JSON-LD output does not properly escape HTML tags

I'll be AFK through the weekend, will follow up on this on Monday.

I think the .<\/p> is happening because the slash is being escaped.

We might need to answer the bigger question which is what HTML should be included via JSON-LD.

This point came up on Schema Metatag: ✨ Escape html instead of stripping html tags Needs work

Twice, in fact: #2989338: SchemaMetatagManager::encodeJsonld Should Allow for HTML →

The MR uses the below flags copied from SchemaMetatagManager::encodeJsonld

JSON_UNESCAPED_SLASHES | JSON_HEX_TAG | JSON_HEX_APOS | JSON_HEX_AMP | JSON_HEX_QUOT | JSON_UNESCAPED_UNICODE

The results JSON-LD inside the <script> tag is valid.

If the exact flags that are supported become a debate, we can easily make the JSON encode flags a configurable settings at a later point.

This resolves one problem - the output is at least consistent now, so that part is RTBC IMHO.

The second issue is how to determine whether the property's value should be rendered as plain text with HTML tags removed, versus escaping the tags. Do you have thoughts on whether you would prefer this handled through hooks, or whether it should be possible to handle it from the UI?

We should absolutely support and default to allowing HTML tags.

I am open to providing a setting to strip or convert HTML tags to plain text

LGTM!

Automatically closed - issue fixed for 2 weeks with no activity.