Not sure I agree with

some large tables often contain images and are not problematic.

While a page like NRHP listings in Providence, RI would continue to count in the listings, that's only incidental to the fact the images aren't in the first column. That seems like a page that could very easily be a few more columns wide (the location and date listed columns particularly) and so should not be ignored by this lint. I anticipate there are many others with an image in the first column that are exactly this kind of list.

Tables inside tables

These are an antipattern anyway, but the complaint in the context of this request seems reasonable to correct.

Refined the rule.

Changed project tags to reflect that the change will be in Parsoid, not the Linter extension which provides the lint database and API and UI for accessing those lints ...the lint generation exists elsewhere, currently only in Parsoid.

I don't see how having an image in the first column, or not, is relevant to the width of a table. If this test is going to work, it needs to test for the actual conditions that make a table too wide for mobile or for a reasonable width of a content column in the most prevalent skins.

According to @Mabualruz the code on English Wikipedia is outdated so moving this task until later.

This Linter tracking should be disabled and refined via queries and reports until the actual problem conditions and criteria are understood.

Change 928674 had a related patch set uploaded (by Sbailey; author: Sbailey):

[mediawiki/services/parsoid@master] WIP Large tables lint test for nested tables

https://gerrit.wikimedia.org/r/928674

Change 928674 abandoned by Sbailey:

[mediawiki/services/parsoid@master] WIP Large tables lint test for nested tables

Reason:

Code used in another patch

https://gerrit.wikimedia.org/r/928674

Can you please turn off this linter condition while it this task is stalled and on the backlog? The false positives for this linter condition result in a lot of noise on quarry reports that ask for pages with any linter condition on them.

@ovasileva I do think we should plan to revive this effort next quarter. It is frequently coming up in community discussions and bug reports and something we need to improve. I think off the back of the night mode autocorrection we have a better idea of how this sort of thing should work.