Agree.
I wonder how this was done. I believe there is no standard for a "separator", and the text is always localized...

I just tried to play with the EmailReplyParser. getVisibleText() method returns the content without a quoted email text, so no quotes will be inside the comments. Another issue - it works with the text_raw only, so getVisibleText() will return GFM formatted content and we will need to parse it back...

I've been playing with this but didn't get any usable results :(

Some further info:

The "original" used by GitHub (Ruby): https://github.com/github/email_reply_parser

I thought we might do this with JS after rendering the page, so I tried https://www.npmjs.org/package/emailreplyparser but it turns out that it also works only on the raw text...

So the only thing that occurs to me is to "split" the comment when it is imported via the CLI script and to store the parts JSON encoded, so we would have to decode them on display.

So the only thing that occurs to me is to "split" the comment when it is imported via the CLI script and to store the parts JSON encoded, so we would have to decode them on display.

Well at least something...

Would https://github.com/willdurand/EmailReplyParser help?