Whitespace in HTML parsed incorrectly #27

New Issue

shadowfacts · 2019-07-01T13:20:07Z

shadowfacts commented

2019-07-01 13:20:07 +00:00

Posts authored in Markdown or HTML have extra whitespace at the beginning of paragraphs.

Screen Shot 2019-07-01 at 9.19.13 AM.png

196 KiB

shadowfacts added this to the 1.0.0 milestone 2019-07-01 13:20:07 +00:00

shadowfacts added the

bug

label 2019-07-01 13:20:07 +00:00

shadowfacts changed title from ~~Extra whitespace in Markdown/HTML formatted posts~~ to Whitespace in HTML parsed incorrectly

2020-01-21 02:08:39 +00:00

shadowfacts commented

2020-01-21 02:17:21 +00:00

SwiftSoup, the HTML parsing library we currently use parses whitespace in between HTML elements incorrectly, e.g. it will pars a\nb as [paragraph element, TextNode containing a space, paragraph element]. The TextNode containing the space is what's showing up when rendering.

This issue only manifests with rich text posts because non-rich text posts don't have their text wrapped in paragraph tags.

The newlines in the raw HTML should be ~~parsed correctly and then~~ collapsed per the CSS whitespace collapsing rules.

Potential solutions to this are:

Use the NSAttributedString HTML initializer

This has the downside of required two HTML parses (first with something else to sanitize the HTML, second to convert into an attributed string) which would be slower than ideal.
Fix SwiftSoup

good luck
Switch to a different HTML parsing library

HTMLReader seems like it could work, but most libraries seem like crap
Manually scan through the attributed string after it's generated and collapse whitespace per the CSS rules

just kind of a pain in the ass

~~SwiftSoup, the HTML parsing library we currently use parses whitespace in between HTML elements incorrectly, e.g. it will pars `a\nb` as [paragraph element, TextNode containing a space, paragraph element]. The TextNode containing the space is what's showing up when rendering.~~ This issue only manifests with rich text posts because non-rich text posts don't have their text wrapped in paragraph tags. The newlines in the raw HTML should be ~~parsed correctly and then~~ collapsed per the [CSS whitespace collapsing rules](https://www.w3.org/TR/css-text-3/#white-space-phase-1). Potential solutions to this are: - Use the [`NSAttributedString` HTML initializer](https://developer.apple.com/documentation/foundation/nsattributedstring/1525953-init) This has the downside of required two HTML parses (first with something else to sanitize the HTML, second to convert into an attributed string) which would be slower than ideal. - Fix SwiftSoup good luck - Switch to a different HTML parsing library [HTMLReader](https://github.com/nolanw/HTMLReader) seems like it could work, but most libraries seem like crap - Manually scan through the attributed string after it's generated and collapse whitespace per the CSS rules just kind of a pain in the ass

shadowfacts referenced this issue from a commit

2020-01-22 02:22:21 +00:00

Collapse whitespace according to CSS spec after converting HTML to attributed string Fixes #27

shadowfacts closed this issue

2020-01-22 02:22:21 +00:00

shadowfacts referenced this issue from a commit

2020-01-22 02:28:17 +00:00

Collapse whitespace according to CSS spec after converting HTML to attributed string Fixes #27

Sign in to join this conversation.