fix(ui): distinguish empty reviews from filtered files

[benvinegar]

Summary

Fixes #1180.

A clean worktree or empty patch currently says “No files match the current filter” even with no active filter. Show “No changes to review.” for empty loaded input; retain the existing filter message when a nonempty review has no visible files.

Approach

• App already owns the unfiltered review and its visible projection, so it selects the presentation-only empty-state message.
• DiffPane renders that message rather than inferring why its projected file list is empty.
• No shared review semantics, extension contracts, or session wire changes.
• Independent of the malformed-patch validation fix (hunk patch silently accepts non-patch input as a successful empty review #1181) and existing sidebar PR fix(ui): keep the selected file visible in the sidebar during held navigation #1085.
• Adds a patch Changeset.

Tests

• App-level empty-review regression fails on untouched main; the existing active-filter behavior remains covered.
• Component tests cover both messages.
• Three real-PTY regressions cover clean worktree, empty stdin, resize, filter exclusion and recovery.

Validation on Linux:

• bun run typecheck — passed.
• bun run lint — passed.
• bun run format:check — passed.
• bun run deps:check — passed.
• bun run test — 5,080 passed, 13 skipped, no failures.
• bun test test/pty/empty-review-integration.test.ts — 3 passed.
• bun run test:integration — 259 passed, 1 skipped, no failures.
• bun run test:tty-smoke — 11 passed, no failures.

CI follow-up

Added a separate test-only commit to address CI failures:

• Website smoke checks now compare the rendered prerelease-label count with the published changelog source instead of assuming there are no prereleases. The previous assertion failed identically on untouched main (6aa52617, Website run). Stable Latest and release ordering/link assertions remain intact. All 103 browser smoke tests passed locally (1 skipped).
• Real Git status fixtures receive a 30-second test budget; Windows CI exceeded the default five seconds and killed Git mid-fixture. No tests or platform checks are skipped.
• A further Windows run exposed the filesystem document suite's default five-second cleanup-hook limit after removing the 10,001-directory fixture. Its hook now gets the same 30-second budget as that bounded-enumeration test; all 17 filesystem-source tests pass locally.

No product behavior, generated release metadata or workflow gating changes are included in this follow-up. The identical test-only commit is also on #1183 so both independent PRs can pass CI.

Real terminal evidence

Ran source hunk diff --no-extensions on a clean disposable worktree and hunk patch --no-extensions with empty file-backed stdin in real tmux PTYs: Linux, 120×32, default terminal theme/auto layout. The corresponding new PTY tests also exercise a 60×12 resize and keyboard filtering/recovery.

Actual tmux screen text (not redirected CLI stdout):

Before:  No files match the current filter.
After:   No changes to review.

No mouse action changes. macOS and Windows not tested locally. The installed npm beta and existing user sessions were left untouched.

This PR description was generated by Pi using gpt-6.1-sol

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated
hunk-web	Ignored	Preview	Oct 13, 2026 8:15pm UTC

[greptile-apps]

Confidence Score: 5/5

[Medium impact] The PR appears safe to merge.

Summary

The PR distinguishes an empty review from a review whose files are hidden.

• Empty reviews and filtered reviews now show different messages.

No actionable issues were found. Tests were not run during this review.

_{Reviews (1) · Last reviewed commit: "fix(ui): distinguish empty reviews from ..." · Reviewed by Greptile}