[tacc-tacc]

@5j9 What do you think of this?

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (d281474) to head (8a63030).
⚠️ Report is 2 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff            @@
##              main      #160   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           16        16           
  Lines         2442      2459   +17     
=========================================
+ Hits          2442      2459   +17     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[tacc-tacc]

@5j9 Will you merge this PR? I think is ready but you have to review everything. Here's the deal:

• Rewrote the LIST_PATTERN_FORMAT with sublist patterns: now it processes ;a:b as ;a, :b instead of ;a:b, :b.
• Also, the sublists must repeat exactly the last pattern so as in :;a::b\n:;c, here :;c is nested in ::b\n and not in :;a. For this, i replaced (?&pattern) with (?P=pattern) so it repeats the captured group exactly. The reason why I enclosed with a new label (pattern_) is to call the spans for the convert method, since bare backreference does not save the spans.
• I think you should shadow everything before parsing lists, or at least with : lists. I think this should include multiline tables since that is how Parsoid works, as I understand. I shadowed almost everything, but you should review that part, specially because it is not clear to me whether the shadows should be added in get_lists as I did or in lists_shadow_ss method.
• regarding templates and parser functions, the result depends on the output, but I think it's better to guess that it won't output a : so we can retain the entity as a whole and avoid chopping it while parsing definition lists.

Please, carefully examine the new testcases I added and compare how it behaves this new version with the previous one. I think it is too much important to be ignored.

[tacc-tacc]

@5j9 Can you please answer this thread instead of merging this on the sly? I found by chance your recent edits, but I think my expectation in the testcase for ;;a:1\n;;b:2\n;c:3\n is right: there is one list with three items, not two. Can you try this on en.wikipedia or in a wiki that has enabled Parsoid by default? You might be renderizing with legacy parser.