fix: bound batch failure revert data (N-01)

[md0x]

Bound TryMulticall failure-data copying and hashing to the first 256 bytes. A child returning a large revert payload no longer forces the parent to allocate and hash the entire payload. Successful return data is discarded; event documentation now specifies prefix-hash semantics.

Base: #77. Tracks FRO-151 (N-01).

Validation: all 52 SignedProposer unit tests pass. The new regression fails on the base and passes with the fix: a successful proposal survives a subsequent 512 KiB revert under a 2M-gas batch limit. Failure tests use nonzero patterned returndata and verify both the selector and the first-256-byte hash. Formatting and diff checks pass.

This bounds failure-metadata processing; existing child gas-starvation and insufficient outer-gas limitations still apply.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-09-14T08:44:57.955697Z	71d608e	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[linear]

FRO-151

[chrismaree]

Reviewed against N-01 / FRO-151 at da6cb0d. The self-delegatecall has a zero-length output area, and failure processing copies only min(returndatasize(), 256) before hashing. This closes the unbounded parent-frame returndata copy/hash path while preserving caller context and child rollback. The temporary memory at the free-memory pointer is consumed within the assembly block; only the selector/hash escape. Empty/short payload handling and successful returndata discard are consistent with the documented behavior.

The patterned-prefix assertions address the earlier feedback, and the limited-gas regression checks that an earlier proposal survives a later 512 KiB revert. Existing child gas-starvation/insufficient outer-gas limits remain explicitly documented. Static code and test review with Solidity 0.8.30 memory-safety documentation; no tests or builds run.

_{Chris's Codex agent · automated}