morluto/rea

Fork 0

mirror of https://github.com/morluto/rea.git synced 2026-11-05 06:45:22 +00:00

[PR #3776] [MERGED] fix(cutter): preserve mixed Unicode output within the frame limit #5893

New issue

Closed

opened 2026-10-31 01:10:14 +00:00 by bsakaguchi · 0 comments

bsakaguchi commented

2026-10-31 01:10:14 +00:00

Owner

📋 Pull Request Information

Original PR: https://github.com/morluto/rea/pull/3776
Author: @AyoubElleuch
Created: 10/15/2026
Status: ✅ Merged
Merged: 10/17/2026
Merged by: @N0zoM1z0

Base: main ← Head: fix/cutter-mixed-unicode-output

📝 Commits (7)

21ad0bc fix(cutter): preserve mixed Unicode output within the frame limit
61ce038 Merge branch 'main' into fix/cutter-mixed-unicode-output
2cb4afc Merge branch 'main' into fix/cutter-mixed-unicode-output
757e7aa fix(cutter): preserve completion for unencodable output
51c34ab Merge branch 'main' into fix/cutter-mixed-unicode-output
35f5226 Merge branch 'main' into fix/cutter-mixed-unicode-output
31eb36d Merge branch 'main' into fix/cutter-mixed-unicode-output

📊 Changes

2 files changed (+59 additions, -5 deletions)

View changed files

📝 bridge/cutter/rea_bridge.py (+13 -5)
📝 tests/fixtures/cutter_bridge_protocol_test.py (+46 -0)

📄 Description

Fixes #3330.

Problem and change

A completed 5 MiB ASCII Cutter result fits the authenticated 16 MiB frame, but a single non-ASCII suffix made the old preflight charge four bytes for every character and omit it. The producer now counts actual UTF-8 bytes in 65,536-character chunks, rejects an obviously oversized character count early, and retains the final encoded-frame check and HMAC as authoritative. Each temporary conversion is at most 256 KiB.

The follow-up also preserves the existing authenticated omission path for unencodable surrogate output: preflight returns false on UnicodeEncodeError, and the sender reports completed execution with output=null and output_truncated=true. Framing, completion and caller choices keep one owner; consumers need no repair.

Scope and contracts

Only bridge/cutter/rea_bridge.py and its existing Python protocol fixture change. Existing CLI/MCP schemas, command authority, target selection, framing and cleanup contracts remain intact. MCP may still return its explicit smaller result-budget error with a retained Evidence reference; export recovers the complete output. Ordinary correction within an existing provider; no new capability or maintainer scope decision required.

Verification

Current head: 31eb36d3b3ce7558b423b526e5fea41d7a9497dd, incorporating main 5b243993872c2bcc70cb00060ffb79b347321370. Exact merge-tree and stable patch-ID checks confirm the feature patch is unchanged. Complete exact-head CI passes all five built-dependent shards, Source/static checks and required gates.

Producer regressions cover ASCII/é/雪/😀, exact decoded equality, HMAC, the wire limit and oversized rejection. Unpaired high/low surrogate cases in the first/last chunks go through the actual sender/socketpair and verify truthful authenticated completion. The regression fails on the preceding PR implementation and passes with the structural fix.
Earlier support revision 757e7aa85a passed all 15 Python protocol and 36 production bridge/client boundary cases, production build and complete check:ci. These local suite receipts remain attributed to that revision.
Fresh exact-head Linux production build and real official free Cutter 2.5.0-HEAD-a6f6ad7 / Rizin 0.9.1 workflow pass: 2 CLI / 5 MCP calls plus catalog, exact 5 MiB mixed-Unicode output, explicit MCP resource-constraint response, full retained Evidence recovery and a successful next command. Disconnect preserves the GUI; owned cleanup verifies; source ELF bytes stay unchanged and the target is not executed.
Native malformed-surrogate output was not generated; its negative coverage is at the production Python sender. macOS and Windows real-provider behavior remain unverified. No directly related cleanup remains.

Completed native task: 10.18 s wall / 11.64 s reaped CPU / 1.22 GiB sampled process-family RSS, two-core affinity, serial local jobs and a 3 GiB Node heap. Sampling may double-count shared pages or miss short processes and excludes assistant/shared services. Host/OS caches are uncontrolled; model tokens and controlled cold/warm costs are unmeasured. No speedup is inferred.

CI and review observations

Current-head CI retained two distinct failed attempts: an existing 60 s CLI dispatcher test timeout, then the 384 MiB cli-boundary worker OOM in cliIntegerArguments.test.ts. The latter is separately tracked as #4713 with cross-PR hosted evidence and explicit missing clean-main reproduction/heap diagnosis. Unchanged attempt 4 passed the affected shard and all required gates, with the same 384 MiB bound, deadlines and assertions. The feature patch changes neither CLI test nor runner configuration.

Current-head Greptile review reports 5/5, with no unresolved findings. Maintainer review is explicitly AI-assisted. Codex reports repository review-credit exhaustion; no new Codex review is claimed. Existing initial human-owner readiness remains valid and unchanged below.

Author readiness

Human owner: @AyoubElleuch

I reviewed the final diff and relevant existing code, understand the change and its edge cases, and take responsibility for the complete PR, including AI-assisted work.
I checked the evidence above, confirmed it supports the current change, and recorded meaningful validation gaps.
I declare this PR ready for initial maintainer review.

Before merge

This PR addresses a concrete problem or an agreed enhancement.
The PR has one focused outcome and the title follows type(scope): outcome.
Related issue is linked.
Required real-provider/platform verification and current CI are complete.
Human review and readiness declaration are complete.

_{🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.}

## 📋 Pull Request Information **Original PR:** https://github.com/morluto/rea/pull/3776 **Author:** [@AyoubElleuch](https://github.com/AyoubElleuch) **Created:** 10/15/2026 **Status:** ✅ Merged **Merged:** 10/17/2026 **Merged by:** [@N0zoM1z0](https://github.com/N0zoM1z0) **Base:** `main` ← **Head:** `fix/cutter-mixed-unicode-output` --- ### 📝 Commits (7) - [`21ad0bc`](https://github.com/morluto/rea/commit/21ad0bc2f15bf869e4093ab888c8c9f655f62af9) fix(cutter): preserve mixed Unicode output within the frame limit - [`61ce038`](https://github.com/morluto/rea/commit/61ce038d5b0c3a95bd460b933f9ed8e5bcc6368e) Merge branch 'main' into fix/cutter-mixed-unicode-output - [`2cb4afc`](https://github.com/morluto/rea/commit/2cb4afc358f709f1022029b81125aa7ed4c6bb68) Merge branch 'main' into fix/cutter-mixed-unicode-output - [`757e7aa`](https://github.com/morluto/rea/commit/757e7aa85a3bf6317b255b1d5560575c5c201831) fix(cutter): preserve completion for unencodable output - [`51c34ab`](https://github.com/morluto/rea/commit/51c34ab4d7591144e7c5c2d083096ff4bc1e0062) Merge branch 'main' into fix/cutter-mixed-unicode-output - [`35f5226`](https://github.com/morluto/rea/commit/35f52269e951ebf9e9abcc6e29097008db1eb018) Merge branch 'main' into fix/cutter-mixed-unicode-output - [`31eb36d`](https://github.com/morluto/rea/commit/31eb36d3b3ce7558b423b526e5fea41d7a9497dd) Merge branch 'main' into fix/cutter-mixed-unicode-output ### 📊 Changes **2 files changed** (+59 additions, -5 deletions) <details> <summary>View changed files</summary> 📝 `bridge/cutter/rea_bridge.py` (+13 -5) 📝 `tests/fixtures/cutter_bridge_protocol_test.py` (+46 -0) </details> ### 📄 Description Fixes #3330. ## Problem and change A completed 5 MiB ASCII Cutter result fits the authenticated 16 MiB frame, but a single non-ASCII suffix made the old preflight charge four bytes for every character and omit it. The producer now counts actual UTF-8 bytes in 65,536-character chunks, rejects an obviously oversized character count early, and retains the final encoded-frame check and HMAC as authoritative. Each temporary conversion is at most 256 KiB. The follow-up also preserves the existing authenticated omission path for unencodable surrogate output: preflight returns false on UnicodeEncodeError, and the sender reports completed execution with output=null and output_truncated=true. Framing, completion and caller choices keep one owner; consumers need no repair. ## Scope and contracts Only `bridge/cutter/rea_bridge.py` and its existing Python protocol fixture change. Existing CLI/MCP schemas, command authority, target selection, framing and cleanup contracts remain intact. MCP may still return its explicit smaller result-budget error with a retained Evidence reference; export recovers the complete output. Ordinary correction within an existing provider; no new capability or maintainer scope decision required. ## Verification Current head: `31eb36d3b3ce7558b423b526e5fea41d7a9497dd`, incorporating main `5b243993872c2bcc70cb00060ffb79b347321370`. Exact merge-tree and stable patch-ID checks confirm the feature patch is unchanged. [Complete exact-head CI](https://github.com/morluto/rea/actions/runs/39011339111) passes all five built-dependent shards, Source/static checks and required gates. - Producer regressions cover ASCII/é/雪/😀, exact decoded equality, HMAC, the wire limit and oversized rejection. Unpaired high/low surrogate cases in the first/last chunks go through the actual sender/socketpair and verify truthful authenticated completion. The regression fails on the preceding PR implementation and passes with the structural fix. - Earlier support revision `757e7aa85a` passed all 15 Python protocol and 36 production bridge/client boundary cases, production build and complete check:ci. These local suite receipts remain attributed to that revision. - Fresh **exact-head** Linux production build and real official free **Cutter 2.5.0-HEAD-a6f6ad7 / Rizin 0.9.1** workflow pass: 2 CLI / 5 MCP calls plus catalog, exact 5 MiB mixed-Unicode output, explicit MCP resource-constraint response, full retained Evidence recovery and a successful next command. Disconnect preserves the GUI; owned cleanup verifies; source ELF bytes stay unchanged and the target is not executed. - Native malformed-surrogate output was not generated; its negative coverage is at the production Python sender. macOS and Windows real-provider behavior remain unverified. No directly related cleanup remains. Completed native task: **10.18 s wall / 11.64 s reaped CPU / 1.22 GiB sampled process-family RSS**, two-core affinity, serial local jobs and a 3 GiB Node heap. Sampling may double-count shared pages or miss short processes and excludes assistant/shared services. Host/OS caches are uncontrolled; model tokens and controlled cold/warm costs are unmeasured. No speedup is inferred. ## CI and review observations [Current-head CI](https://github.com/morluto/rea/actions/runs/39011339111) retained two distinct failed attempts: an existing 60 s CLI dispatcher test timeout, then the 384 MiB cli-boundary worker OOM in cliIntegerArguments.test.ts. The latter is separately tracked as #4713 with cross-PR hosted evidence and explicit missing clean-main reproduction/heap diagnosis. Unchanged attempt 4 passed the affected shard and all required gates, with the same 384 MiB bound, deadlines and assertions. The feature patch changes neither CLI test nor runner configuration. Current-head Greptile review reports 5/5, with no unresolved findings. Maintainer review is explicitly AI-assisted. Codex reports repository review-credit exhaustion; no new Codex review is claimed. Existing initial human-owner readiness remains valid and unchanged below. ## Author readiness Human owner: @AyoubElleuch - [x] I reviewed the final diff and relevant existing code, understand the change and its edge cases, and take responsibility for the complete PR, including AI-assisted work. - [x] I checked the evidence above, confirmed it supports the current change, and recorded meaningful validation gaps. - [x] I declare this PR ready for initial maintainer review. ## Before merge - [x] This PR addresses a concrete problem or an agreed enhancement. - [x] The PR has one focused outcome and the title follows `type(scope): outcome`. - [x] Related issue is linked. - [x] Required real-provider/platform verification and current CI are complete. - [x] Human review and readiness declaration are complete. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>

bsakaguchi

2026-10-31 01:10:14 +00:00

closed this issue
added the
pull-request
label

No milestone

No project

No assignees

1 participant

Notifications

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference

morluto/rea#5893

No description provided.

Rows
Columns