HMTL being included in post Metadata, and not scrubbed when removed from post.

Describe the bug

When using html for inserting images, video or even just centering images, the html code is being included in the metadata desriptions. This is viewable both on google search results, and posted links to sites like Mastodon. Futhermore, deleting the HTML code from the post does not remove the html text from the metadata.

Steps to reproduce (if necessary)

Steps to reproduce the behavior:

1. Publish a post with html code, within first few lines of post.
2. Click on publish
3. post link to post on site such as Mastodon. Code will show in description.

Expected behavior

Code should be ignored and not included in metadata descriptions.

Application configuration

https://write.as

• Federation enabled? yes

Version or last commit:
https://write.as

[thebaer]

Hey, thanks for reporting this. Right now we strip out Markdown for the post descriptions, so we'll need to strip out HTML, as well.

As for the metadata not updating after the HTML has been removed, is that happening when you view the same post on sites like Mastodon? Or when you view source on the post itself? If it only happens on sites like Mastodon, that's likely just because it's caching the metadata.

I have checked google search results again, and the html is now not present within the metadata descriptions. So it looks like google has recrawled the post and updated this. So yes, this leaves just the Mastodon posts with the link. Even when i deleted the original Mastodon post, and then posted a new updated toot, it kept the old metadata.

…

________________________________ From: Matt Baer <notifications@github.com> Sent: 13 March 2019 17:50 To: writeas/writefreely Cc: digitalgyoza; Author Subject: Re: [writeas/writefreely] HMTL being included in post Metadata, and not scrubbed when removed from post. (#83) Hey, thanks for reporting this. Right now we strip out Markdown for the post descriptions, so we'll need to strip out HTML, as well. As for the metadata not updating after the HTML has been removed, is that happening when you view the same post on sites like Mastodon? Or when you view source on the post itself? If it only happens on sites like Mastodon, that's likely just because it's caching the metadata. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub<#83 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AuJ12eOtWYllCTt-OjmpCbKacTtM2T7-ks5vWTpkgaJpZM4bm-LE>.

[thebaer]

Yeah, sounds like that's Mastodon caching the metadata. I'm not sure how to invalidate it, unfortunately... But I fixed this, so that won't happen anymore. The change should be deployed on Write.as within the next day.