[nwparker]

What

react-markdown's <Markdown> has no internal memoization: every render rebuilds the whole unified remark→rehype→highlight→katex processor and re-parses the document (react-markdown v10 lib/index.js runs createProcessor + parse + runSync per render). MarkdownPreview re-renders on internal state that doesn't affect the rendered output — most visibly, every keystroke in Find (query / match-index state) — so a large doc with code blocks + math re-ran the full parse + syntax-highlight + KaTeX pass on each keypress, making Find laggy.

This:

• hoists the two fully-static plugin arrays to module scope (a fresh array identity per render would defeat the memo), and
• renders the body through a React.memo'd MarkdownBody keyed on content + components.

The pipeline now re-runs only when the rendered content or the components map actually changes. The components map was already memoized, so its identity is stable across Find/body-unrelated re-renders. (Inline-annotation/review state deliberately rebuilds components and thus still re-runs the pipeline — required to keep the live annotation markup correct.)

ELI5

Every time you typed a letter in the preview's Find box, Orca re-cooked the entire markdown document from scratch — re-parsing it, re-syntax-highlighting every code block, re-rendering every math formula — just to draw a search highlight that doesn't change the document at all. On a big doc that made Find stutter. Now the cooked document is remembered and only re-cooked when the document actually changes, so Find (and other unrelated UI blips) are smooth.

Proof of zero regression

• Pure memoization + hoisting static arrays; no output change. The components memo and its deps are untouched, so inline annotation cards / review-note highlights stay live (they intentionally rebuild components).
• renderedContent is primitive string state; Object.is compares by value, so equal text safely skips and any real content change re-renders — no stale preview.
• MarkdownBody adds no host DOM node and stays the same child under the translate="no" bodyRef container, so the OS-page-translation crash guard (fix(markdown-preview): paint Find matches without mutating react DOM (crash 237acef1) #8678) and the Find CSS-Custom-Highlight painting (which reads bodyRef) are unaffected — in fact the body DOM is now more stable during search.
• 106 existing MarkdownPreview tests pass. Independent adversarial review (frontier model) additionally ran the full configured web typecheck (clean) and Vitest suite (33,840 tests pass) with the change present, and found no regression (only a comment-accuracy nit, now fixed).

Proof of perf improvement

The triage confirmed by reading node_modules/react-markdown/lib/index.js that the sync Markdown() has no memoization and rebuilds the processor + re-parses on every render. With the memoized body, Find keystrokes (and review-pulse/store re-renders that don't touch content/components) no longer trigger that O(document-size) parse + highlight + KaTeX pass — each avoided pass is the full pipeline over the whole document. The win scales with document size and the number of code/math blocks.

Made with Orca 🐋

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 62029dd1-a182-4e66-b9b9-9ed9ee34c46e

📥 Commits

Reviewing files that changed from the base of the PR and between a6d3e6a and d4d0d75.

📒 Files selected for processing (1)

• src/renderer/src/components/editor/MarkdownPreview.tsx

🚧 Files skipped from review as they are similar to previous changes (1)

• src/renderer/src/components/editor/MarkdownPreview.tsx

---

📝 Walkthrough

Walkthrough

MarkdownPreview now uses a memoized MarkdownBody component to render markdown. Remark and rehype plugin arrays are hoisted to module scope and typed with ReactMarkdownOptions, keeping their identities stable across renders. The inline Markdown render was replaced with MarkdownBody, passing the rendered content and component mapping.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Description check	⚠️ Warning	The description has strong summary/testing details, but it omits required Screenshots, AI Review Report, Security Audit, and Notes sections.	Add the missing template sections, including screenshot status, AI review findings, a security audit summary, and any platform-specific notes.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title is concise and accurately describes the main change: memoizing the markdown preview pipeline.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}