in any case i am wondering about BIP158 and how much privacy does that offer in practice since it seems to only make it harder to analyze not impossible. and more importantly how does the implementation deal with this BIP since i have yet to dissect Wasabi
Generally, the privacy is definitely better than with BIP37. However it still depends on the implementation on how much better it is.
Not leaking your wallet addresses, but downloading full blocks from nodes in the bitcoin p2p network, already makes it much better.
Obviously, there would be some attacks like monitoring which blocks are being downloaded and compare that to the transaction graph as mentioned in the bitconi-dev mailing list.
However, by connecting via TOR, downloading each block from a different node through a new circuit (what wasabi is actually doing), as well as choosing an appropriate false positive rate (for downloading blocks which do not contain any address of yours), that's a huge gain in privacy since a simple comparison to the transaction graph is not that easily possible anymore (downloaded via different IP's / tor circuits & downloaded from different nodes & false positive rate).
With that, it is no longer trivial to establish a connection between a specific wallet and its addresses.
