Merged
Conversation
gregorydemay
commented
Feb 14, 2025
gregorydemay
commented
Feb 14, 2025
lpahlavi
reviewed
Feb 18, 2025
Contributor
lpahlavi
left a comment
lpahlavi
left a comment
There was a problem hiding this comment.
Thanks for the PR @gregorydemay! A few nits regarding mostly the documentation, but otherwise looks good! Had to wrap my head a bit around the number of generics but the end result is quite cool.
Co-authored-by: Louis Pahlavi <louis.pahlavi@gmail.com>
ninegua
reviewed
Feb 24, 2025
gregorydemay
commented
Feb 24, 2025
Contributor
Author
gregorydemay
left a comment
gregorydemay
left a comment
There was a problem hiding this comment.
Thanks @lpahlavi for the review! Should be ready for another round.
gregorydemay
commented
Feb 24, 2025
Contributor
Author
gregorydemay
left a comment
gregorydemay
left a comment
There was a problem hiding this comment.
Thanks @ninegua for the review! Should also be ready for a 2nd round
ninegua
approved these changes
Feb 24, 2025
lpahlavi
approved these changes
Feb 25, 2025
Contributor
lpahlavi
left a comment
lpahlavi
left a comment
There was a problem hiding this comment.
Thanks again for this PR @gregorydemay LGTM!
lpahlavi
approved these changes
Feb 25, 2025
gregorydemay
added a commit
that referenced
this pull request
Feb 28, 2025
Follow-up on #370 and #364 to add a conversion layer between request and responses so that the caller only has to deal with types from the Rust [`http`](https://crates.io/crates/http) crate. This has several advantages from the caller's point of view: 1. Can re-use existing types like `Request::builder` or `StatusCode`. 2. Requests are automatically sanitized and canonicalized (e.g. header names are validated and lower cased). 3. Can re-use existing middlewares, like from the `[tower-http`](https://crates.io/crates/tower-http) crate, since many of them require a client that uses the [`http`](https://crates.io/crates/http) crate.
gregorydemay
added a commit
that referenced
this pull request
Mar 7, 2025
Follow-up on #370, #364, and #374 to add a conversion layer between request and responses to handle [JSON RPC](https://www.jsonrpc.org/specification) over HTTP. Concretely: 1. Refactor existing infrastructure to handle errors explicitly, instead of having to downcast `BoxError`. This is beneficial since errors may happen at different layers and they all need to be handled by the above observability layer, e.g., add the corresponding metric in case the response has a non successful status. 2. Refactor existing infrastructure to transform request and response to use a single `Convert` trait to convert one type into another with some potential error. 3. Add a layer for `http::Request` and `http::Response` to transform their body into a JSON RPC type. This uses the above `Convert` trait. This conversion layer is behind a feature flag (`json`) since it requires `serde_json` to handle the serialization/deserialization 4. Add a new layer to filter out non-successful HTTP responses.
Merged
gregorydemay
added a commit
that referenced
this pull request
Mar 11, 2025
gregorydemay
added a commit
to dfinity/canhttp
that referenced
this pull request
Jul 9, 2025
Follow-up on dfinity/evm-rpc-canister#364 to add a new observability layer to take care of logging and metrics. Unfortunately, [`tower_http::trace`](https://docs.rs/tower-http/latest/tower_http/trace/index.html) cannot be used in a canister environment because it measures call latency with `Instant::now`. The proposed layer, while inspired by [`tower_http::trace`](https://docs.rs/tower-http/latest/tower_http/trace/index.html), is also simpler because it does not have to deal with streaming responses. --------- Co-authored-by: Louis Pahlavi <louis.pahlavi@gmail.com>
gregorydemay
added a commit
to dfinity/canhttp
that referenced
this pull request
Jul 9, 2025
Follow-up on dfinity/evm-rpc-canister#370 and dfinity/evm-rpc-canister#364 to add a conversion layer between request and responses so that the caller only has to deal with types from the Rust [`http`](https://crates.io/crates/http) crate. This has several advantages from the caller's point of view: 1. Can re-use existing types like `Request::builder` or `StatusCode`. 2. Requests are automatically sanitized and canonicalized (e.g. header names are validated and lower cased). 3. Can re-use existing middlewares, like from the `[tower-http`](https://crates.io/crates/tower-http) crate, since many of them require a client that uses the [`http`](https://crates.io/crates/http) crate.
gregorydemay
added a commit
to dfinity/canhttp
that referenced
this pull request
Jul 9, 2025
Follow-up on dfinity/evm-rpc-canister#370, dfinity/evm-rpc-canister#364, and dfinity/evm-rpc-canister#374 to add a conversion layer between request and responses to handle [JSON RPC](https://www.jsonrpc.org/specification) over HTTP. Concretely: 1. Refactor existing infrastructure to handle errors explicitly, instead of having to downcast `BoxError`. This is beneficial since errors may happen at different layers and they all need to be handled by the above observability layer, e.g., add the corresponding metric in case the response has a non successful status. 2. Refactor existing infrastructure to transform request and response to use a single `Convert` trait to convert one type into another with some potential error. 3. Add a layer for `http::Request` and `http::Response` to transform their body into a JSON RPC type. This uses the above `Convert` trait. This conversion layer is behind a feature flag (`json`) since it requires `serde_json` to handle the serialization/deserialization 4. Add a new layer to filter out non-successful HTTP responses.
gregorydemay
added a commit
to dfinity/canhttp
that referenced
this pull request
Jul 9, 2025
Follow-up on dfinity/evm-rpc-canister#370, dfinity/evm-rpc-canister#364, dfinity/evm-rpc-canister#374, and dfinity/evm-rpc-canister#375 to automatically retry requests by doubling its `max_response_bytes` when the response was too big.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Follow-up on #364 to add a new observability layer to take care of logging and metrics. Unfortunately,
tower_http::tracecannot be used in a canister environment because it measures call latency withInstant::now. The proposed layer, while inspired bytower_http::trace, is also simpler because it does not have to deal with streaming responses.