mirror of
https://github.com/rtk-ai/rtk.git
synced 2026-09-19 07:33:17 +08:00
Matt Van Horn
d1a0d9c2c1
- Drop the needless borrow at the format_diff_changes call site; with warnings = "deny" this was failing cargo clippy --all-targets and blocking the clippy job that build, test and release depend on. - Keep the original both-files dump as the timer.track baseline. The synthesized classic diff is empty for identical files, which recorded input_tokens = 0 and reported ~0% instead of ~46% in rtk gain --history. - Un-gate render_file_diff: fold it into render_diff rather than compiling a production helper only under cfg(test). - Update the scripts/test-all.sh assertion to match the real output.