Compiling with Hardhat generates two files per compiled contract (not each
.sol file): an artifact and a debug file.
An artifact has all the information that is necessary to deploy and interact with the contract. These are compatible with most tools, including Truffle's artifact format. Each artifact consists of a json with the following properties:
contractName: A string with the contract's name.
"0x"-prefixed hex string of the unlinked deployment bytecode. If the contract is not deployable, this has the string
"0x"-prefixed hex string of the unlinked runtime/deployed bytecode. If the contract is not deployable, this has the string
linkReferences: The bytecode's link references object as returned by solc. If the contract doesn't need to be linked, this value contains an empty object.
deployedLinkReferences: The deployed bytecode's link references object as returned by solc. If the contract doesn't need to be linked, this value contains an empty object.
The debug file has all the information that is necessary to reproduce the compilation and to debug the contracts: this includes the original solc input and output, and the solc version used to compile it.
Hardhat optimizes compilation by compiling the smallest possible set of files at a time. Files that are compiled together have the same solc input and output. Since having this in each debug file would be meaningfully wasteful, this information is deduplicated in build info files that are placed in
artifacts/build-info. Each contract debug file contains a relative path to its build info file, and each build info file contains the solc input, solc output and the solc version used.
You shouldn't interact with these files directly.
The [HRE] has an
artifacts object with helper methods. For example, you can get a list with the paths to all artifacts by calling
You can also read an artifact using the name of the contract by calling
hre.artifacts.readArtifact("Bar"), which will return the content of the artifact for the
Bar contract. This would only work if there was just one contract named
Bar in the whole project; it would throw an error if there were two. To disambiguate this case, you would have to use the Fully Qualified Name of the contract:
artifacts/ directory has a structure that follows the original directory structure of the contracts. For example, if your contracts look like this:
contracts ├── Foo.sol ├── Bar.sol └── Qux.sol
then the structure of your artifact directory would look like this:
artifacts └── contracts ├── Foo.sol │ ├── Foo.json │ ├── Foo.dbg.json │ ├── Foo2.json │ └── Foo2.dbg.json ├── Bar.sol │ ├── Bar.json │ └── Bar.dbg.json └── Qux.sol ├── Foo.json └── Foo.dbg.json
Each Solidity file in your source will get a directory in the artifacts structure. Each of these directories contains one artifact (
.json) file and one debug (
.dbg.json) file for each contract in that file.
Foo.sol, for example, contains two contracts inside.
Two Solidity files can have contracts with the same name, and this structure allows for that.