When building examples, the linker fails with multiple definition of entrypoint error because both spl-token and spl-token-2022 define the entrypoint() function by default.
Error:
error: ld returned 1 exit status
multiple definition of 'entrypoint'
spl_token-*.rlib: previous definition here
This SDK is an off-chain client library, not an on-chain program deployed to validators. The entrypoint() function is only needed for on-chain Solana programs. Client libraries only need the types and structs for parsing account data.
Fix: Add no-entrypoint feature to exclude the entrypoint symbol from compilation.
Before:
spl-token = "9.0.0"
spl-token-2022 = "10.0.0"
After:
spl-token = { version = "9.0.0", default-features = false, features = ["no-entrypoint"] }
spl-token-2022 = { version = "10.0.0", default-features = false, features = ["no-entrypoint"] }
PumpFun and PumpSwap have additional buy instruction variants that were not being parsed:
- PumpFun: `buy_exact_sol_in` (discriminator: [56, 252, 116, 8, 158, 223, 205, 95])
- PumpSwap: `buy_exact_quote_in` (discriminator: [198, 46, 21, 82, 180, 217, 232, 112])
These discriminators are from the official PumpFun IDL:
https://github.com/pump-fun/pump-public-docs/tree/main/idl
Without this fix, approximately 4-5% of trades were being missed.
- Upgrade Solana dependencies from 3.0.0 to 3.1.4
- Upgrade solana-hash to 4.0.1
- Update spl-token and spl-token-2022 to latest versions
- Update other dependencies (yellowstone-grpc, rustls, wide, etc.)
- Fix wide library API changes: replace cmp_eq() with simd_eq()
Update package version from 1.1.4 to 1.1.5 in Cargo.toml and documentation.
This release includes recent improvements from merged pull requests:
- Increase minimum accounts check from 11 to 13 for PumpSwap parser
- Refactor event parser core implementation
Add support for parsing SetComputeUnitLimit and SetComputeUnitPrice instructions from Solana's Compute Budget Program. Update event parser to handle compute budget instructions and bump version to 1.1.3.
- Replace instance-based EventParser with static method design
- Introduce ParserCache module for protocol config caching
- Simplify EventProcessor to pure functional handlers
- Optimize metrics: remove min/max tracking, keep last + avg only
- Reduce atomic operations in hot path for better performance
- Convert UnifiedEvent from trait object (Box<dyn UnifiedEvent>) to concrete enum type
- Remove match_event! macro in favor of native match expressions
- Simplify event metadata access: event.event_type() -> event.metadata().event_type
- Unify event callback signatures: Fn(Box<dyn UnifiedEvent>) -> Fn(UnifiedEvent)
- Update all example code to use the new enum-based event system
- Optimize type system to reduce runtime overhead and improve type safety
Affected scope:
- Core event parser and processor
- All protocol event definitions (PumpFun, PumpSwap, Bonk, Raydium series, etc.)
- gRPC and Shred streaming modules
- All example code
modernize dependencies
- Upgrade all Solana SDK dependencies
from v2.3.x to v3.0.0
- Update yellowstone-grpc dependencies
from v8.0.0 to v9.0.0
- Upgrade tonic from v0.12.3 to
v0.14.2 with transport features
- Update prost and prost-types to
v0.14.1
- Add solana-commitment-config and
tonic-prost dependencies
- Migrate from solana-sdk::instruction
to message::compiled_instruction
- Update imports for SPL token pack
compatibility
- Fix type compatibility issues with
new gRPC and protobuf versions
- Regenerate proto files with updated
codegen and server implementations
- Introduce EventSource enum (Grpc, Shred) for better event source management
- Refactor start_block_processing_thread to conditionally spawn threads based on source type
- Update callers to specify event source type
OpenMindTeh