Html Serializer retains excess whitespace between tags

For example,

Html.deserialize(`<div>
    <strong>Test</strong> 
</div>`)

Will render Test with four leading spaces. I’m not sure if you want to leave this up to the developer to implement their own rules, but I believe we should provide a foundation that follows spec.

I would recommend changing the default text node rule at html.js#L41 to something like:

         if (el.type == 'text') {
             let text = el.data

             // Strip extra spaces
             if (!text.trim()) text = el.next && el.prev ? ' ' : ''

             return {
                 kind: 'text',
                 text: text,
             }
         }

[roadscape]

Whitespace normalization is not very straightforward; your change may break formatting in <pre> elements for example.

In serializers/html.js#L94, you can try changing

const $ = cheerio.load(html).root()

to

const $ = cheerio.load(html, {normalizeWhitespace: true}).root()

cheerio will pass this option directly to the underlying htmlparser2, unfortunately it does not do a proper job.

I had much better results running incoming html through htmlclean before passing it to cheerio. But it seems to rely on regex and if this is the case I wouldn’t trust it unless you knew incoming HTML was always well-formed.

[zhujinxuan]

Got it. Thank you very much. On Mon, Jan 8, 2018 at 19:48 Ian Storm Taylor [e-mail] wrote:

@zhujinxuan https://github.com/zhujinxuan I think this is something that should be handled outside of Slate if it’s something you need. For the most part, the HTML serializer is used with programmatically generated HTML where whitespace is significant.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/ianstormtaylor/slate/issues/407#issuecomment-356144140, or mute the thread https://github.com/notifications/unsubscribe-auth/AAtir7XCecz7EoEnwlAERE_8WYpAav6Uks5tIrdUgaJpZM4KgQYv .