Has support been fully implemented now? Note that https://www.wikidata.org/wiki/Wikidata:Pywikibot_-_Python_3_Tutorial still claims that pywikibot currently doesn't support lexemes at all, which doesn't appear to be the case anymore.

I would like to add another use case: On the Wikidata wiki we have Template:Query page, which is currently coded with plain-old templates and the logic is split across 12 subpages of the template, making it very hard to extend. I rewrote the logic in Lua however unfortunately the code I wrote is useless because Scribunto does not let me access the grandparent frame ... frame:getParent():getParent() is always nil.

I think this change broke a bunch of links that lived outside of the documenation (e.g. on www.mediawiki.org). I fixed a couple of them but I am not sure if these are all because the MediaWiki search for "md_doc" did not seem to find everything.

Can't we just set this new flag? I think it would be nice to address this unnecessary hurdle.

I managed to come up with a workaround for my extension using the ParserPreSaveTransformComplete hook: https://git.push-f.com/mw-vote/commit/?id=166d0fbe2f0355d81669b5fde9c2a9bc576da88a

Well I think that it would have to be a tag-specific setting. Because for some tags like <pre> and <syntaxhighlight> I think it's good that no preSaveTransform take place, but for other specific tags they would be desirable.