html message line breaks are not rendered

[maralorn]

When using email2matrix with SkipMarkdown = false I get matrix messages with the formatted_body field containing html, but \n line breaks. At least riot doesn’t render those. Is it perhaps necessary to replace them with <br> tags? Or does this mean the email (from prometheus in my example) is not standard conform?

[spantaleev]

# Subject\n\nBody seems to yield formatted_body of:

<h1>Subject</h1>\n\n<p>Body</p>

You'd like the HTML in formatted_body to not have \n\n in it at all, but be a single line of HTML tags?

I think the above is pretty natural and also how a human would write the HTML. Humans typically don't put all their HTML tags on the same line. Whitespace (new lines) like this don't cause issues.

[maralorn]

Thanks for the reply.

No the both newlines at the beginning of the formatted_body template are no problem at all.

The problem is that I got an html-formatted email which got simplified by email2matrix as the formatted body. In the <p>Body</p> part it now contains \n which are not rendered (by riot-web at least) but should.
e.g. I got an email containing a table which got simplified to all rows being in one line.
When I set SkipMarkdown = true I get working linebreaks in the message. (Because in the body riot renders \n as linebreaks.)

[spantaleev]

e.g. I got an email containing a table which got simplified to all rows being in one line.

Ahh.. this is because of the way we do things.

Email2Matrix only deals with the plain-text part of the email message.
We use the enmime library to help us with that.

As per its documentation:

The Envelope contains both the plain text and HTML portions of the email. If there was no plain text Part available, the HTML Part will be down-converted using the html2text library[1]. The root of the Part tree, as well as slices of the inline and attachment Parts are also available.

I imagine that your email has no plain-text part, which causes its HTML to be down-converted to plain-text. Which probably yields an ugly "table" in any case.

Depending on SkipMarkdown, different things may happen:

• if SkipMarkdown = true, we just send that plain-text body pretty much as-is, leaving riot-web (or whatever client you use) to interpret it as it wishes - "\n being a line-break
• if SkipMarkdown = false, we pass the message through the blackfriday library, which likely does something weird for you

---

In any case, this whole thing makes me wonder whether we should have an option for simply relaying the HTML part of an email as-is. Right now we're pretty insistent on turning it into plain-text. I'm not sure how useful that is though.. Perhaps you (or others) could share some feedback on what kind of stuff you'd like to send. In the end, relaying HTML as-is may not work great either, as riot-web (or other clients) may choke on more complicated parts.

[maralorn]

Yep, that’s the problem. I must admit I am not entirely sure if the problem is in enmime or blackfriday.

I have no idea which portion of HTML riot is willing to render. But obviously actually passing the table through might look much nicer.

[maralorn]

To elaborate on that: If enmime makes just one linebreak where there is actually supposed to be a rendered linebreak that will look fine in the plaintext version. But a markdown renderer as blackfriday will assume that two lines without a blank line in between are in the same paragraph and remove the newline. So using first a html2text library and then treating the text as markdown might be the wrong approach.