Line break in formatting not respected

[vwkd]

Turndown seems to not respect line breaks inside formatting tags like in <strong><br></strong>.

For example

<span>foo</span><strong><br></strong><span>bar</span>

foobar

I'd expect it to generate the same as the following without formatting tags.

<span>foo</span><br><span>bar</span>

foo
bar

[taythebot]

I added <br> parsing with a rule myself

turndownService.addRule('br', {
    filter: 'br',
    replacement: () => '\n',
});

[bensquire]

Seeing the same issue, but a bit more comlpex:

<p>As a <strong>user<br /></strong>I want to</p>

Generates:

As a userI want to

When it should generate:

As a user I want to

We've worked rounf this by doing this before the md conversion:

.replace(/<br \/>?(<\/(strong|em|s|code)>)/gi, '$1<br />')

[martincizek]

Turndown generally doesn't support HTML to MD conversions, where the HTML cannot be represented in MD. This way Turndown can remain quite fast with straightforward rules.

@bensquire Yes, custom preprocessing is the right way to go. Although I'd do it on DOM level, see below.

@taythebot You're converting hard line breaks to soft breaks - i.e. visible in MD source, but not rendered. Which is OK, but probably not what most users want. :) You can achieve the same with setting options.br = ''.

I'd like to introduce a way for users to define easy preprocessing rules on top of DOM. It'd be outside of the Turndown core project. But the preprocessed DOM can be directly passed to Turndown, so there won't be that much overhead.