Support for Factory Contract Dependency in forge create #162

soul022 · 2025-06-26T15:14:53Z

Motivation

Solution

This change enhances forge create on Substrate-based networks by automatically detecting, encoding, and uploading any factory-declared child contracts before instantiating the main contract:

New dependencies
Adds hex = "0.4.3" and parity-scale-codec (with derive) in crates/forge/Cargo.toml to support SCALE encoding and hex utilities.

find_contract_by_hash helper
Implements a utility in create.rs that scans compiled artifacts, computes each bytecode’s Keccak-256 hash, and returns the matching raw bytes to locate dependencies at runtime.

handle_factory_dependencies flow
Introduces an async routine in create.rs that:

Reads each artifact’s factoryDependencies map.
Filters out proxy “child” contracts.
ABI-encodes each dependency’s bytecode and a Compact storage limit, prefixes with the EIP-1167 marker, and uploads via the Alloy RPC.
Ensures all dependencies are on-chain before deploying the main contract.

Cleanup & lockfile update
Imports Compact, Encode, and BTreeMap, removes unused code, and bumps Cargo.lock for the new crates.

With this PR, users can simply declare child-contract dependencies in their artifacts and rely on forge create to handle the full upload and instantiation sequence automatically.

PR Checklist

Added Tests
Added Documentation
Breaking changes

Test Evidance

Contract Tested -
https://gist.github.com/soul022/268ff439b8381bb02839161121066b0a

Logs -
https://gist.github.com/soul022/63b440a61e442ff57081132afdd2da09

closes: #130

filip-parity · 2025-07-07T06:43:13Z

High level comments:

The pull requests mentions that tests were added, but no test code is included in the diff...
The find_initialized_contracts_with_paths function has a good docstring, but other new functions (e.g., find_contract_initializations, upload_child_contract_alloy) lack documentation.

smiasojed · 2025-07-07T07:04:39Z

Thank you, @soul022, for this pull request.

Could you please link this PR to the relevant GitHub issue? You can do this by adding closes: #<issue_number> in the description.

Regarding the parsing of factory dependencies from Solidity code: The idea from the ticket description was that this information would be available directly in the resolc compilation output in factoryDependencies field. Could you clarify why parsing it from the Solidity code is necessary in this context?

crates/forge/src/cmd/create.rs

soul022 · 2025-07-07T10:38:17Z

closes: #130

…utput

soul022 · 2025-07-07T17:34:45Z

Thank you, @soul022, for this pull request.

Could you please link this PR to the relevant GitHub issue? You can do this by adding closes: #<issue_number> in the description.

Regarding the parsing of factory dependencies from Solidity code: The idea from the ticket description was that this information would be available directly in the resolc compilation output in factoryDependencies field. Could you clarify why parsing it from the Solidity code is necessary in this context?

Linked it to issue.
Changed code to read bytecode from compiler output, instead of parsing it out, please check.

soul022 · 2025-07-07T17:40:26Z

upload_child_contract_alloy

added comments to upload_child_contract_alloy, removed find_contract_initializations based on new logic

smiasojed · 2025-07-07T18:17:28Z

crates/cli/src/utils/cmd.rs

+}
+
+/// Get child contracts from compiled output
+pub fn get_child_contracts(


In ResolcContract we have
pub factory_dependencies: Option<BTreeMap<String, String>>,
Is there any reason why you do not use it?

There is no ResolcContract in workspace.
Also, Since I am already getting child contracts in pre-existing output object from compiler in
"let output: foundry_compilers::ProjectCompileOutput =
compile::compile_target(&target_path, &project, shell::is_json())?;"
I used it in our usecase

I have just looping through the artifacts from the compiled output to get the child contracts. Attaching below sample artifact array that we get from output, it has both child and parent contract details, get_child_contracts just extracts the child contracts from existing output -
https://gist.github.com/soul022/8f8dc0ef4c0811b05d382d7668f07e40

How will foundry behave when you compile the following code:

// SPDX-License-Identifier: GPL-3.0 pragma solidity >=0.7.0 <0.9.0; library Assert { function equal(uint256 a, uint256 b) internal pure returns (bool result) { result = (a == b); } } contract TestAssert { function checkEquality(uint256 a, uint256 b) public pure returns (string memory) { Assert.equal(a, b); return "Values are equal"; } }

Will foundry upload the Assert library as a child contract?

What about the case where deploy time linking is being used - will your solution still work?

Since this is internal library, the compiler simply embeds its code directly into your contract. There’s no separate library contract on-chain and nothing extra to deploy or link at runtime.

My code to get child contracts was treating it as a contract and uploading it, I have modified code in get_child_contracts and added a check to filter the same, so it does not gets uploaded

for deploy time linking i.e external library, it should work as is, As per foundry book -we need to deploy external library first and use linker options (https://paritytech.github.io/foundry-book-polkadot/reference/forge/forge-create.html#linker-options) to create the contract using that external library.

I tried testing linker option for forge create for below contract, but it failed. seems like linker option is not working in foundry-polkadot.

contract - https://gist.github.com/soul022/9ef77e6ac0cf10365b8949fd9bc310fa
logs - https://gist.github.com/soul022/eccb3fc929953e207adafcfd62cba1c9

My point is that the logic you're implementing already exists in the Revive compiler. I think we should find a way to make factory_dependencies available.
In the near future, we’ll probably also need to expose unlinked_dependencies for deploy time linking. WDYT?

@pkhry do you have any opinion on this? You have been looking into the linking already.

just add a field to solc contract artifact in foundry-compilers. factory-deps are already present, you just need to forward them. unlinked deps should probably forwarded the same way as a field on solc::Contract.

something like this below a la how metadata is currently forwarded:

struct Contract { ... #[serde(flatten)] auxiliary: Option<BTreeMap<String, serde_json::Value>>,

also yeah, not a fan of doing something manually if it's already done by the compiler.

smiasojed · 2025-07-07T18:48:19Z

crates/forge/src/cmd/create.rs

+                                    .clone()
+                                    .ok_or_eyre("Private key not provided")?;
+
+                                let tx_hash = upload_child_contract_alloy(


Should we first check if the contract is already present on-chain?

Only possible way that I am aware to check if contract is uploaded/exists is to check if storage deposit limit value is 0 using revive api dynamically, there is no possible way to do it right now as it needs substrate node url.

I think the process is already optimized, it won't take up any funds if we try to upload already existing/uploaded contract

@pgherveou Do you plan to allow for such a check, or is the current approach good enough in your opinion?

We should, you would still pay for a transaction that does nothing otherwise
let me look into it

sorry didn't look at this, I think you could check that the code exist using eth_getCode first, that should fix it

@pgherveou , seems like when uploading via magic address, it is not returning contract address, without it we cannot call eth_getCode.

{
"status": "0x1",
"cumulativeGasUsed": "0x0",
"logs": [],
"logsBloom": "0x00000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000",
"type": "0x2",
"transactionHash": "0x2889879e6939bc6a7fde7eb2392329961a9681feb4b60b0ccc3a358f01315bf8",
"transactionIndex": "0x2",
"blockHash": "0xa0f9fd2baca909f5839587037a11203b9831b5401dcb59fdadb61fb406178904",
"blockNumber": "0xa03df",
"gasUsed": "0x57c8d1df958a",
"effectiveGasPrice": "0x4b0",
"from": "0xf24ff3a9cf04c71dbc94d0b566f7a27b94566cac",
"to": "0x6d6f646c70792f70616464720000000000000000",
"contractAddress": null
}

my bad I thought you had an address already, yeah this just upload the code without instantiating it

@smiasojed @pgherveou, based on the last discussion, to avoid extra transactions for already uploaded bytecode, we probably need some precompile/check, which would determine if the given bytecode is already present on-chain. For now, @pgherveou suggested to move ahead and this can be picked sooner.

Mostly this would be a blocker to merge it, as it incurs extra cost for user

pkhry · 2025-07-08T20:11:46Z

crates/forge/src/cmd/create.rs

+        let output: foundry_compilers::ProjectCompileOutput =
+            compile::compile_target(&target_path, &project, shell::is_json())?;
+
+        match get_child_contracts(output.clone(), &self.contract.name) {


this introduces new codepath even for solidity contracts, ie should break compat between resolc and non resolc versions.

…racts

smiasojed · 2025-09-19T10:07:31Z

crates/forge/src/cmd/create.rs

+            let scaled_encoded_bytes = bytecode.encode();
+            let storage_deposit_limit = Compact(10000000000u128);
+            let encoded_storage_deposit_limit = storage_deposit_limit.encode();
+            let combined_hex = "0x3c04".to_string() +


It should be based on node metadata. Pallet index may be different between runtimes

smiasojed · 2025-09-19T10:07:56Z

crates/forge/src/cmd/create.rs

+        if let Some(bytecode) = bytecode {
+            // Skip child contracts (those with 0x3c04 prefix)
+            let bytecode_slice = bytecode.as_ref();
+            if bytecode_slice.len() >= 2 && bytecode_slice[0] == 0x3c && bytecode_slice[1] == 0x04 {


Why do we need this check?

smiasojed · 2025-09-19T10:11:46Z

TODO:

use subxt with node metadata
get metadata through eth-proxy - API missing
add check if code is present on chain
get deposit limit instead hardcoding it

soul022 added 9 commits June 26, 2025 20:40

wip

f96402d

[issue 130] - clean up

c01f6b2

[issue 130] - fix fmt issues

b6eb1e3

[issue 130] - fix fmt issues

c3be5c8

[issue - 130] - fix clippy issues

f25cf41

Merge branch 'master' into soul022/factoryContracts

869de53

[issue - 130] - fix fmt issues

56bd04b

[issue - 130] - fix fmt issues

850206f

[issue - 130] - fix clippy issues

0efe3a4

soul022 changed the title ~~[WIP] - Factory Dependecy fix~~ Support for Factory Contract Dependency in forge create Jul 6, 2025

soul022 linked an issue Jul 6, 2025 that may be closed by this pull request

Factory contracts deployment support #130

Open

soul022 marked this pull request as ready for review July 6, 2025 16:59

smiasojed reviewed Jul 7, 2025

View reviewed changes

crates/forge/src/cmd/create.rs Outdated Show resolved Hide resolved

soul022 added 2 commits July 7, 2025 22:43

[issue - 130] - updated code to get bytecode contract from compiler o…

4543076

…utput

[issue - 130] - rustfmt fixes

b6328da

[issue - 130] - added comments to upload_child_contract_alloy

8b65934

[issue - 130] - rustfmt fixes

cccbe25

smiasojed reviewed Jul 7, 2025

View reviewed changes

soul022 added 2 commits July 8, 2025 23:30

[issue - 130] - remove internal libraries used as child contracts

bb249c9

[issue - 130] - clippy fix

0552ce5

pkhry reviewed Jul 8, 2025

View reviewed changes

soul022 added 3 commits July 13, 2025 23:41

[issue - 130] - reading factory contract from compilers

64fa04d

[issue - 130] - remove unused code

3411a24

[issue - 130] - fix fmt issues

af8036c

soul022 and others added 10 commits July 13, 2025 23:58

[issue - 130] - fix fmt issues

053f1e4

[issue - 130] - fix fmt issues

145f6da

[issue-130] - factory dependency changed to btree

ff4e4d1

[issue - 130] - extensions changes

90d9554

[issue - 130] - extensions fixes

1c8fe0a

[issue - 130] - fmt fixes

4c53953

Merge branch 'master' into soul022/factoryContracts

f06ccf9

Merge remote-tracking branch 'origin/master' into soul022/factoryCont…

05fbbcc

…racts

Fmt

2dc0615

use factory deps when revive build is enabled

c36e11c

smiasojed reviewed Sep 19, 2025

View reviewed changes

Add forge create tests

c58e525

Support for Factory Contract Dependency in forge create #162

Are you sure you want to change the base?

Support for Factory Contract Dependency in forge create #162

Conversation

soul022 commented Jun 26, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Motivation

Solution

PR Checklist

Test Evidance

Uh oh!

filip-parity commented Jul 7, 2025

Uh oh!

smiasojed commented Jul 7, 2025

Uh oh!

Uh oh!

soul022 commented Jul 7, 2025

Uh oh!

soul022 commented Jul 7, 2025

Uh oh!

soul022 commented Jul 7, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

soul022 Jul 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

soul022 Jul 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

smiasojed Jul 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

soul022 Jul 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

smiasojed Sep 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

smiasojed commented Sep 19, 2025

Uh oh!

Uh oh!

soul022 commented Jun 26, 2025 •

edited

Loading

soul022 Jul 8, 2025 •

edited

Loading

soul022 Jul 8, 2025 •

edited

Loading

smiasojed Jul 8, 2025 •

edited

Loading

soul022 Jul 8, 2025 •

edited

Loading

smiasojed Sep 19, 2025 •

edited

Loading