> I don't understand why Mozilla don't offer this themselves though, is it too complicated to put the effort in given the size of the userbase?

That's the short answer to the question.

A longer answer: the underlying Exchange protocol is another IMAP-level of complexity of protocol, and the IMAP protocol implementation in Thunderbird is itself running somewhere in the 50KLOC region. On top of that, the protocol itself (I believe) largely works in Outlook's not-a-regular-RFC 822-email-message representation, which means you need to build in a translation layer to make it work well. Finally, a lot of the logic in Thunderbird isn't really designed for pluggable account types, and that it's even possible to a degree is up to work that rkent and I did to make it barely possible for extensions to do so; incorporating the code in core would require properly fixing several of the places rather than continuing the various levels of hacks in place.

Now what's the benefit of adding it core? If you're using TB for your small/medium enterprise, then the better answer is to just switch on IMAP access on the server, which is much better supported in TB (and other clients). So the value-add is comparatively minimal, especially compared to doing something like supporting JMAP (which is somewhat closer to a sane modernization of IMAP than EWS is). It's going to rank lower priority than other fixes, then, especially given the technical debt TB suffers from.