Patrick Ohly

Thanks to some optimization (not allocating a "not found" error when the caller in `add` just discards it again) the current implementation is now faster than the original one.

The `AssumeCache` under client-go could be implemented without the "older than" comparison. It would still be useful for my purposes. But can it then be used by the volume binding...

> You can use .generation for this if your changes are in the spec. No, volume binding also needs to cover status changes.

@lavalamp, @liggitt: in order to move forward with this PR, I made the following changes: - The final commit only compares opaque ResourceVersion strings, without the "is newer" comparison. It...

> Watches aren't guaranteed to show you every revision, though? The "is equal" comparison does not depend on that. The logic is "if the AssumeCache currently holds the same object...

Perhaps two examples in the documentation of AssumeCache make this clearer. I've pushed an update...

> Yeah, there's at least one case where this technique can cause backwards time travel, which is the one thing clients basically can't tolerate at all. [...] I think we...

In my case, the controller triggers two external changes, allocation and deallocation of a claim, and records the result in the claim status. When A = "not allocated, pending" and...

> Maybe the trigger will be exceedingly rare or never happen in your case, I think it is indeed. I can easily trigger the situation where the lack of an...

This was discussed in the [September 21st 2022](https://www.youtube.com/watch?v=Z-BqxaML_JE) SIG api-machinery meeting. There was a certain tendency to just allow parsing ResourceVersion, but also concerns that it might not always be...