Closed lenary closed 10 years ago
So the change to update/3
makes this change fuuuun, because of EQC's symbolic and dynamic values.
I think I may have to add a model_update_error_expected(...)
callback or something similar.
I think @russelldb's new EQC on #83 supersedes any of this work. Let's reopen if/when this needs revisiting.
"Better" remains to be seen, but at least intent was clearer, and it tested far more behaviour of the CRDT itself, which was nice (including serialisation, merges, updates).
It's loltastically out of date. I've done a bit of work when I removed everything related to GC, but there are still some bugs - indeed the only module i've written callbacks for so far is failing right now.
update/3
signature, including precondition failsThis is up so @rdb can get eyes on it tomorrow and tell me if I'm an idiot or not.