chore: remove dead code — evaluation APIs, chunk_by_paragraphs, bundled installer, Reasoning.safety (#648) by sypsyp97 · Pull Request #675 · nearai/ironclaw

sypsyp97 · 2026-03-07T19:36:25Z

Summary

Closes #648.

Remove four pieces of dead code:

1. Evaluation module unused APIs (`src/evaluation/success.rs`)

Removed RuleBasedEvaluator::with_min_success_rate() and with_max_failures() — never called
Removed LlmEvaluator::new() — never called
Replaced RuleBasedEvaluator::new() with a default_evaluator() test helper (constructor was only used in tests)
Removed associated #[allow(dead_code)] annotations

2. `chunk_by_paragraphs()` (`src/workspace/chunker.rs`)

Removed the unused function and its two test functions
Only chunk_document() is used in production

3. `install_bundled_channel_from_artifacts()` (`src/extensions/manager.rs`)

Removed the function marked "Used by upcoming hot-activation flow" — never called anywhere

4. `Reasoning::safety` field (`src/llm/reasoning.rs`)

Removed the unused safety: Arc<SafetyLayer> field from Reasoning struct
This cascades through all callers: Reasoning::new() no longer takes a safety parameter
Updated 12 call sites across dispatcher, compaction, heartbeat, commands, worker, thread_ops, tool_builder, and tests

Verification

cargo clippy --all-targets — zero warnings
cargo test — all tests pass

Stats

17 files changed, 69 insertions(+), 287 deletions(-)

gemini-code-assist · 2026-03-07T19:36:52Z

Summary of Changes

Hello, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request focuses on significant code cleanup by removing several unused or dead code components across the codebase. The changes streamline the project by eliminating unnecessary evaluation APIs, an unused text chunking strategy, an unutilized extension installation mechanism, and a redundant safety field within the reasoning engine. This refactoring improves maintainability, reduces the project's footprint, and ensures that the codebase only contains actively used and relevant components.

Highlights

Evaluation Module Cleanup: Removed the unused LlmEvaluator struct and its SuccessEvaluator trait implementation. The RuleBasedEvaluator was moved to a test-only configuration (#[cfg(test)]) as it had no production callers, and its builder methods (with_min_success_rate, with_max_failures) were removed in favor of a simplified with_config() test helper.
Chunking Function Removal: Deleted the chunk_by_paragraphs() function and its two associated test cases from src/workspace/chunker.rs, as only chunk_document() was found to be in production use.
Bundled Channel Installer Removal: Removed the install_bundled_channel_from_artifacts() function from src/extensions/manager.rs, as its intended 'hot-activation flow' was never implemented or wired up.
Reasoning Safety Field Removal: Eliminated the safety: Arc<SafetyLayer> field and its corresponding constructor parameter from the Reasoning struct in src/llm/reasoning.rs. This change was cascaded through all affected callers, including dispatcher, compaction, heartbeat, commands, runtime, builder/core, tools/registry, and various test files, as tool output sanitization is handled elsewhere.

Changelog

src/agent/agent_loop.rs
- Removed safety parameter from HeartbeatRunner::new call.
src/agent/commands.rs
- Removed safety parameter from HeartbeatRunner::new and Reasoning::new calls.
src/agent/compaction.rs
- Removed SafetyLayer import.
- Removed safety field from ContextCompactor struct.
- Removed safety parameter from ContextCompactor::new and Reasoning::new calls.
- Simplified make_compactor test helper by removing SafetyLayer instantiation.
src/agent/dispatcher.rs
- Removed safety parameter from Reasoning::new calls and related test setup.
src/agent/heartbeat.rs
- Removed SafetyLayer import.
- Removed safety field from HeartbeatRunner struct.
- Removed safety parameter from HeartbeatRunner::new and Reasoning::new calls.
src/agent/thread_ops.rs
- Removed safety parameter from ContextCompactor::new calls.
src/agent/worker.rs
- Removed safety parameter from Reasoning::new call.
src/app.rs
- Removed safety parameter from register_builder_tool call.
src/evaluation/success.rs
- Removed LlmEvaluator struct and its SuccessEvaluator implementation.
- Moved RuleBasedEvaluator struct and its implementation into a #[cfg(test)] block.
- Removed with_min_success_rate and with_max_failures builder methods from RuleBasedEvaluator.
- Added with_config test helper to RuleBasedEvaluator for simplified test setup.
- Updated test cases to use the new RuleBasedEvaluator structure and with_config helper.
src/extensions/manager.rs
- Removed the install_bundled_channel_from_artifacts function.
src/llm/reasoning.rs
- Removed SafetyLayer import.
- Removed safety field from Reasoning struct.
- Removed safety parameter from Reasoning::new constructor.
- Simplified make_test_reasoning helper by removing SafetyLayer instantiation.
src/tools/builder/core.rs
- Removed SafetyLayer import.
- Removed safety field from LlmSoftwareBuilder struct.
- Removed safety parameter from LlmSoftwareBuilder::new and Reasoning::new calls.
src/tools/registry.rs
- Removed SafetyLayer import.
- Removed safety parameter from register_builder_tool function.
src/worker/runtime.rs
- Removed safety parameter from Reasoning::new call.
src/workspace/chunker.rs
- Removed the chunk_by_paragraphs function and its associated test cases.
tests/e2e_routine_heartbeat.rs
- Removed SafetyConfig and SafetyLayer imports.
- Removed safety parameter from HeartbeatRunner::new calls and related test setup.
tests/heartbeat_integration.rs
- Removed SafetyLayer import.
- Removed safety parameter from HeartbeatRunner::new call.

Activity

The pull request author, sypsyp97, has performed a comprehensive cleanup of dead code, resulting in a net reduction of 460 lines of code.
The author verified the changes by running cargo clippy --all --tests -- -D warnings, which reported 0 warnings.
All tests passed after the changes, as confirmed by cargo test.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	`/gemini review`	Performs a code review for the current pull request in its current state.
Pull Request Summary	`/gemini summary`	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	`/gemini help`	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

gemini-code-assist

Code Review

This is an excellent pull request that significantly improves the codebase by removing dead code and simplifying several components. The removal of the unused safety field from Reasoning and ContextCompactor is a major cleanup, and the cascading changes have been applied consistently across all affected files. Moving the RuleBasedEvaluator to be test-only and deleting the unused LlmEvaluator and chunk_by_paragraphs function are great improvements to code organization and maintainability. The changes are well-documented and the tests have been updated appropriately. Overall, this is a high-quality contribution.

_{Note: Security Review did not run due to the size of the PR.}

Copilot

Pull request overview

Removes multiple unused APIs and dead code paths to reduce maintenance overhead and eliminate #[allow(dead_code)] annotations, including simplifying Reasoning construction now that tool-output sanitization is handled outside Reasoning.

Changes:

Deleted unused evaluation APIs and moved the rule-based evaluator to test-only code with a with_config() helper.
Removed unused chunk_by_paragraphs() implementation and its tests.
Deleted an unused bundled channel installer helper and removed the unused SafetyLayer field/parameter from Reasoning, cascading signature updates through callers and tests.

Reviewed changes

Copilot reviewed 17 out of 17 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
tests/heartbeat_integration.rs	Updates heartbeat test to use new `HeartbeatRunner::new` signature (no `SafetyLayer`).
tests/e2e_routine_heartbeat.rs	Removes safety setup from routine heartbeat E2E tests and updates runner construction.
src/workspace/chunker.rs	Deletes unused `chunk_by_paragraphs()` and associated tests.
src/worker/runtime.rs	Updates `Reasoning::new` call site to new signature.
src/tools/registry.rs	Removes `SafetyLayer` from builder-tool registration API and updates builder construction.
src/tools/builder/core.rs	Removes unused `SafetyLayer` threading and updates `Reasoning::new` call sites.
src/llm/reasoning.rs	Removes `SafetyLayer` field/constructor parameter from `Reasoning` and updates internal tests.
src/extensions/manager.rs	Deletes unused bundled channel installer helper.
src/evaluation/success.rs	Removes unused evaluator APIs; makes rule-based evaluator test-only and adjusts tests.
src/app.rs	Updates builder-tool registration call site for new signature.
src/agent/worker.rs	Updates `Reasoning::new` call site for new signature.
src/agent/thread_ops.rs	Updates `ContextCompactor::new` call sites for new signature.
src/agent/heartbeat.rs	Removes `SafetyLayer` from `HeartbeatRunner` and updates reasoning construction.
src/agent/dispatcher.rs	Updates `Reasoning::new` call sites in dispatcher and tests.
src/agent/compaction.rs	Removes `SafetyLayer` from `ContextCompactor` and updates reasoning construction + tests.
src/agent/commands.rs	Updates heartbeat runner and `Reasoning::new` call sites for new signatures.
src/agent/agent_loop.rs	Updates heartbeat spawn call to match new `spawn_heartbeat` signature.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot · 2026-03-07T19:40:41Z

src/agent/compaction.rs

+    pub fn new(llm: Arc<dyn LlmProvider>) -> Self {
+        Self { llm }


ContextCompactor::new is public, and removing the SafetyLayer parameter is a breaking API change for any external callers. Consider leaving a deprecated constructor overload/shim that accepts the old parameter list (and ignores it), or ensure the change is called out in the changelog/versioning if this is intended to be breaking.

Copilot · 2026-03-07T19:40:41Z

src/agent/heartbeat.rs

    pub fn new(
        config: HeartbeatConfig,
        hygiene_config: HygieneConfig,
        workspace: Arc<Workspace>,
        llm: Arc<dyn LlmProvider>,
-        safety: Arc<SafetyLayer>,
    ) -> Self {


HeartbeatRunner::new is public, and dropping the SafetyLayer argument is a breaking API change for downstream callers (including any external integrations). If maintaining API stability matters, consider adding a deprecated new(..., _safety) shim or documenting the breaking change + version bump.

Copilot · 2026-03-07T19:40:42Z

src/tools/registry.rs

    pub async fn register_builder_tool(
        self: &Arc<Self>,
        llm: Arc<dyn LlmProvider>,
-        safety: Arc<SafetyLayer>,
        config: Option<BuilderConfig>,
    ) {


ToolRegistry::register_builder_tool is part of the public API (re-exported via the prelude) and its signature change (removing SafetyLayer) is breaking for downstream crates. Consider adding a deprecated wrapper with the old signature (ignoring the safety argument) to preserve compatibility, or ensure the change is reflected in changelog/versioning.

Copilot · 2026-03-07T19:40:42Z

src/tools/builder/core.rs

    pub fn new(
        config: BuilderConfig,
        llm: Arc<dyn LlmProvider>,
-        safety: Arc<SafetyLayer>,
        tools: Arc<ToolRegistry>,
    ) -> Self {


LlmSoftwareBuilder::new is a public constructor; removing the SafetyLayer parameter is a breaking change for any external callers constructing builders directly. If API stability is desired, consider a deprecated shim that accepts the old parameter list (and ignores it) or documenting the breaking change in changelog/versioning.

Copilot · 2026-03-07T19:40:42Z

src/llm/reasoning.rs

 impl Reasoning {
    /// Create a new reasoning engine.
-    pub fn new(llm: Arc<dyn LlmProvider>, safety: Arc<SafetyLayer>) -> Self {
+    pub fn new(llm: Arc<dyn LlmProvider>) -> Self {


Reasoning::new is a public constructor (re-exported from crate::llm) and its signature change is a breaking API change for downstream users. If this crate is consumed as a library, consider keeping a backwards-compatible/deprecated constructor (e.g., new_with_safety(llm, safety) or new(llm) plus a deprecated new(llm, _safety) shim) and/or documenting the breaking change in the changelog with an appropriate version bump.

…ed installer, Reasoning.safety (nearai#648) Remove four pieces of dead code identified in nearai#648: 1. evaluation/success.rs: Remove unused RuleBasedEvaluator methods (with_min_success_rate, with_max_failures) and LlmEvaluator::new(). Replace RuleBasedEvaluator::new() with a default_evaluator() test helper since the constructor was only used in tests. 2. workspace/chunker.rs: Remove unused chunk_by_paragraphs() function and its tests. Only chunk_document() is used in production. 3. extensions/manager.rs: Remove install_bundled_channel_from_artifacts() which was marked for "upcoming hot-activation flow" but never called. 4. llm/reasoning.rs: Remove unused safety field from Reasoning struct. This cascades through all callers (dispatcher, compaction, heartbeat, commands, worker, thread_ops, tool_builder) removing the now-unnecessary Arc<SafetyLayer> parameter from Reasoning::new(). All #[allow(dead_code)] annotations on removed items are also removed. cargo clippy --all-targets produces zero warnings. cargo test passes all tests.

Copilot AI review requested due to automatic review settings March 7, 2026 19:36

Copilot started reviewing on behalf of sypsyp97 March 7, 2026 19:36 View session

gemini-code-assist bot reviewed Mar 7, 2026

View reviewed changes

Copilot AI reviewed Mar 7, 2026

View reviewed changes

sypsyp97 force-pushed the fix/648-clean-dead-code branch from 9b82a0a to 4b9ec1b Compare March 7, 2026 22:58

github-actions bot added size: L 200-499 changed lines and removed size: XL 500+ changed lines labels Mar 7, 2026

sypsyp97 changed the title ~~fix(#648): clean up dead code — unused evaluator APIs, chunk_by_paragraphs, bundled channel installer, reasoning safety field~~ chore: remove dead code — evaluation APIs, chunk_by_paragraphs, bundled installer, Reasoning.safety (#648) Mar 7, 2026

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

chore: remove dead code — evaluation APIs, chunk_by_paragraphs, bundled installer, Reasoning.safety (#648)#675

chore: remove dead code — evaluation APIs, chunk_by_paragraphs, bundled installer, Reasoning.safety (#648)#675
sypsyp97 wants to merge 1 commit intonearai:mainfrom
sypsyp97:fix/648-clean-dead-code

sypsyp97 commented Mar 7, 2026 •

edited

Loading

Uh oh!

gemini-code-assist bot commented Mar 7, 2026

Uh oh!

gemini-code-assist bot left a comment

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Mar 7, 2026

Uh oh!

Copilot AI Mar 7, 2026

Uh oh!

Copilot AI Mar 7, 2026

Uh oh!

Copilot AI Mar 7, 2026

Uh oh!

Copilot AI Mar 7, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

sypsyp97 commented Mar 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

1. Evaluation module unused APIs (src/evaluation/success.rs)

2. chunk_by_paragraphs() (src/workspace/chunker.rs)

3. install_bundled_channel_from_artifacts() (src/extensions/manager.rs)

4. Reasoning::safety field (src/llm/reasoning.rs)

Verification

Stats

Uh oh!

gemini-code-assist bot commented Mar 7, 2026

Summary of Changes

Highlights

Footnotes

Uh oh!

gemini-code-assist bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Mar 7, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 7, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 7, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 7, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Mar 7, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

sypsyp97 commented Mar 7, 2026 •

edited

Loading

1. Evaluation module unused APIs (`src/evaluation/success.rs`)

2. `chunk_by_paragraphs()` (`src/workspace/chunker.rs`)

3. `install_bundled_channel_from_artifacts()` (`src/extensions/manager.rs`)

4. `Reasoning::safety` field (`src/llm/reasoning.rs`)