Support gzipped WASM inputs in utils::parse_wasm and utils::parse_wasm_file
#55
Conversation
src/utils.rs
Outdated
| } else if bytes.starts_with(GZIPPED_WASM_MAGIC_BYTES) { | ||
| Err(io::Error::new( | ||
| io::ErrorKind::InvalidInput, | ||
| "Please use parse_wasm_robust with gzipped inputs.", |
There was a problem hiding this comment.
This doesn't seem to be a breaking change, why not just decompress here, instead of a new function?
There was a problem hiding this comment.
The original public function parse_wasm's signature has bytes: &[u8], changing it to bytes: Vec<u8> would be a breaking change, right?
If we keep it &[u8], then the decompress function's result type (Result<Vec<u8>) is incompatible with it in one of the branches. So either we would need to inline decompress or have another function that takes Vec<u8> to begin with.
What's your preference? Am I missing a better possible implementation?
There was a problem hiding this comment.
If we keep it &[u8], then the decompress function's result type (Result<Vec) is incompatible with it in one of the branches.
We can make both branches to return Cow::Borrowed and Cow::Owned, so that they end up with the same type.
There was a problem hiding this comment.
Seems to have worked right away. Thanks for the tip!
Good question. In my view, this should be controlled separately. I.e., we could add a feature to enable emitting gzipped WASMs, but it doesn't have to be part of this feature. |
aterga
changed the title
utils::parse_wasm and utils::parse_wasm_file
The goal of this PR is to support gzipped WASMs as inputs in the ic-wasm crate. This is aligned with the IC interface spec that says:
https://internetcomputer.org/docs/current/developer-docs/smart-contracts/deploy/larger-wasm/