perf(logging): write formatted span/event messages into buffer with write! to avoid nested allocations

[danteboe]

Motivation

• Nested format! calls created transient String allocations in hot logging code paths.

What changed

• Replaced nested format! with write! into preallocated String buffers and integrated with the thread-local buffer.

Performance impact

• Reduces transient allocation and heap churn under heavy logging.

Testing

• cargo check and unit tests for logging code path.

Risk and compatibility

• Low to medium: minor API changes inside logging; validated caller call-sites adjusted. Recommend running the full Rust CI.

Checklist

• cargo check / clippy
• Reviewer: backend/rust, infra/logging

Summary by Sourcery

Optimize logging and related data serialization while improving Flutter view title UI performance and updating tooling configuration.

Bug Fixes:

• Ensure the view rename popover uses the latest bloc state to avoid stale data when renaming views.

Enhancements:

• Cache and lazily render Flutter workspace title bar breadcrumbs to reduce rebuilds and widget construction in large hierarchies.
• Reuse a thread-local buffer and use streaming writes for JSON log serialization to reduce transient allocations and heap churn in Rust logging.
• Avoid intermediate String allocations when formatting span and event messages in the Rust logging layer.
• Optimize RAG ID JSON serialization in chat persistence to avoid intermediate collections and reduce allocation overhead.
• Allow the perf commit type in commitlint configuration for performance-focused changes.

[sourcery-ai]

Reviewer's Guide

Optimizes Rust logging and related code paths to reduce transient allocations by reusing buffers and using write!-based formatting, while also adding Flutter UI caching for breadcrumbs, improving ViewTitle popover state handling, simplifying RAG ID serialization, and allowing perf commit types.

Sequence diagram for Rust logging buffer reuse in FlowyFormattingLayer

sequenceDiagram
  participant TracingSubscriber
  participant FlowyFormattingLayer
  participant LOG_BUFFER
  participant MakeWriter
  participant Writer

  TracingSubscriber->>FlowyFormattingLayer: on_event(event, ctx)
  FlowyFormattingLayer->>LOG_BUFFER: borrow_mut()
  LOG_BUFFER-->>FlowyFormattingLayer: &mut Vec<u8>
  FlowyFormattingLayer->>FlowyFormattingLayer: serde_json::Serializer::new(&mut *buf)
  FlowyFormattingLayer->>FlowyFormattingLayer: serialize_map(None)
  FlowyFormattingLayer->>FlowyFormattingLayer: format_event_message(...)
  FlowyFormattingLayer->>FlowyFormattingLayer: serialize_entry(MESSAGE, message)
  FlowyFormattingLayer->>FlowyFormattingLayer: map_serializer.end()
  FlowyFormattingLayer->>LOG_BUFFER: write_all(b"\n")
  FlowyFormattingLayer->>MakeWriter: make_writer()
  MakeWriter-->>FlowyFormattingLayer: Writer
  FlowyFormattingLayer->>Writer: write_all(&buf)
  Writer-->>FlowyFormattingLayer: Result
  FlowyFormattingLayer-->>TracingSubscriber: return

Loading

File-Level Changes

Change	Details	Files
Make Flutter ViewTitleBar stateful and cache breadcrumb widgets while switching to a lazy horizontal ListView, and improve ViewTitle rename popover state handling.	Change ViewTitleBar from StatelessWidget to StatefulWidget with a private _ViewTitleBarState. Introduce cached breadcrumb widgets and a cache key that depends on ancestor IDs, deletion state, and access-level properties, invalidated when the view changes. Replace eager SingleChildScrollView+Row breadcrumb rendering with a horizontally scrolling ListView.separated that skips the workspace root and reuses cached widgets. Refactor ViewTitle rename popover building into a dedicated method that uses the latest bloc state or falls back to reading state from context.	frontend/appflowy_flutter/lib/workspace/presentation/widgets/view_title_bar.dart
Refactor Rust logging formatting layer to reuse thread-local buffers and avoid nested String allocations via write! and shared Vec.	Add a thread-local LOG_BUFFER Vec for reusable logging serialization buffers. Change span serialization to write JSON directly into the thread-local buffer, append a newline, and emit via the writer without allocating intermediate Vecs. Update format_span_context and format_event_message to use write! into preallocated Strings instead of nested format! calls and to pre-size combined messages. Refactor event logging to reuse the thread-local buffer for JSON serialization and write directly to the writer, simplifying error handling and removing now-redundant emit calls from span lifecycle hooks.	frontend/rust-lib/lib-log/src/layer.rs
Optimize RAG ID serialization in chat persistence to avoid intermediate Vec allocation.	Introduce an internal helper that serializes a slice of Uuid into a JSON array using serde_json::Serializer over a Vec. Update ChatTable::new to serialize rag_ids directly from &[Uuid] using the new helper instead of first mapping into Vec. Preserve existing metadata serialization and ChatTable fields while changing only the rag_ids path.	frontend/rust-lib/flowy-ai-pub/src/persistence/chat_sql.rs
Allow performance-focused commit types in commitlint configuration.	Extend commitlint type-enum rule to include the 'perf' commit type alongside existing types.	commitlint.config.js
Housekeeping: remove unused or generated files from the repo.	Remove unused localization/translation, font, build artifact, and configuration files that are no longer needed or should not be tracked.	frontend/appflowy_flutter/assets/translations/mr-IN.json frontend/appflowy_flutter/macos/build/ios/XCBuildData/PIFCache/project/PROJECT@v11_mod=a7fbf46937053896f73cc7c7ec6baefb_hash=bfdfe7dc352907fc980b868725387e98plugins=1OJSG6M1FOV3XYQCBH7Z29RZ0FPR9XDE1-json frontend/rust-lib/.cargo/config.toml frontend/rust-lib/flowy-sqlite/.env frontend/appflowy_flutter/assets/google_fonts/Poppins/OFL.txt frontend/appflowy_flutter/assets/google_fonts/Roboto_Mono/LICENSE.txt frontend/resources/translations/ur.json frontend/rust-lib/event-integration-test/tests/asset/project.csv

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[CLAassistant]

All committers have signed the CLA.

[sourcery-ai]

Hey - I've found 2 issues, and left some high level feedback:

• In ViewTitleBar, _cachedBreadcrumbs is populated but never actually used by the ListView.separated (except indirectly via _buildDeletedTitle), so either wire the cache into the list rendering or remove the caching to avoid unused complexity and potential confusion.
• The previous breadcrumb row included _buildLockPageStatus(context) after the titles, but the new ListView.separated version no longer renders the lock status; if that indicator is still needed, reintroduce it outside the list so it’s not lost in the refactor.
• In ChatTable::new, the new serialize_rag_ids_from_uuids helper is defined inside the function and uses multiple unwrap/unwrap_or_default calls; consider lifting it to module scope and handling serialization/UTF-8 errors explicitly instead of potentially panicking or silently returning an empty string.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- In `ViewTitleBar`, `_cachedBreadcrumbs` is populated but never actually used by the `ListView.separated` (except indirectly via `_buildDeletedTitle`), so either wire the cache into the list rendering or remove the caching to avoid unused complexity and potential confusion.
- The previous breadcrumb row included `_buildLockPageStatus(context)` after the titles, but the new `ListView.separated` version no longer renders the lock status; if that indicator is still needed, reintroduce it outside the list so it’s not lost in the refactor.
- In `ChatTable::new`, the new `serialize_rag_ids_from_uuids` helper is defined inside the function and uses multiple `unwrap`/`unwrap_or_default` calls; consider lifting it to module scope and handling serialization/UTF-8 errors explicitly instead of potentially panicking or silently returning an empty string.

## Individual Comments

### Comment 1
<location path="frontend/appflowy_flutter/lib/workspace/presentation/widgets/view_title_bar.dart" line_range="105-112" />
<code_context>
+              // Build a cache key from inputs that influence breadcrumb widgets.
+              final currentKey = '${ancestors.map((a) => a.id).join(',')}|${state.isDeleted}|${pageAccessLevelState.isEditable}|${pageAccessLevelState.sectionType.name}';
+
+              if (_cachedKey != currentKey || _cachedBreadcrumbs.isEmpty) {
+                _cachedBreadcrumbs = _buildViewTitles(
+                  context,
+                  ancestors,
+                  state.isDeleted,
+                  pageAccessLevelState.isEditable,
+                  pageAccessLevelState,
+                );
+                _cachedKey = currentKey;
+              }
</code_context>
<issue_to_address>
**issue:** Breadcrumb caching is computed but not actually used in the ListView builder.

`_cachedBreadcrumbs` and `_cachedKey` are updated, but `itemBuilder` still rebuilds breadcrumbs from `ancestors` each time and never reads `_cachedBreadcrumbs`, so the cache has no effect and only adds complexity. Either use `_cachedBreadcrumbs` to drive `itemCount`/`itemBuilder` (e.g., via precomputed widgets or view models), or remove the caching and let `ListView` handle the work.
</issue_to_address>

### Comment 2
<location path="frontend/rust-lib/lib-log/src/layer.rs" line_range="110" />
<code_context>
+  fn serialize_span<
</code_context>
<issue_to_address>
**suggestion:** Span serialization now performs writing directly, but `emit` still exists and the responsibilities are split awkwardly.

`serialize_span` now both serializes and writes via `LOG_BUFFER`, while `emit` still exists as a separate writer over a `Vec<u8>` but is no longer used for spans (and barely for events). This split makes the API harder to reason about and easy to misuse. It would be clearer to consolidate responsibilities (e.g., keep `serialize_span` pure and let `emit` handle I/O, or the opposite) and remove the now‑redundant path so the logging flow and buffer ownership are unambiguous.
</issue_to_address>

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[sourcery-ai]

issue: Breadcrumb caching is computed but not actually used in the ListView builder.

_cachedBreadcrumbs and _cachedKey are updated, but itemBuilder still rebuilds breadcrumbs from ancestors each time and never reads _cachedBreadcrumbs, so the cache has no effect and only adds complexity. Either use _cachedBreadcrumbs to drive itemCount/itemBuilder (e.g., via precomputed widgets or view models), or remove the caching and let ListView handle the work.

[sourcery-ai]

suggestion: Span serialization now performs writing directly, but emit still exists and the responsibilities are split awkwardly.

serialize_span now both serializes and writes via LOG_BUFFER, while emit still exists as a separate writer over a Vec<u8> but is no longer used for spans (and barely for events). This split makes the API harder to reason about and easy to misuse. It would be clearer to consolidate responsibilities (e.g., keep serialize_span pure and let emit handle I/O, or the opposite) and remove the now‑redundant path so the logging flow and buffer ownership are unambiguous.