umputun / agterm / 28854941433
98%

DEFAULT BRANCH: master
Ran
Jobs 1
Files 56
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 99.013%. Remained the same
28854941433

push

github

web-flow 
fix: clear notification badge when refocusing the app on a visible session (#164)

* fix: clear notification badge when refocusing the app on a visible session

The unseen-count badge cleared only on a first-responder transition
(onFocusChange(true)) or a session select. AppKit's per-window first
responder never resigns when the app is merely backgrounded, so
returning focus to agterm without switching sessions left the badge
stuck until you switched away and back.

Extend the surface's didBecomeKey observer to re-run the same "seen"
clear on the become-key edge, gated on liveFocus so it fires only for
the focused pane of the now-key window (the inverse of notification
suppression). A background session's pill is untouched.

Fix #155

* test(review-loop): guard the liveFocus invariant of the refocus badge clear

Final cleanup pass from code-review-loop iteration 1 — addressed 1 confirmed
test-gap finding.

- [test-gap] [ControlSidebarStatusUITests.swift] the positive refocus test
  asserted only that the focused session's badge clears; a regression dropping
  the liveFocus guard (clearing every session's badge on didBecomeKey) would
  still pass it. Add testRefocusingWindowKeepsNonFocusedSessionBadge: a
  non-focused badged session in the same window must survive the refocus.
  Verified it fails with the guard removed and passes with it.

* test: address review — verify non-focused badge, poll window active

- restructure testRefocusingWindowKeepsNonFocusedSessionBadge to badge the
  NON-focused session (tree-verifiable before and after) instead of the focused
  one, so the survives-refocus assertion can't pass trivially when the notify
  fails to land. Re-verified it still fails when the liveFocus guard is removed.
- both refocus tests now poll window.list for the target window becoming active
  before asserting the didBecomeKey-driven clear, so a delayed key transition
  under XCUITest can't flake the assertion.

3710 of 3747 relevant lines covered (99.01%)

8023849.78 hits per line

Jobs
ID Job ID Ran Files Coverage
1 28854941433.1 56
99.01
 GitHub Action Run
Source Files on build 28854941433