Update #3
Merged
Update #3
Conversation
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
…a-xyz#196) #### Problem There are now a lot of dead crates in the sdk, mostly around the precompiles. #### Summary of changes Remove all references to precompiles, including solana-precompiles, the verify implementation for all precompiles (secp256k1, secp256r1, and ed25519). Also remove solana-feature-set, solana-feature-set-interface, and solana-reserved-account-keys.
* add conversion logic from and into compressed types * fix variable name
* remove DecodeError * fmt toml
#### Problem The publish job against maintenance/v2.x failed to generate a changelog because we look for changes to `master`, but we really want changes up to the current HEAD. #### Summary of changes Use `HEAD` instead of `master`.
…e so that it works with any projective, affine, and compressed types (anza-xyz#202) * define traits to convert signature, pubkey, and proof-of-possession types to projective * implement signaturre and proof of possession verification using traits * remove todo regarding type conversions * use a generic implementation of the verification traits * move verification logic outside of the traits to prevent circular dependencies * directly convert uncompressed types to projective to avoid unnecessary array allocation * remove AsProjective... and Verifiable... traits * add `As*Projective` and `Verifiable*` traits to simplify the function signatures * fix a typo * use a macro for the type `From` and `TryFrom` implementations for projective, affine, and compressed types * move `As...Projective` implementation into a macro * include `BlsError` directly in the macro
* Add solana base3 encoding crate * clippy * use little-endian instead of big-endian * fix big-endian to little-endian in the docs * use `BASE3_SYMBOL_PER_CHUNK` for 80 * add a `max_len` parameter * use u8 to translate the symbols * keep the ternary-bits in ascending order * update the tests for `u8` encoding * make bitvec crate dependency optional * clippy * cargo fmt * cargo sort * add a clarifying comment about extra bytes * add a comment noting that we use 2 bytes to hold the bit length
#### Problem Rust has been upgrading, but not the version used in the sdk repo. Also, with the upcoming breaking changes to the sdk, it's a good time to bump the rust edition to 2024, but we should upgrade to the newest Rust first. #### Summary of changes Bump the version in rust-toolchain.toml and scripts/rust-version.sh, then fix any issues reported.
* turn BorshIoError variants into unit variants make std optional in solana-program-error extract solana-pubkey-error crate add missing feature activation extract instruction-error crate remove unnecessary docs target from pubkey-error make num-traits optional in solana-program-error missing feature activation in solana-program move min_specialization to instruction-error missing feature activation in solana-instruction remove redundant DecodeError constraint from PrintProgramError remove unncecessary Error constraint from PrintProgramError remove num-traits dep and rename the num-traits feature to solana-msg add missing changes after removing num-traits make pubkey-error optional move instruction error codes to program-error and fix pubkey-error import fmt move PubkeyError back to solana-pubkey and move ProgramError conversion to solana-pubkey enable std if solana-msg is enabled add ToStr trait post-rebase fixes fmt toml * Fix CI, remove solana-instruction where not needed * Run cargo-sort --------- Co-authored-by: Jon C <me@jonc.dev>
* Improve sysvar get error handling * Improve use of constants * Fix clippy * Qualify program error use
* refactor out hard-coded test data * clean up test imports * refactor out integration tests * add a clarifying comment regarding the length check in the tests
* refactor out test constants * move bench from sdk into `big-mod-exp` and use criterion * cargo fmt * use `black_box` around the function result rather than the inputs
…ion to be generic (anza-xyz#213) * make the `aggregate_with` and `aggregate` functions generic and add support for dyn types * remove `Default` implementation for public key and signatures; instead add an identity function * cargo fmt
* add benches for base3 encoding * clippy
* vote-interface: rename `VoteState` to `VoteStateV3` * add deprecatio attr
* docs: Fix all warnings in docs build #### Problem Although we check that all of the rustdocs build correctly for the whole sdk, there are warnings in the build. #### Summary of changes Fix the warnings, and forbid warnings in CI. * deny warnings in check-doc.sh * Use non-bare links --------- Co-authored-by: Kevin Heavey <heaveykevin2@gmail.com>
* typo lib.rs * typo abi_example.rs
* docs: Fix tests, run tests in CI #### Problem The doctests are currently very broken, but it wasn't detected because we only run checks on the docs. #### Summary of changes Fix all of the doctests, and add a step to test them in CI. * Fix circular deps
#### Problem There's a maintenance branch for v2 crates in the SDK repo called `maintenance/v2.x`, but there's no way to automatically backport PRs to that branch. #### Summary of changes Copy / paste how Kit does things and include a backporting job. There's also some new info about the maintenance branch and how to perform backports.
* Remove some unused dependencies * return solana-logger * return getrandom
* borsh!: Remove borsh v0.10 support #### Problem The sdk has implemented the borsh v0.10 traits for years, even though v1 has been out for awhile. This was to avoid breaking people who were still on the older crate version. #### Summary of changes It's been two years, and we're making breaking changes, so remove all borsh 0.10 support! * Fix doc build by referencing borsh1
#### Problem The SDK depends on tokio v1.43.0, which has a small security issue. Although this isn't directly an issue since this repo only has libraries, there's no reason to keep around the issue. #### Summary of changes Simply bump to 1.43.1.
* Remove deprecated functions and structs #### Problem There are a lot of deprecated functions and structs all over the SDK that can be removed now that we're making breaking changes. #### Summary of changes Remove it all! * Fixup CI errors
* deps: Update env_logger to v0.11 #### Problem The SDK and Agave are still using env_logger v0.9, which has a dependency on an unmaintained crate, atty. #### Summary of changes Update to the newest version of env_logger. * Specify a slightly older version for minimal-versions to pass
* wasm-js: Remove system program dependency #### Problem While working on the refactor of all wasm-js logic into sdk-wasm-js, I noticed that no matter what I did, `SystemInstruction` still made it into the final build. That made it impossible to build the package with breaking changes. #### Summary of changes Remove the usage of `SystemInstruction` in the tests. This required exposing `Instruction` and `AccountMeta` through wasm-bindgen to be able to create instructions. The API doesn't need to be perfect, just usable, so I went with a builder pattern for adding `AccountMeta`s to an `Instruction`, which hugely simplifies the logic. Otherwise we'd need to perform conversions between JS arrays and `Vec<AccountMeta>`, which I was too lazy to do properly because wasm-js crate is more of a proof-of-concept. While doing that, I moved the wasm-js deps to the component crates that were required for the tests. * Regenerate cargo.lock, fixup wasmbindgen attribute * No builder pattern
* account-info: Rename `rent_epoch` to `unused` #### Problem The concept of rent-epoch doesn't exist anymore in the runtime, but the account-info type for on-chain programs requires some 8-byte field where the rent-epoch is currently. #### Summary of changes Rename the field, skip it in the program entrypoint. Also, move the layout test into solana-account-info. BREAKING: `solana_account_info::Account` trait removes `Epoch` * Rename unused -> _unused, deprecate it
…nza-xyz#239) * message: Remove dependency on solana-system-interface #### Problem While debugging changes to the wasm-js build, I noticed that `solana-transaction` and `solana-message` were pulling in `solana-system-interface` in a lot of situations, even though there's a lot of code in place to make sure that doesn't happen. #### Summary of changes Totally remove the dependency on solana-system-interface by inlining some nonce-related code in solana-message. * Run cargo-sort * Remove bincode feature-gating where it's no longer needed * Fixup imports * Remove solana_bincode dep, use bincode directly
* hash!: Unify all hash types, introduce "offchain" feature #### Problem There are many repeated `Hash` types in all of the different hasher crates, which brings in unnecessary dependencies and copied code. #### Summary of changes Unify all hashes under `solana_hash::Hash`, by re-exporting that type from keccak and blake3. While doing this work, I also cleaned up the dependencies, which were unnecessarily bringing in the crate for the offchain implementation. When integrating the SDK with Pinocchio, we'll have many situations in which Pinocchio doesn't want to have an offchain fallback implementation available, but that fallback is still used by other downstream users. To satisfy both uses, this PR also introduces a general `"offchain"` crate feature. For example, `Pubkey::find_program_address` could use the offchain feature for deriving the address without using the syscall, which is useful for most offchain applications. I could also split off the offchain feature in a different PR if that's clearer. * Guard Hasher re-export with not(target_os = "solana") * Fixup CI * Properly use offchain feature * Oops, missing semi-colon * Enable offchain feature in offchain-message * Fix doctests * Improve panic message * Require offchain feature for tests * Revert feature names
* create `solana-program-stubs` crate * create `solana-epoch-stake` crate * Revert "create `solana-program-stubs` crate" This reverts commit ebfa4e0. * make epoch-stake crate on-chain only * update patch script * add black box
* Bump ed25519-dalek to 2.1.1, ed25519-dalek-bip32 to 0.3.0 * unwrap where error is impossible * use from_keypair_bytes * nit: change unwrap to ? Co-authored-by: samkim-crypto <skim13@cs.stanford.edu> * fmt * don't accept slices longer than 64 bytes --------- Co-authored-by: samkim-crypto <skim13@cs.stanford.edu>
* docs: fix memory layout comment in instructions sysvar * rewrite
#### Problem The SDK repo provides libraries, which should allow downstream users some flexibility when choosing what dependencies they want to use. However, we currently pin two dependencies: array-bytes and ed25519-dalek. #### Summary of changes Unpin both of them! array-bytes is only used in tests, so it shouldn't have any impact. ed25519-dalek is definitely more sensitive, but Agave and other downstream users should decide what version they want to use, and we shouldn't force that at the library level.
…nza-xyz#87) * factor out SysvarGet from Sysvar * move Default constraint to SysvarGet * fmt * rename original Sysvar to SysvarSerialize, and SysvarGet to Sysvar * fmt * fix doc links * fmt * fix doctests * remove SysvarSerialize::get
* secp256k1-program: Fix doc tests #### Problem The doctests for secp256k1-program fail when run on the crate individually because of a missing feature declaration on `solana-keccak-hasher`. #### Summary of changes Add the feature. At the same time, run the doctests with `cargo-hack` to catch these in the future. The sdk-wasm-js package fails to run doctests because it doesn't have an rlib target, so add that too. * Use nightly and install hack for doc test job
#### Problem As noted in anza-xyz/agave#7028, the `extend_and_hash` function is unnecessary and creates another allocation, so we inlined it in Agave. #### Summary of changes Deprecate it formally in the crate. This PR will be backported to v2.x and published, so we can remove the function in a follow-up PR.
#### Problem As mentioned in anza-xyz/agave#7028, the `extend_and_hash` function is unnecessary and creates another allocation, so we inlined it in Agave. #### Summary of changes Since it's been properly marked as deprecated, remove it.
remove `once-cell`
* sdk!: Remove `program` feature #### Problem The `program` feature has existed since v1.3 so that people can use solana-sdk in on-chain programs. `cargo-build-sbf` even had a special case where it would enable the `program` feature when building solana-sdk, but it was removed in anza-xyz/agave#5870. This means that our current build-sbf script fails when using v2.3 of the tools. #### Summary of changes Remove the `program` feature from the sdk, and fix the build-sbf script to exclude the sdk, and then try to build without default features. Rather than adding more excludes, I also changes the build-sbf script to use `cargo hack`, which is a bit cleaner. * Exclude solana-sdk
* replace std with core for Error * Also adds rust-version = "1.81.0" for all such crates * Removes the redundant std attribute * Fix CI failures --------- Co-authored-by: Jon C <me@jonc.dev>
wasm-js: Move all implementations into sdk-wasm-js #### Problem As outlined in anza-xyz#118, the sdk incorrectly assumes that all wasm builds are targeting a js environment, which breaks builds in non-js targets. #### Summary of changes The wasm-js code isn't strictly needed to make things work, so instead, move all of the code into different modules. This causes some copy-pasta, but keeps wasm builds lean. If people are using wasm-js, most likely they'll roll their own code anyway. As part of this, since solana-keypair now uses rand 0.8, we can upgrade the getrandom backend to 0.2.
|
If this PR represents a change to the sysvar layout, please open a follow-up PR to update the JavaScript client |
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.
No description provided.