Hi Ahmed and Thomas,
Thanks for the continued work regarding the document corruption. I see some parts of the current patch in PR #251 that I believe can be safely included as increments to facilitate progress (I have to admit the complexity of the change and the fact that it mixes several related concerns make it relatively hard to move forward, so smaller chunks may help):
Specifically, LSPEclipseUtils.applyEditsWithVersionCheck() is a very good addition. It would be nice if this could be added in a separate PR with an example of direct consumer (eg LSForatter.applyEdit or something like that).
Also, I may have missed it; but I'm expecting a Test case that would emphasize the current issue(s?) you're trying to solve; on a end-user grain (eg edit;format;edit from different threads). Even if this test can be made to fail 1 time out of 50, that would help to go for increments that can fix it first, and then refactorings that can improve the code. The current tests i see do validate some ideas of the proposal, but I don't think they actually reproduce the issue. Don't hesitate to add tweaks to eg DocumentContentSynchronizer or LSFormatterHandler to artificially add a delay that would emphasize the issue more deterministically.
Cheers,
Mickael
