resolc crate (#328)

- Factor the YUL crate out of `revive-solidity`.
- `revive-solidity` is in reality not a Solidity implementation but the
revive solidity compiler driver (`resolc`). By renaming we not only get
this straight but also a binary with the same name as the crate which
should be less confusing.

---------

Signed-off-by: Cyrill Leutwiler <bigcyrill@hotmail.com>
This commit is contained in:
xermicus
2025-05-27 09:48:43 +02:00
committed by GitHub
parent 090e3ac13c
commit bd4e108bb0
99 changed files with 599 additions and 624 deletions
@@ -0,0 +1,21 @@
//! The contract LLVM IR source code.
use serde::Deserialize;
use serde::Serialize;
/// The contract LLVM IR source code.
#[derive(Debug, Serialize, Deserialize, Clone)]
#[allow(clippy::upper_case_acronyms)]
pub struct LLVMIR {
/// The LLVM IR file path.
pub path: String,
/// The LLVM IR source code.
pub source: String,
}
impl LLVMIR {
/// A shortcut constructor.
pub fn new(path: String, source: String) -> Self {
Self { path, source }
}
}
@@ -0,0 +1,66 @@
//! The contract source code.
pub mod llvm_ir;
pub mod yul;
use std::collections::HashSet;
use serde::Deserialize;
use serde::Serialize;
use revive_yul::parser::statement::object::Object;
use self::llvm_ir::LLVMIR;
use self::yul::Yul;
/// The contract source code.
#[derive(Debug, Serialize, Deserialize, Clone)]
#[allow(clippy::upper_case_acronyms)]
pub enum IR {
/// The Yul source code.
Yul(Yul),
/// The LLVM IR source code.
LLVMIR(LLVMIR),
}
impl IR {
/// A shortcut constructor.
pub fn new_yul(source_code: String, object: Object) -> Self {
Self::Yul(Yul::new(source_code, object))
}
/// A shortcut constructor.
pub fn new_llvm_ir(path: String, source: String) -> Self {
Self::LLVMIR(LLVMIR::new(path, source))
}
/// Get the list of missing deployable libraries.
pub fn get_missing_libraries(&self) -> HashSet<String> {
match self {
Self::Yul(inner) => inner.get_missing_libraries(),
Self::LLVMIR(_inner) => HashSet::new(),
}
}
}
impl<D> revive_llvm_context::PolkaVMWriteLLVM<D> for IR
where
D: revive_llvm_context::PolkaVMDependency + Clone,
{
fn declare(
&mut self,
context: &mut revive_llvm_context::PolkaVMContext<D>,
) -> anyhow::Result<()> {
match self {
Self::Yul(inner) => inner.declare(context),
Self::LLVMIR(_inner) => Ok(()),
}
}
fn into_llvm(self, context: &mut revive_llvm_context::PolkaVMContext<D>) -> anyhow::Result<()> {
match self {
Self::Yul(inner) => inner.into_llvm(context),
Self::LLVMIR(_inner) => Ok(()),
}
}
}
@@ -0,0 +1,48 @@
//! The contract Yul source code.
use std::collections::HashSet;
use serde::Deserialize;
use serde::Serialize;
use revive_yul::parser::statement::object::Object;
/// The contract Yul source code.
#[derive(Debug, Serialize, Deserialize, Clone)]
pub struct Yul {
/// The Yul source code.
pub source_code: String,
/// The Yul AST object.
pub object: Object,
}
impl Yul {
/// A shortcut constructor.
pub fn new(source_code: String, object: Object) -> Self {
Self {
source_code,
object,
}
}
/// Get the list of missing deployable libraries.
pub fn get_missing_libraries(&self) -> HashSet<String> {
self.object.get_missing_libraries()
}
}
impl<D> revive_llvm_context::PolkaVMWriteLLVM<D> for Yul
where
D: revive_llvm_context::PolkaVMDependency + Clone,
{
fn declare(
&mut self,
context: &mut revive_llvm_context::PolkaVMContext<D>,
) -> anyhow::Result<()> {
self.object.declare(context)
}
fn into_llvm(self, context: &mut revive_llvm_context::PolkaVMContext<D>) -> anyhow::Result<()> {
self.object.into_llvm(context)
}
}
@@ -0,0 +1,43 @@
//! The Solidity contract metadata.
use serde::Serialize;
use crate::ResolcVersion;
/// The Solidity contract metadata.
/// Is used to append the metadata hash to the contract bytecode.
#[derive(Debug, Serialize)]
pub struct Metadata {
/// The `solc` metadata.
pub solc_metadata: serde_json::Value,
/// The `solc` version.
pub solc_version: String,
/// The pallet revive edition.
pub revive_pallet_version: Option<semver::Version>,
/// The PolkaVM compiler version.
pub revive_version: String,
/// The PolkaVM compiler optimizer settings.
pub optimizer_settings: revive_llvm_context::OptimizerSettings,
/// The extra LLVM arguments give used for manual control.
pub llvm_arguments: Vec<String>,
}
impl Metadata {
/// A shortcut constructor.
pub fn new(
solc_metadata: serde_json::Value,
solc_version: String,
revive_pallet_version: Option<semver::Version>,
optimizer_settings: revive_llvm_context::OptimizerSettings,
llvm_arguments: Vec<String>,
) -> Self {
Self {
solc_metadata,
solc_version,
revive_pallet_version,
revive_version: ResolcVersion::default().long,
optimizer_settings,
llvm_arguments,
}
}
}
+190
View File
@@ -0,0 +1,190 @@
//! The contract data.
pub mod ir;
pub mod metadata;
use std::collections::HashSet;
use revive_solc_json_interface::SolcStandardJsonInputSettingsPolkaVMMemory;
use serde::Deserialize;
use serde::Serialize;
use sha3::Digest;
use revive_llvm_context::PolkaVMWriteLLVM;
use crate::build::contract::Contract as ContractBuild;
use crate::project::Project;
use crate::solc::version::Version as SolcVersion;
use self::ir::IR;
use self::metadata::Metadata;
/// The contract data.
#[derive(Debug, Serialize, Deserialize, Clone)]
pub struct Contract {
/// The absolute file path.
pub path: String,
/// The IR source code data.
pub ir: IR,
/// The metadata JSON.
pub metadata_json: serde_json::Value,
}
impl Contract {
/// A shortcut constructor.
pub fn new(
path: String,
source_hash: [u8; revive_common::BYTE_LENGTH_WORD],
source_version: SolcVersion,
ir: IR,
metadata_json: Option<serde_json::Value>,
) -> Self {
let metadata_json = metadata_json.unwrap_or_else(|| {
serde_json::json!({
"source_hash": hex::encode(source_hash.as_slice()),
"source_version": serde_json::to_value(&source_version).expect("Always valid"),
})
});
Self {
path,
ir,
metadata_json,
}
}
/// Returns the contract identifier, which is:
/// - the Yul object identifier for Yul
/// - the module name for LLVM IR
pub fn identifier(&self) -> &str {
match self.ir {
IR::Yul(ref yul) => yul.object.identifier.as_str(),
IR::LLVMIR(ref llvm_ir) => llvm_ir.path.as_str(),
}
}
/// Extract factory dependencies.
pub fn drain_factory_dependencies(&mut self) -> HashSet<String> {
match self.ir {
IR::Yul(ref mut yul) => yul.object.factory_dependencies.drain().collect(),
IR::LLVMIR(_) => HashSet::new(),
}
}
/// Compiles the specified contract, setting its build artifacts.
pub fn compile(
mut self,
project: Project,
optimizer_settings: revive_llvm_context::OptimizerSettings,
include_metadata_hash: bool,
debug_config: revive_llvm_context::DebugConfig,
llvm_arguments: &[String],
memory_config: SolcStandardJsonInputSettingsPolkaVMMemory,
) -> anyhow::Result<ContractBuild> {
let llvm = inkwell::context::Context::create();
let optimizer = revive_llvm_context::Optimizer::new(optimizer_settings);
let version = project.version.clone();
let identifier = self.identifier().to_owned();
let metadata = Metadata::new(
self.metadata_json.take(),
version.long.clone(),
version.l2_revision.clone(),
optimizer.settings().to_owned(),
llvm_arguments.to_vec(),
);
let metadata_json = serde_json::to_value(&metadata).expect("Always valid");
let metadata_hash: Option<[u8; revive_common::BYTE_LENGTH_WORD]> = if include_metadata_hash
{
let metadata_string = serde_json::to_string(&metadata).expect("Always valid");
Some(sha3::Keccak256::digest(metadata_string.as_bytes()).into())
} else {
None
};
let module = match self.ir {
IR::LLVMIR(ref llvm_ir) => {
// Create the output module
let memory_buffer =
inkwell::memory_buffer::MemoryBuffer::create_from_memory_range_copy(
llvm_ir.source.as_bytes(),
self.path.as_str(),
);
llvm.create_module_from_ir(memory_buffer)
.map_err(|error| anyhow::anyhow!(error.to_string()))?
}
_ => llvm.create_module(self.path.as_str()),
};
let mut context = revive_llvm_context::PolkaVMContext::new(
&llvm,
module,
optimizer,
Some(project),
include_metadata_hash,
debug_config,
llvm_arguments,
memory_config,
);
context.set_solidity_data(revive_llvm_context::PolkaVMContextSolidityData::default());
match self.ir {
IR::Yul(_) => {
context.set_yul_data(Default::default());
}
IR::LLVMIR(_) => {}
}
let factory_dependencies = self.drain_factory_dependencies();
self.ir.declare(&mut context).map_err(|error| {
anyhow::anyhow!(
"The contract `{}` LLVM IR generator declaration pass error: {}",
self.path,
error
)
})?;
self.ir.into_llvm(&mut context).map_err(|error| {
anyhow::anyhow!(
"The contract `{}` LLVM IR generator definition pass error: {}",
self.path,
error
)
})?;
if let Some(debug_info) = context.debug_info() {
debug_info.finalize_module()
}
let build = context.build(self.path.as_str(), metadata_hash)?;
Ok(ContractBuild::new(
self.path,
identifier,
build,
metadata_json,
factory_dependencies,
))
}
/// Get the list of missing deployable libraries.
pub fn get_missing_libraries(&self) -> HashSet<String> {
self.ir.get_missing_libraries()
}
}
impl<D> PolkaVMWriteLLVM<D> for Contract
where
D: revive_llvm_context::PolkaVMDependency + Clone,
{
fn declare(
&mut self,
context: &mut revive_llvm_context::PolkaVMContext<D>,
) -> anyhow::Result<()> {
self.ir.declare(context)
}
fn into_llvm(self, context: &mut revive_llvm_context::PolkaVMContext<D>) -> anyhow::Result<()> {
self.ir.into_llvm(context)
}
}
+382
View File
@@ -0,0 +1,382 @@
//! The processed input data.
pub mod contract;
use std::collections::BTreeMap;
use std::collections::HashMap;
use std::collections::HashSet;
use std::path::Path;
#[cfg(feature = "parallel")]
use rayon::iter::{IntoParallelIterator, ParallelIterator};
use serde::Deserialize;
use serde::Serialize;
use sha3::Digest;
use revive_solc_json_interface::SolcStandardJsonOutput;
use revive_yul::lexer::Lexer;
use revive_yul::parser::statement::object::Object;
use crate::build::contract::Contract as ContractBuild;
use crate::build::Build;
use crate::missing_libraries::MissingLibraries;
use crate::process::input::Input as ProcessInput;
use crate::process::Process;
use crate::project::contract::ir::IR;
use crate::solc::version::Version as SolcVersion;
use crate::solc::Compiler;
use self::contract::Contract;
/// The processes input data.
#[derive(Debug, Serialize, Deserialize, Clone)]
pub struct Project {
/// The source code version.
pub version: SolcVersion,
/// The project contracts,
pub contracts: BTreeMap<String, Contract>,
/// The mapping of auxiliary identifiers, e.g. Yul object names, to full contract paths.
pub identifier_paths: BTreeMap<String, String>,
/// The library addresses.
pub libraries: BTreeMap<String, BTreeMap<String, String>>,
}
impl Project {
/// A shortcut constructor.
pub fn new(
version: SolcVersion,
contracts: BTreeMap<String, Contract>,
libraries: BTreeMap<String, BTreeMap<String, String>>,
) -> Self {
let mut identifier_paths = BTreeMap::new();
for (path, contract) in contracts.iter() {
identifier_paths.insert(contract.identifier().to_owned(), path.to_owned());
}
Self {
version,
contracts,
identifier_paths,
libraries,
}
}
/// Compiles all contracts, returning their build artifacts.
pub fn compile(
self,
optimizer_settings: revive_llvm_context::OptimizerSettings,
include_metadata_hash: bool,
debug_config: revive_llvm_context::DebugConfig,
llvm_arguments: &[String],
memory_config: revive_solc_json_interface::SolcStandardJsonInputSettingsPolkaVMMemory,
) -> anyhow::Result<Build> {
let project = self.clone();
#[cfg(feature = "parallel")]
let iter = self.contracts.into_par_iter();
#[cfg(not(feature = "parallel"))]
let iter = self.contracts.into_iter();
let results: BTreeMap<String, anyhow::Result<ContractBuild>> = iter
.map(|(full_path, contract)| {
let process_input = ProcessInput::new(
contract,
project.clone(),
include_metadata_hash,
optimizer_settings.clone(),
debug_config.clone(),
llvm_arguments.to_vec(),
memory_config,
);
let process_output = {
#[cfg(target_os = "emscripten")]
{
crate::WorkerProcess::call(process_input)
}
#[cfg(not(target_os = "emscripten"))]
{
crate::NativeProcess::call(process_input)
}
};
(full_path, process_output.map(|output| output.build))
})
.collect();
let mut build = Build::default();
let mut hashes = HashMap::with_capacity(results.len());
for (path, result) in results.iter() {
match result {
Ok(contract) => {
hashes.insert(path.to_owned(), contract.build.bytecode_hash.to_owned());
}
Err(error) => {
anyhow::bail!("Contract `{}` compiling error: {:?}", path, error);
}
}
}
for (path, result) in results.into_iter() {
match result {
Ok(mut contract) => {
for dependency in contract.factory_dependencies.drain() {
let dependency_path = project
.identifier_paths
.get(dependency.as_str())
.cloned()
.unwrap_or_else(|| {
panic!("Dependency `{dependency}` full path not found")
});
let hash = match hashes.get(dependency_path.as_str()) {
Some(hash) => hash.to_owned(),
None => anyhow::bail!(
"Dependency contract `{}` not found in the project",
dependency_path
),
};
contract
.build
.factory_dependencies
.insert(hash, dependency_path);
}
build.contracts.insert(path, contract);
}
Err(error) => {
anyhow::bail!("Contract `{}` compiling error: {:?}", path, error);
}
}
}
Ok(build)
}
/// Get the list of missing deployable libraries.
pub fn get_missing_libraries(&self) -> MissingLibraries {
let deployed_libraries = self
.libraries
.iter()
.flat_map(|(file, names)| {
names
.iter()
.map(|(name, _address)| format!("{file}:{name}"))
.collect::<HashSet<String>>()
})
.collect::<HashSet<String>>();
let mut missing_deployable_libraries = BTreeMap::new();
for (contract_path, contract) in self.contracts.iter() {
let missing_libraries = contract
.get_missing_libraries()
.into_iter()
.filter(|library| !deployed_libraries.contains(library))
.collect::<HashSet<String>>();
missing_deployable_libraries.insert(contract_path.to_owned(), missing_libraries);
}
MissingLibraries::new(missing_deployable_libraries)
}
/// Parses the Yul source code file and returns the source data.
pub fn try_from_yul_path<T: Compiler>(
path: &Path,
solc_validator: Option<&T>,
) -> anyhow::Result<Self> {
let source_code = std::fs::read_to_string(path)
.map_err(|error| anyhow::anyhow!("Yul file {:?} reading error: {}", path, error))?;
Self::try_from_yul_string(path, source_code.as_str(), solc_validator)
}
/// Parses the test Yul source code string and returns the source data.
/// Only for integration testing purposes.
pub fn try_from_yul_string<T: Compiler>(
path: &Path,
source_code: &str,
solc_validator: Option<&T>,
) -> anyhow::Result<Self> {
if let Some(solc) = solc_validator {
solc.validate_yul(path)?;
}
let source_version = SolcVersion::new_simple(crate::solc::LAST_SUPPORTED_VERSION);
let path = path.to_string_lossy().to_string();
let source_hash = sha3::Keccak256::digest(source_code.as_bytes()).into();
let mut lexer = Lexer::new(source_code.to_owned());
let object = Object::parse(&mut lexer, None)
.map_err(|error| anyhow::anyhow!("Yul object `{}` parsing error: {}", path, error))?;
let mut project_contracts = BTreeMap::new();
project_contracts.insert(
path.to_owned(),
Contract::new(
path,
source_hash,
source_version.clone(),
IR::new_yul(source_code.to_owned(), object),
None,
),
);
Ok(Self::new(
source_version,
project_contracts,
BTreeMap::new(),
))
}
/// Parses the LLVM IR source code file and returns the source data.
pub fn try_from_llvm_ir_path(path: &Path) -> anyhow::Result<Self> {
let source_code = std::fs::read_to_string(path)
.map_err(|error| anyhow::anyhow!("LLVM IR file {:?} reading error: {}", path, error))?;
let source_hash = sha3::Keccak256::digest(source_code.as_bytes()).into();
let source_version =
SolcVersion::new_simple(revive_llvm_context::polkavm_const::LLVM_VERSION);
let path = path.to_string_lossy().to_string();
let mut project_contracts = BTreeMap::new();
project_contracts.insert(
path.clone(),
Contract::new(
path.clone(),
source_hash,
source_version.clone(),
IR::new_llvm_ir(path, source_code),
None,
),
);
Ok(Self::new(
source_version,
project_contracts,
BTreeMap::new(),
))
}
/// Converts the `solc` JSON output into a convenient project.
pub fn try_from_standard_json_output(
output: &SolcStandardJsonOutput,
source_code_files: BTreeMap<String, String>,
libraries: BTreeMap<String, BTreeMap<String, String>>,
solc_version: &SolcVersion,
debug_config: &revive_llvm_context::DebugConfig,
) -> anyhow::Result<Self> {
let files = match output.contracts.as_ref() {
Some(files) => files,
None => match &output.errors {
Some(errors) if errors.iter().any(|e| e.severity == "error") => {
anyhow::bail!(serde_json::to_string_pretty(errors).expect("Always valid"));
}
_ => &BTreeMap::new(),
},
};
let mut project_contracts = BTreeMap::new();
for (path, contracts) in files.iter() {
for (name, contract) in contracts.iter() {
let full_path = format!("{path}:{name}");
let ir_optimized = match contract.ir_optimized.to_owned() {
Some(ir_optimized) => ir_optimized,
None => continue,
};
if ir_optimized.is_empty() {
continue;
}
debug_config.dump_yul(full_path.as_str(), ir_optimized.as_str())?;
let mut lexer = Lexer::new(ir_optimized.to_owned());
let object = Object::parse(&mut lexer, None).map_err(|error| {
anyhow::anyhow!("Contract `{}` parsing error: {:?}", full_path, error)
})?;
let source = IR::new_yul(ir_optimized.to_owned(), object);
let source_code = source_code_files
.get(path.as_str())
.ok_or_else(|| anyhow::anyhow!("Source code for path `{}` not found", path))?;
let source_hash = sha3::Keccak256::digest(source_code.as_bytes()).into();
let project_contract = Contract::new(
full_path.clone(),
source_hash,
solc_version.to_owned(),
source,
contract.metadata.to_owned(),
);
project_contracts.insert(full_path, project_contract);
}
}
Ok(Project::new(
solc_version.to_owned(),
project_contracts,
libraries,
))
}
}
impl revive_llvm_context::PolkaVMDependency for Project {
fn compile(
project: Self,
identifier: &str,
optimizer_settings: revive_llvm_context::OptimizerSettings,
include_metadata_hash: bool,
debug_config: revive_llvm_context::DebugConfig,
llvm_arguments: &[String],
memory_config: revive_solc_json_interface::SolcStandardJsonInputSettingsPolkaVMMemory,
) -> anyhow::Result<String> {
let contract_path = project.resolve_path(identifier)?;
let contract = project
.contracts
.get(contract_path.as_str())
.cloned()
.ok_or_else(|| {
anyhow::anyhow!(
"Dependency contract `{}` not found in the project",
contract_path
)
})?;
contract
.compile(
project,
optimizer_settings,
include_metadata_hash,
debug_config,
llvm_arguments,
memory_config,
)
.map_err(|error| {
anyhow::anyhow!(
"Dependency contract `{}` compiling error: {}",
identifier,
error
)
})
.map(|contract| contract.build.bytecode_hash)
}
fn resolve_path(&self, identifier: &str) -> anyhow::Result<String> {
self.identifier_paths
.get(identifier.strip_suffix("_deployed").unwrap_or(identifier))
.cloned()
.ok_or_else(|| {
anyhow::anyhow!(
"Contract with identifier `{}` not found in the project",
identifier
)
})
}
fn resolve_library(&self, path: &str) -> anyhow::Result<String> {
for (file_path, contracts) in self.libraries.iter() {
for (contract_name, address) in contracts.iter() {
let key = format!("{file_path}:{contract_name}");
if key.as_str() == path {
return Ok(address["0x".len()..].to_owned());
}
}
}
anyhow::bail!("Library `{}` not found in the project", path);
}
}