Parsoid closes unclosed <del> tag while Tidy doesn't
Closed, DeclinedPublic
Actions

Assigned To

None

Authored By

	Tacsipacsi
	Sep 20 2017, 8:15 PM

Description

Parsoid version ([[https://hu.wikipedia.org/w/api.php?action=visualeditor&paction=parse&page=Wikip%C3%A9dia%3AJavaslatok+j%C3%B3+sz%C3%B3cikkekre%2FA+mes%C3%A9k+mes%C3%A9je&oldid=19163140|visualeditor API module result]]) closes unclosed <del> tag on end of paragraph, while the default page version (Tidy) doesn’t. It’s very confusing as the new wikitext editor uses Parsoid for preview, but the Tidy version appears after saving.

Mentioned In: T176363: New high-priority Linter category for a subset of misnested tags whose behavior will change with a HTML5 parser

This is not specific to Parsoid, but HTML5 vs HTML4. We are in the process of replacing Tidy with Remex (see https://www.mediawiki.org/wiki/Parsing/Replacing_Tidy/FAQ).

You can see that with Remex, the problem you see with Parsoid will show up as well.

The HTML5 spec ( https://www.w3.org/TR/2011/WD-html5-20110525/edits.html#the-del-element ) says that the del element should not cross (implied) paragraph boundaries. So, that is what Parsoid and RemexHTML do. So, the unclosed <del> tag gets closed at the end of the paragraph.

The right fix here is to update the wikitext to add these tags around every paragraph that needs them.