It is hard to triage extension support and rendering diffs without knowing whether or not they are used on our initial "Parsoid read views for DiscussionTools" wikis.
We should pick a particular wiki to target for the first turn-on of Parsoid read views for DiscussionTools, so that we can then run visual diff testing against that wiki and reprioritize (if needed) any extension or rendering-fidelity work which impacts that particular wiki's Talk pages.
Ideally our first wiki will be (a) relatively low impact, so broken rendering won't be front page news on a national newspaper, (b) use few extensions within Talk pages with active discussions, and (c) have enough users of DiscussionTools that our userbase is meaningful (and so we'll get good bug reports).