Open edsko opened 5 months ago
Lowering the priority of this issue to low, since the probability that a peer sends us messages that have compression errors in them is very low.
If we implement this, we should think about whether the difficulty of streaming decompression whilst checking for errors is relevant or not (@FinleyMcIlwaine has some discussion of this issue at https://hackage.haskell.org/package/snappy-c-0.1.0/docs/Codec-Compression-SnappyC-Framed.html#v:decompressWithParams-39-).
Compression
is currently defined asThe type of
decompress
should instead beAnd these decompression failures then need to propagate everywhere, and the specific instances of
Compression
(such asgzip
) need to be modified.