Permalink
Comparing changes
Choose two branches to see what’s changed or to start a new pull request.
If you need to, you can also or
learn more about diff comparisons.
Open a pull request
Create a new pull request by comparing changes across two branches. If you need to, you can also .
Learn more about diff comparisons here.
...
- 9 commits
- 26 files changed
- 3 contributors
Commits on Mar 23, 2026
-
docs: tighten README and simplify session docs (#74)
* docs: tighten README and simplify session docs * docs: mention OpenTUI and Pierre * docs: restore quickstart usage headings * docs: clarify git-style review entrypoints * docs: tighten git workflow wording * docs: shorten git workflow intro * docs: tighten feature comparison table * docs: add diff-so-fancy to comparison table * docs: simplify comparison summary
-
-
-
-
docs: group README setup under Advanced (#82)
* docs: group README setup under Advanced * docs: format Hunk skill prompt as block
-
perf: benchmark and optimize large split review streams (#76)
* test: benchmark large split review streams * test: flush benchmark highlight work * Baseline note-enabled large split-stream scroll benchmark on PR branch before optimization; current 18-tick note-enabled scroll cost is 16406.23ms and still shows act warnings from deferred highlight work.\n\nResult: {"status":"keep","note_scroll_ticks_ms":16406.23} * Scope visible agent notes to the selected hunk and keep placeholder windowing active for the rest of the review stream; note-enabled 18-tick scroll drops from 16406.23ms to 647.48ms while the interaction test now follows notes as focus moves.\n\nResult: {"status":"keep","note_scroll_ticks_ms":647.48} * Count visible inline note rows in split-mode section metrics so placeholder windowing keeps accurate heights for the selected note-bearing file; note-enabled 18-tick scroll drops from 647.48ms to 231.25ms and the section-height helper now has note-row coverage.\n\nResult: {"status":"keep","note_scroll_ticks_ms":231.25} * Validation rerun after note-height-aware windowing confirms the win: note-enabled 18-tick scroll improves further to 197.32ms, though the benchmark still occasionally surfaces deferred Pierre highlight act warnings that should be cleaned up separately.\n\nResult: {"status":"keep","note_scroll_ticks_ms":197.32} * Baseline after restoring viewport-scoped visible notes rather than selected-hunk-only notes; under the broader visible-note policy the current 18-tick large-stream scroll cost is 2009.15ms.\n\nResult: {"status":"keep","note_scroll_ticks_ms":2009.15} * Cache note-aware section metrics by visible note set so viewport-scoped visible notes reuse placeholder height plans across scroll ticks; under the restored always-visible note policy, the 18-tick large-stream scroll cost drops from 2009.15ms to 301.54ms while viewport-visible notes stay on screen.\n\nResult: {"status":"keep","note_scroll_ticks_ms":301.54} * chore: remove autoresearch log -
fix: daemon autostart for prebuilt npm binaries (#84)
* fix: daemon autostart for prebuilt npm binaries Bun-compiled single-file executables set process.execPath to the embedded Bun runtime, not the binary itself. When the daemon launcher tried to spawn `process.execPath mcp serve`, it was invoking bare Bun with invalid args instead of the prebuilt hunk binary. Fix: when argv[0] differs from execPath, use argv[0] as the command — that's the actual compiled binary the user invoked. Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> * style: format daemon launcher test Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]> --------- Co-authored-by: Claude Opus 4.6 (1M context) <[email protected]>
-
-
Loading
This comparison is taking too long to generate.
Unfortunately it looks like we can’t render this comparison for you right now. It might be too big, or there might be something weird with your repository.
You can try running this command locally to see the comparison on your machine:
git diff v0.5.1...v0.6.0-beta.1