Implement gate-featured serde for exported structures #38
Conversation
This file contains 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
The serde implementation sets the `Fp` as 48 bytes, then encodes it and does the decoding and decompression on the deserialization. The deserialization will fail if the `Fp` was encoded incorrectly or the bytes are less than 48. Added tests for G1Affine roundtrip de/serialization. We implement this serde support for `Fp` since it will be necessary so later is easier for us to de/serialize `G2Prepared` which does not carry any `to_bytes` impl.
The serde implementation sets the `Fp2` as 2 - 48 bytes `Fp`s and then encodes them and does the decoding and decompression on the deserialization. The deserialization will fail if any of the `Fp`s was encoded incorrectly or the bytes are less than 48 for each of them. Added tests for roundtrip de/serialization. We implement this serde support for `Fp2` since it will be necessary so later is easier for us to de/serialize `G2Prepared` which does not carry any `to_bytes` impl.
The serde implementation sets the `Fp6` struct as a combination of 3 `Fp2` structs de/serializing them. The deserialization will fail if any of the `Fp2`s were encoded incorrectly. Added tests for roundtrip de/serialization. Implement Serde for Fp2 We implement this serde support for `Fp6` since it will be necessary so later is easier for us to de/serialize `G2Prepared` which does not carry any `to_bytes` impl.
- The serialization is done for `G2Prepared` as for a struct which fields have Serde implemented. - The `subtle::Choice` is encoded as an `u8` and built from it later on the deserialization. - Added roundtrip tests for G2Prepared.
The implementation compresses the point to 48 bytes, then encodes it and does the decoding and decompression on the deserialization. Added tests for G1Affine roundtrip de/serialization.
- The serialization compresses the `G2Affine` point to 96 bytes which are then serialized. - The deserialization then, deserializes the 96 bytes and tries to re-build the point checking it's correctness. If the point is not cannonically encoded or not valid the deserialization process will fail. - Added roundtrip tests for G2Affine.
|
You should add Scalar serialization as well |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
A few weeks ago, I needed
serdeimplementations for the types that this library exports, so I coded them on our fork: https://github.com/dusk-network/bls12_381Then I saw the suggestion raised in #35 by @hdevalence and thought that it would be nice to convert our impl to a
feature-gated behind a serde feature.So I modified our impl according to the gate-feature suggestion and made the PR.
Adds serde impl for:
G1AffineG2AffineG2Preparedwhich then requiredFp,Fp2&Fp6serde impls.All the implementations will make the deserialization fail if any of the items is not
canonically-encoded.Appart from that, I think that implement
WriteandReadfor these types would be also nice, since sometimes, you just want to move the point structures as bytes but without doing any compresions/decompressions, having them "serialized" as bytes (which would be much more performant). See: dusk-network/plonk#190 where aProoftakes a considerable amount of time to be deserialized just because we calldecompressionfor each point we store on it computing the security checks.I added the
serdefeature to thedefaultfeatures group, but feel free to remove it from there since Idk if everyone will want the feature enabled by default.