fix: resolve false positive constructor arguments

[zachdaniel]

Resolves #1873

Changelog

Bug Fixes

• We only want to take everything up to the swarm hash.

• I added an entry to CHANGELOG.md with this PR
• If I added new functionality, I added tests covering it.
• If I fixed a bug, I added a regression test to prevent the bug from silently reappearing again.
• I checked whether I should update the docs and did so if necessary

[goodsoft]

Are you really sure there can't be any false positives after the 0029 we need?
If the constructor arguments are there, are you sure they can't contain 0029 themselves?
Or, even, x0 02 9x, which would match as well?

[coveralls]

Pull Request Test Coverage Report for Build 25fc482d-5f11-4b03-bb82-c87979d3fcfb

• 11 of 13 (84.62%) changed or added relevant lines in 3 files are covered.
• 7 unchanged lines in 4 files lost coverage.
• Overall coverage decreased (-0.04%) to 81.947%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
apps/explorer/lib/explorer/smart_contract/verifier.ex	5	6	83.33%
apps/explorer/lib/explorer/smart_contract/verifier/constructor_arguments.ex	2	3	66.67%

Files with Coverage Reduction	New Missed Lines	%
apps/explorer/lib/explorer/smart_contract/solidity/code_compiler.ex	1	95.83%
apps/indexer/lib/indexer/fetcher/token.ex	1	78.57%
apps/indexer/lib/indexer/fetcher/token_balance.ex	2	87.1%
apps/indexer/lib/indexer/block/fetcher.ex	3	86.57%

Totals
Change from base Build ef7c3a4c-fe72-4c65-a1d9-4d1f537e9755:	-0.04%
Covered Lines:	4630
Relevant Lines:	5650

---

💛 - Coveralls

[acravenho]

@zachdaniel @goodsoft Here is an example of a false positive where we would run into an issue: https://etherscan.io/address/0x64d14595152b430cf6940da15c6e39545c7c5b7e#code

Please see the constructor arguments and the 0029.

[zachdaniel]

@goodsoft definitely not really sure. I only just started looking into it this morning, and had a similar thought myself, but I figured I would get the bare bones solution for that specific contract up in front of people.

[zachdaniel]

Okay, so the swarm source starts with a known value, I think. So perhaps what we do is skip to the swarm source and split there?

[zachdaniel]

https://solidity.readthedocs.io/en/v0.4.21/metadata.html#encoding-of-the-metadata-hash-in-the-bytecode

[acravenho]

Yes, that's correct. Are you able to identify the swarm source without the 0029?

[acravenho]

Nice! That should work.

[zachdaniel]

We could use the length + bzzr0 bytes? Although I think its implying that the mapping itself might change at anytime.

[zachdaniel]

I think it looks reliable enough for now. If they "support other ways to retrieve contract metadata in the future", this can be updated.

[acravenho]

{a165627a7a72305820} - {32 bytes} - {0029} - {constructor arguments}

[acravenho]

Comparing now against these addresses:

• 0xada931ac30aff9e51148d6aaf565fc89e7e90bc0 ✅
• 0x8f639224f992e466ecdffc3f5880194a73f1d93b ✅
• 0x1155c5605d08585594a428f473c77eeb33f28bc1 ✅
• 0x71320e08c267746616456aeca4efde73ed98caf3 ❌
• 0x99831db28f7bb7c03a1a6c517a54ac3e80da2be1 ✅
• 0x7c6ef33ab884862dadf412f7271d669fd3cb8875 ✅
• 0x5152345f9722bc1bef5fd4ec0428946d5eb18c82 ✅
• 0xe241db15493adfe3c49884255f4edf012389bcc8 ✅
• 0x61fe6d3692abdb010347d89d7ec391b719115519 ✅
• 0xbf67ebc297685863ae7e0e6f88eaffe8052a0d3b ✅
• 0x74dd3ee13dc6714de5670198f6e5cd586690b79a ❌
• 0x348c19f7de78acf3c89c7a5fd67b3f4b515bf69a ✅
• 0x2020fe35fc3a2ad06073b5083917b284f927f7b3 ✅

[acravenho]

The two contracts that failed above don't seem to have a swarm source hash. Looks like they were created prior to when swarm source was created.

[zachdaniel]

Do they also have constructor arguments? If so, then we'll need another heuristic for those. I'm almost done with the code to use the swarm hash.

[acravenho]

No constructor arguments. They fail when trying to verify them but probably for some other reason. Looks like an edge case.

[zachdaniel]

Well for now, if we don't find a swarm hash I'll have it use the whole thing.

[zachdaniel]

It was asserting that part of the swarm metadata hash would come back, but that would be incorrect.

[zachdaniel]

Alright @vbaranov @acravenho @goodsoft this should be ready to go for real now.

[acravenho]

LGTM 👍

[ayrat555]

what is a165627a7a72305820?

[zachdaniel]

It’s what the swarm hash starts with.