diff options
| author | Yifeng He <yfhe.prsn@gmail.com> | 2026-04-28 03:13:53 -0700 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2026-04-28 13:13:53 +0300 |
| commit | 22d7bef5680a1d55c28ff6246518d96487e95d5e (patch) | |
| tree | 7e793059880f7f5c2aba9afc45cfdf578c4c16e9 /src/session.rs | |
| parent | b6308ef3bec7ebd46ac830b3d36910171e94edf2 (diff) | |
fix: resize w claude code no flicker (#61)
* fix: mouse resize panel failed with claude-code no flicker
* test: run pane mouse reporting drag tests inside tokio runtime
Switch the new horizontal/vertical drag tests to `#[tokio::test]` so
`PaneRuntime::test_with_screen_bytes` has a reactor when it calls
`tokio::spawn`. CI was panicking with "there is no reactor running".
* test: refresh git state in workspace drop slot test
Workspace::test_new caches the parent repo's git branch at construction
time. On master CI the herdr checkout has ref: refs/heads/master in HEAD,
so the cache is populated and workspace_row_height returns 2. On PR CI
actions/checkout uses a detached merge ref, git_branch returns None, row
height becomes 1, and the drop-slot tie-break flips workspace_drop_index_at_row(2)
from Some(0) to Some(1). Refresh git state after pointing identity_cwd at
the per-test temp repo so the cache reflects the temp repo's HEAD.
Diffstat (limited to 'src/session.rs')
0 files changed, 0 insertions, 0 deletions
