ws-export is adding unwanted line from the header template module in the exported Epub
Closed, ResolvedPublicBUG REPORT
Actions

Description

While using Module:Header template in Bangla Wikisource or similar template in French Wikisource, the exported Epubs are having the following issue.

For ভানুসিংহের পত্রাবলী from Bangla Wikisource, the title page of in the Epub file has this line added at the top.

This line is coming from the following code from the module

container:tag('link')
    :attr('itemprop', 'mainEntityOfPage')
    :attr('href', page:fullUrl(nil, 'canonical'))
container:tag('meta')
    :attr('itemprop', 'inLanguage')
    :attr('content', 'bn')

This issue was noticed very recently, it was not there a day ago (most probably).

Screenshot from 2025-11-29 20-49-23.png (963×595 px, 96 KB)

Related Objects

Mentioned In: T411280: Links in <mapframe> broken when using Parsoid
Mentioned Here: T408915: visualdiff testing: Escaped link elements are shown in page view on fr.wikisource.org
T411280: Links in <mapframe> broken when using Parsoid

Event Timeline

Bodhisattwa created this task.Nov 29 2025, 3:22 PM

Restricted Application added a project: Community-Tech. · View Herald TranscriptNov 29 2025, 3:22 PM

Restricted Application added a subscriber: Aklapper. · View Herald Transcript

Same problem on books exported from French Wikisource. Some tags are urlencoded and therefore not recognised.

I don't think this is a WS Export issue, the wayward HTML is also visible in the Parsoid output at e.g.: https://bn.wikisource.org/api/rest_v1/page/html/%E0%A6%AD%E0%A6%BE%E0%A6%A8%E0%A7%81%E0%A6%B8%E0%A6%BF%E0%A6%82%E0%A6%B9%E0%A7%87%E0%A6%B0_%E0%A6%AA%E0%A6%A4%E0%A7%8D%E0%A6%B0%E0%A6%BE%E0%A6%AC%E0%A6%B2%E0%A7%80

Or if viewed with useparsoid=1, e.g. https://bn.wikisource.org/w/index.php?title=%E0%A6%AD%E0%A6%BE%E0%A6%A8%E0%A7%81%E0%A6%B8%E0%A6%BF%E0%A6%82%E0%A6%B9%E0%A7%87%E0%A6%B0_%E0%A6%AA%E0%A6%A4%E0%A7%8D%E0%A6%B0%E0%A6%BE%E0%A6%AC%E0%A6%B2%E0%A7%80/%E0%A7%A7&useparsoid=1

Denis_Gagne52 mentioned this in T411280: Links in <mapframe> broken when using Parsoid.Nov 30 2025, 2:47 PM

May be related with T411280

Bodhisattwa added a project: Parsoid.Nov 30 2025, 6:25 PM

@cscott I know you that y'all on the Parsoid team did a bunch of work on ProofreadPage and the <pages /> tag recently, this might be a regression that made it through ?

Soda added a project: ProofreadPage.Dec 2 2025, 7:15 PM

Soda removed a subscriber: Bodhisattwa.

Soda added a subscriber: Bodhisattwa.

This looks similar to this one: https://phabricator.wikimedia.org/T408915
I've already added some comments while investigating it here: https://phabricator.wikimedia.org/T408915#11331410

Very similar indeed but T408915 was registered on October 31 while the problem appeared on ws-export only on November 28.
The books downloaded before November 28 are well constructed. Afterwards, several [ '>', '<' ] need to be replaced with [ '>', '<' ] to reactivate the link tags in each chapter of an epub file.
This is the opposite of the replacement found here : https://gerrit.wikimedia.org/r/c/mediawiki/core/+/993051/5/includes/Parser/CoreTagHooks.php#85).

MSantos added a project: Content-Transform-Team (Work In Progress).Dec 4 2025, 3:22 PM

MSantos moved this task from Needs Triage to Bugs & Crashers on the Parsoid board.

MSantos moved this task from Backlog to Needs Investigation on the Content-Transform-Team (Work In Progress) board.

Hello, There is the same problem while exporting in "pdf" in French Wikisource, for example after exporting the book : https://fr.wikisource.org/wiki/Le_Ruban_(Feydeau)
I had this morning on the first page :
<link itemprop='mainEntityOfPage'
href='https://fr.wikisource.org/wiki/Le_Ruban_(Feydeau)'
/><meta itemprop='inLanguage' content='fr' /><meta
itemprop='http://purl.org/library/placeOfPublication'
content='Paris' /><link itemprop='mainEntityOfPage'
href='https://fr.wikisource.org/wiki/Fichier:Feydeau_-
_Th%C3%A9%C3%A2tre_complet,_volume_8,_1948.djvu
' /><meta itemprop='width' content='546' /><meta
itemprop='height' content='857' /><meta
itemprop='fileFormat' content='image/vnd.djvu' />

I think this is more specific to the issue:
https://gerrit.wikimedia.org/g/mediawiki/services/parsoid/+/09d3da067bd346f41a1b1763e298b7d31b86985f/tests/parser/parserTests-knownFailures.json#133

So if this is under known failures then we need to fix it on parsoid.

As this is affecting reading experience on Wikisource reader app, tagging to track the updates there

Bodhisattwa moved this task from General to Android Reader Experience on the Wikisource Reader App (Android) board.Dec 10 2025, 5:25 PM

After many days beeing down, WS Export is running again... But we have still the same problem on some pages at the begining of the files exported.

For exemple, when exporting in "pdf" the book : https://fr.wikisource.org/wiki/Le_Ruban_(Feydeau)

We have on the third page oh the export :
<link itemprop='mainEntityOfPage'
href='https://fr.wikisource.org/wiki/Le_Ruban_(Feydeau)'
/><meta itemprop='inLanguage' content='fr' /><meta
itemprop='http://purl.org/library/placeOfPublication'
content='Paris' /><link itemprop='mainEntityOfPage'
href='https://fr.wikisource.org/wiki/Fichier:Feydeau_-
_Th%C3%A9%C3%A2tre_complet,_volume_8,_1948.djvu
' /><meta itemprop='width' content='546' /><meta
itemprop='height' content='857' /><meta
itemprop='fileFormat' content='image/vnd.djvu' />

Apparently the issue is gone because of this reversal - https://gerrit.wikimedia.org/r/c/operations/mediawiki-config/+/1225613 . If the issue reappears at some point in future, will reopen again.

Bodhisattwa reassigned this task from Bodhisattwa to Jgiannelos.Jan 17 2026, 8:38 AM

Bodhisattwa moved this task from Backlog to Closed on the Bangla-Sites board.Jan 19 2026, 4:10 PM

	F70771426: Screenshot from 2025-11-29 20-49-23.png
	Nov 29 2025, 3:22 PM

ws-export is adding unwanted line from the header template module in the exported EpubClosed, ResolvedPublicBUG REPORTActions

Description

Related Objects

Event Timeline

ws-export is adding unwanted line from the header template module in the exported Epub
Closed, ResolvedPublicBUG REPORT
Actions