custom value</mention></strong><strong> test</strong></p>

Expected behavior
Bold attribute listener should not mess up the value. jsonInsert needs to process before the attributes

Additional context
To fix the problem I did following:

1. overwrite bold attribute listener to prevent jsonInsert processing

class CustomBold extends InlineListener
{
    /**
     * {@inheritDoc}
     */
    public function process(Line $line)
    {
        if ($line->getAttribute('bold') && !$line->isJsonInsert()) {
            $this->updateInput($line, '<strong>'.$line->getUnsafeInput().'</strong>');
        }
    }
}

$lexer->overwriteListener(new Bold(), new CustomBold());

2. add the attribute processor to custom Listener processor

public function process(Line $line)
    {
        $mention = $line->insertJsonKey('mention');
        if ($contractNumber) {
            $this->updateInput($line, '<span>'.$mention['value'].'</span>');
        }
        if ($line->getAttribute('bold')) {
            $this->updateInput($line, '<strong>'.$line->getInput().'</strong>');
        }
    }

[nadar]

Would you like to provide a PR to fix this issue? Would be nice 👍

[Jafa80]

Would you like to provide a PR to fix this issue? Would be nice 👍

This is my first interaction with the library, so it will take me a bit time to figure out the solution, to fix the root cause.
I would assume that there needs to be a prioritization on JsonInsert listeners being processed before attribute listeners as it works perfectly fine with my workaround. I will tinker more and see if I can wrap my head around processor behavior.