- šŗšøUnited States jameslp
I've opened an issue with CKEditor on GitHub.
- Status changed to Postponed
about 1 year ago 10:39am 2 November 2023 - š§šŖBelgium wim leers Ghent š§šŖšŖšŗ
Thank you! š We'll need test coverage on the Drupal side to ensure CKEditor 5 never regresses against this without us knowing. Would you be interested in contributing that, @jameslp? š This would be a great addition to
\Drupal\Tests\ckeditor5\FunctionalJavascript\TableTest
.There's metadata loss here, not "hard" data loss. So reducing priority to .
Please vote with a š there if you want to see it solved sooner.
- š³šæNew Zealand jweowu
There's metadata loss here, not "hard" data loss.
It's certainly data corruption and a hard loss of semantic meaning.
I'd agree that some HTML markup could be classed as metadata for most intents and purposes, but table markup is so integral to the semantic meaning of that content that I don't think you can consider it as anything other than "data".
This issue might be treated as less important on the basis of it being likely to affect a relatively small set of users, but for those affected users I think "data loss" is an appropriate description.
- š§šŖBelgium wim leers Ghent š§šŖšŖšŗ
Forgot to update this issue ā¦ but I agree with you, @jweowu :)
See š [upstream] [Table] CKEditor 5 changes and keeps only a single , which impedes screen reader accessibility of HTML tables Postponed and the current issue summary: I've categorized this (with @lauriii's +1) as the current highest priority for the CKEditor team to fix as far as the Drupal project is concerned :)
- š³šæNew Zealand ericgsmith
Cross posting for visibility - but I have added comment and a draft MR with a PoC to resolve the tbody part of problem on https://github.com/ckeditor/ckeditor5/issues/14911#issuecomment-1815402805
Would love for anybody impacted by this issue to head over to the CKEditor issue to take a look and provide any relevant feedback.
- šŗšøUnited States bkosborne New Jersey, USA
Wim, in #8 I think you meant to tag a separate issue but you tagged this one
- š§šŖBelgium wim leers Ghent š§šŖšŖšŗ
Indeed! š³š Fixed š
- šŖšØEcuador jwilson3
There's metadata loss here, not "hard" data loss. So reducing priority to Major.
There's an argument here that this should be left as critical
Iām responding to an RFP for a city government website proposal, and one question completely disqualifies Drupal 10 with CKE5 as inconformant:
To comply with the U.S. Access Board Web-based Intranet and Internet Information and Applications (1194.22) provisions, the solution provides the following:
- The solution allows markup to be used to associate data cells and header cells for data tables that have two or more logical levels of row or column headers.
The CKEditor 5 demo breaks the table example from Curriculum for Web Content Accessibility Guidelines 1.0: 5.2 - For data tables that have two or more logical levels of row or column headers, use markup to associate data cells and header cells. (created in year 2000) (Paste the table in via the Source mode button).
- š§šŖBelgium wim leers Ghent š§šŖšŖšŗ
This already is the most important remaining prioritized upstream blocker (see š± [meta] [upstream] Prioritized CKEditor 5 upstream blockers Active ). The CKEditor 5 team knows.
You already did: https://github.com/ckeditor/ckeditor5/issues/14911#issuecomment-1942852909 šš Thank you!
Bumping this to .
Thank you very much, @jwilson3! Crediting you š
- šŖšØEcuador jwilson3
No problem at all. Thanks to you, Wim, for your continued and undying efforts here in general.
Has there been any update regarding this issue. We've newly upgraded to CKEDITOR 5 and we're seriously considering rolling back due to how it mangles tables.