Description

I vaguely recall this issue being well(?) known, but the Parsoid output requires a particular URL setup.

Given a URL of the form: http://localhost/.../mediawiki/index.php?title=Main_Page&variant=sr-ec (for instance) the legacy parser will generate full absolute URLs in its output, eg:

<a href="/.../mediawiki/index.php/Tricky_Question_Marks%3F/Subpath/Child" title="Tricky Question Marks?/Subpath/Child">Tricky Question Marks?/Subpath/Child</a>

however Parsoid generates relative URLs (./Foo) which resolve incorrectly when the base href is '/index.php' because the proper URL *should* be /index.php/Foo.

This is an issue when including Parsoid content on a page, as the base href for the entire page needs to match what Parsoid expects.

Details

	Subject	Repo	Branch	Lines +/-
	[parsoid] Fix Parsoid relative links, take 2	mediawiki/core	master	+19 -15

Customize query in gerrit

Related Objects
Search...

Status	Subtype	Assigned	Task
Stalled		None	T324931 Clean up open RESTBase related tickets
In Progress		None	T262315 <CORE TECHNOLOGY> API Migration & RESTBase Sunset
Resolved		MSantos	T264669 Move VE API from RESTBase into core.
Resolved		MSantos	T301371 Preemptively warm caches for Parsoid output
Open	Release	None	T84936 Release VisualEditor-MediaWiki as "1.0"
Open		None	T50429 [Epic] Support editing parts of a page in VisualEditor-MediaWiki
Open		None	T54365 Explore performance gains from progressive (JIT?) de-alienation in VisualEditor
Open		None	T174303 Copy-pasting linked ISBN numbers from view mode HTML into VisualEditor inserts wikitext links to Special:BookSources (it should turn them into magic links?)
Open	Feature	None	T54091 The read HTML should have hinting to allow full DOM copying (as opposed to just rich copying) from read mode into VE surfaces
Open		None	T55784 [EPIC] Use Parsoid HTML for all page views
Resolved		cscott	T311527 Implement stampede protection for parsoid rendering
Open		None	T272987 [EPIC] Parsoid read views for DiscussionTools
Open		cscott	T367074 Deprecate and remove ParsoidOutputAccess
Open		ssastry	T317018 Make HtmlOutputRendererHelper use ParserOutputAccess
Resolved		daniel	T359426 REST: endpoints that return rendered page content should support all content models
Resolved		cscott	T311648 Allow WikitextContentHandler to use Parsoid for rendering wikitext
Resolved		cscott	T300190 Ensure that if 'useParsoid' flag is present in ParserOptions that we use Parsoid to parse the page.
Open		None	T270219 VisualEditor does not support wikis without short URLs
Open		None	T310520 Parsoid content not compatible with `index.php?title=` URLs

Event Timeline

cscott created this task.Jun 13 2022, 4:00 PM

Not that Parsoid will also append querystrings with delimiter to the end, like ?lang=de

the legacy parser will generate full absolute URLs in its output

See core's Title::getLocalURL()

I vaguely recall this issue being well(?) known, but the Parsoid output requires a particular URL setup.

There are some notes about this in T270219 and T261711.

matmarex added a parent task: T270219: VisualEditor does not support wikis without short URLs.Jun 13 2022, 6:34 PM

ssastry moved this task from Needs Triage to Missing Functionality on the Parsoid board.Jun 14 2022, 9:18 PM

ssastry triaged this task as Medium priority.Jun 14 2022, 10:29 PM

ppelberg moved this task from Backlog to External on the Editing-team (Tracking) board.Jun 16 2022, 3:08 PM

ppelberg moved this task from Backlog to Triaged on the DiscussionTools board.

ssastry removed a project: Parsoid-Read-Views (Phase 1 - DiscussionTools support).Jul 19 2022, 2:36 PM

cscott mentioned this in T350952: After fixing subpage links in Parsoid read views, TOC links are broken.Nov 10 2023, 4:32 PM

It looks like this is more-or-less fixed in Parsoid (but see T350952) in so far as SiteConfig::relativeLinkPrefix() in core jumps through extreme hoops in order to generate a base href and a relativeUrlPrefix() which work even for wikis with ?title= URL setups. But this probably breaks VE (T270219) which still only looks for & strips the basic ./ prefix and gets confused when Parsoid generates hrefs with (for example) ./index.php?title= as the "relative link prefix" instead of ./.