Merged
Conversation
Upgrade existing assert to run in release
When testing ECDSA signatures with a random private key, we were logging the fixed private key. This PR fixes the logging so that we log the randomly generated private key.
Fixes minor OOB issue with RAM / ROM: circuit should not crash due to OOB.
## Summary PR #22061 switched `ci-barretenberg` and `ci-barretenberg-full` from calling `bootstrap.sh ci` (which builds AND runs tests via `parallelize`) to Make targets that only collect test commands into `/tmp/test_cmds` without ever starting the test engine to execute them. This meant BB tests silently stopped running on all `merge-train/barretenberg` PRs. **Fix**: Revert to calling `bootstrap.sh ci` directly. - `ci-barretenberg`: Add explicit CRS download before `barretenberg/cpp/bootstrap.sh ci` (preserving the CRS fix that #22061 was trying to solve) - `ci-barretenberg-full`: Revert to `barretenberg/bootstrap.sh ci` (which already downloads CRS via `bootstrap_all`) ## Test plan - [ ] `ci-barretenberg` label triggers build + test execution (not just build) - [ ] `ci-barretenberg-full` label triggers full build + test execution - [ ] CRS tests pass (the original motivation for #22061) ClaudeBox log: https://claudebox.work/s/0b1cdb0a4d7a7f5a?run=2
- Reject trailing msgpack data in `PrivateExecutionStepRaw::load` and `parse_uncompressed`, with red/green tests in `bbapi.test.cpp` - Add `BB_ASSERT_GTE` guards in `Polynomial::shiftable()` to prevent unsigned underflow when `size < NUM_ZERO_ROWS` - Add `BB_ASSERT` in `scale_by_generator` that `generator_size % num_threads == 0` to prevent silent element skip - Fix `row_disabling_polynomial.hpp` `evaluate_at_challenge` to take `std::span<const FF>` instead of by-value `std::vector<FF>` / non-const `std::span<FF>` - Pass `msg` to assertions in `evaluate_linear_identity` and `evaluate_polynomial_identity` constant paths - Early return in `field_t::invert()` for constant inputs to avoid fragile null context dereference
When the bb helper process crashes, the socket was destroyed but not nulled. This meant subsequent calls to call() would pass the null check and attempt to write to a destroyed socket. While existing error handlers would eventually reject the callbacks, nulling the socket reference makes the guard clause catch dead sockets immediately.
## Summary Defensive hardening of the native socket backend used by `ChonkBatchVerifier`. When the bb helper process crashes (e.g. SIGABRT), the socket was destroyed but the reference was not nulled. This meant subsequent calls to `call()` would pass the `if (!this.socket)` guard and attempt to write to a destroyed socket. While the existing socket error/end handlers would eventually reject callbacks via async error events, nulling the socket makes the guard clause reject immediately — fail-fast rather than relying on async error propagation. Changes: - Null `this.socket` after `destroy()` in process `error` handler - Null `this.socket` after `destroy()` in process `exit` handler - Null `this.socket` after `destroy()` in `cleanup()` method ## Test plan - Existing unit tests should pass — this is a defensive null assignment, no behavior change for healthy paths - On process crash, calls to `call()` now throw `Socket not connected` immediately instead of writing to a destroyed socket and waiting for the async error event ClaudeBox log: https://claudebox.work/s/1164aa0ab601318b?run=5
Any commits made after this event will not be merged.
Collaborator
Author
|
🤖 Auto-merge enabled after 4 hours of inactivity. This PR will be merged automatically once all checks pass. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
BEGIN_COMMIT_OVERRIDE
chore: upgrade uintx assert to run in release (#22289)
chore: Fix log statement in ECDSA test (#22285)
fix: avoid oob access in RAM / ROM and enable soft-fail (#22301)
fix: restore test execution in ci-barretenberg CI modes (#22303)
fix: minor fixes pt. 5 (#22306)
chore: comment about standard behavior for de bruijn (#22310)
fix: null socket on bb process exit/error in native_socket.ts (#22309)
END_COMMIT_OVERRIDE