Putting the same resource that already exists should not result in a change

[abalmos]

OADA should compute the difference between the put body and the existing result. Empty diffs should translate into a no-op, not update rev, nor create a change. This would make writing data sync microservices much easier / correct.

[aultac]

I think that would add a decent amount of overhead, especially for larger documents for which you are only updating a small portion. Maybe it could compare the current body with the current rev's change document instead? i.e. if it is an exact repeat of the previous change, then it is most likely a duplicate to be discarded?

[abalmos]

I think getting the resource and saving a smaller patch is likely to be an overall save on processing as compared to committing an unnecessarily large change document (and, as currently written, committing the same document to the resource node too). It is also much smaller than many watchers GETting the full resource as the result of a small change. Finally, it seems too early to be worrying about performance over function. As is a common function is quite unfriendly and significantly complicates syncer services.

This should be somewhat "easy" to implement if we use what is already existing. They could be some issues around concurrency. Something like: https://www.npmjs.com/package/fast-json-patch