[replay] Add transactions to skip for V2 loader

[georgemitenkov]

Description

There are well-known differences between existing loader and new loader & code cache, see here: https://docs.google.com/spreadsheets/d/16IVAe72g3Rcf3LqMKaj80a0Isa8ma3SaJFMVzOtVOfI/edit?usp=sharing. We should skip those.

Replay run: https://github.com/aptos-labs/aptos-core/actions/runs/11727539596

How Has This Been Tested?

Key Areas to Review

Type of Change

• New feature
• Bug fix
• Breaking change
• Performance improvement
• Refactoring
• Dependency update
• Documentation update
• Tests

Which Components or Systems Does This Change Impact?

• Validator Node
• Full Node (API, Indexer, etc.)
• Move/Aptos Virtual Machine
• Aptos Framework
• Aptos CLI/SDK
• Developer Infrastructure
• Move Compiler
• Other (specify)

Checklist

• I have read and followed the CONTRIBUTING doc
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I identified and added all stakeholders and component owners affected by this change as reviewers
• I tested both happy and unhappy path of the functionality
• I have made corresponding changes to the documentation

[trunk-io]

⏱️ 275h 39m total CI duration on this PR

Slowest 15 Jobs	Cumulative Duration	Recent Runs
replay-mainnet / replay-verify (11)	4h 48m	🟩 🟩 🟩
replay-mainnet / replay-verify (33)	4h 46m	🟩 🟥 🟩
replay-mainnet / replay-verify (12)	4h 45m	🟩 🟩 🟩
replay-mainnet / replay-verify (35)	4h 45m	⬜ 🟩 🟩
replay-mainnet / replay-verify (18)	4h 45m	🟩 🟩 🟩
replay-mainnet / replay-verify (36)	4h 44m	🟩 🟥 🟩
replay-mainnet / replay-verify (24)	4h 43m	🟥 🟩 🟩
replay-mainnet / replay-verify (17)	4h 41m	🟩 🟩 🟩
replay-mainnet / replay-verify (10)	4h 41m	🟥 🟩 🟩
replay-mainnet / replay-verify (13)	4h 40m	🟥 🟥 🟩
replay-testnet / replay-verify (17)	4h 37m	🟥 🟩 🟩
replay-mainnet / replay-verify (32)	4h 37m	🟩 🟩 🟩
replay-testnet / replay-verify (12)	4h 37m	🟥 🟩 🟩
replay-mainnet / replay-verify (21)	4h 36m	🟥 🟩 🟩
replay-mainnet / replay-verify (14)	4h 33m	🟩 🟩 🟩

_{settings ⋅ feedback ⋅ docs ⋅ learn more about trunk.io}

[georgemitenkov]

• [replay] Add transactions to skip for V2 loader #15207 👈 (View in Graphite)
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

[igor-aptos]

@sherry-x are we fine with skipping transactions on mainnet?

@georgemitenkov how recent are these transactions? If leader v2 is onchain flag - what do these represent?

[georgemitenkov]

For this range, the first skip is 27.03.24 and the last is basically now: 17.10.24. The reason why we want to skip is that we can completely remove old V1 code then, and the trade off few dozen of transactions that publish code seems good.

See the spreadsheet for actual divergences, those are only due to bugs in old loader (went to through them one by one). For instance, we were flushing module cache and losing error messages, there are a number of cases where the difference is None vs Some(msg) in VMStatus. There are also 2 other cases: 1) init_module not linking correctly before, 2) not topologically sorted bundles not linking correctly and failing to publish.

Ideally, I would not skip but specify the difference in write set, but we cannot do it with the current replay, hence skipping and documenting in a spreadsheet what is the diff.