Page MenuHomePhabricator

Weird list corruption in a DiscussionTools edit (2020-10-29)
Closed, ResolvedPublic

Description

https://nl.wikipedia.org/?diff=57433885

There is a numbered list beginning with #<li value="5">, which (I think) is supported in a hacky way in Parsoid. In the edit, items of a nested definition list were pulled out of that numbered list (the leading # disappeared).

Event Timeline

This is probably some old edge case bug (nothing to do with selser or discussion tools)

[subbu@earth:~/work/wmf/parsoid] cat /tmp/lwt
#<li value="5"> a
## b
[subbu@earth:~/work/wmf/parsoid] php bin/parse.php --wt2wt < /tmp/lwt
#<li value="5"> a
#b

[subbu@earth:~/work/wmf/parsoid] cat /tmp/html
<ol><li data-parsoid='{"stx":"html","liHackSrc":"#","autoInsertedEnd":true}'>a<ol><li>b</li></ol></li></ol>

<ol><li>a<ol><li>b</li></ol></li></ol>
[subbu@earth:~/work/wmf/parsoid] php bin/parse.php --html2wt < /tmp/html
#<li>a
# b

# a
## b

I am inclined to remove the parent task but not sure if it impacts any of your tracking on discussion tools reported issues.

ssastry triaged this task as Medium priority.

Change 639280 had a related patch set uploaded (by Subramanya Sastry; owner: Subramanya Sastry):
[mediawiki/services/parsoid@master] Fix old wt2wt bug for listed nested in li-hack lists

https://gerrit.wikimedia.org/r/639280

Change 639280 merged by jenkins-bot:
[mediawiki/services/parsoid@master] Fix old wt2wt bug for lists nested in li-hack lists

https://gerrit.wikimedia.org/r/639280

Will deploy in the next train (week of Nov 16).

Change 641308 had a related patch set uploaded (by Subramanya Sastry; owner: Subramanya Sastry):
[mediawiki/vendor@master] Bump wikimedia/parsoid to 0.13.0-a17

https://gerrit.wikimedia.org/r/641308

Change 641308 merged by jenkins-bot:
[mediawiki/vendor@master] Bump wikimedia/parsoid to 0.13.0-a17

https://gerrit.wikimedia.org/r/641308

Very similar issue from 13 November: https://vi.wikipedia.org/?diff=64043608 (this is before the patch was deployed)