Support new stash methods

ivoleitao / stash

Key-value store abstraction with plain and cache driven semantics and a pluggable backend architecture.

MIT License

87 stars 16 forks source link

Support new stash methods #22

Closed ivoleitao closed 2 years ago

ivoleitao commented 3 years ago

Map<K, V> getAll(Set<? extends K> keys);
- void putAll(java.util.Map<? extends K, ? extends V> map);
- void removeAll(Set<? extends K> keys);

rlch commented 2 years ago

@ivoleitao Any updates on this?

ivoleitao commented 2 years ago

Hello, still in the pipeline. I will try to add it on the next version

naturallymitchell commented 2 years ago

https://github.com/ivoleitao/stash/commit/cf6aac31f7da4923d4b6f5c507911a1cb57c82dd & https://github.com/ivoleitao/stash/commit/f700cfc050cf6caa56ff1917cea6b7b3804765a4

@ivoleitao could you please push 4.0.2?

ivoleitao commented 2 years ago

I've found a concurrency problem, that I have to fix. It's an important one that was already present. In a nutshell the current solution had a race condition in the store creation / removal that appeared during the testing of getAll and putAll.

Unfortunately the solution is causing problems in sembast so I was unable to push it during the weekend.

I will try again next weekend.

ivoleitao commented 2 years ago

The problem is now solved in the latest version and the new features published. Unfortunately I had to do a breaking change to fix this concurrency issue. But It shouldn't create that much of a problem. Here’s what I did

BREAKING CHANGE: The creation of vaults and caches now returns a Future<Vault<T>> and Future<Cache<T>>
BREAKING CHANGE: The creation of stores now returns a Future<XXXStore> where XXX is the specific store

There was no other way to fix the issue without compromising the quality of the package