Keep window opacity constant when the window loses focus

[thdxg]

Why

With liquid glass enabled, unfocusing the window faded in a desaturation tint over the glass (mirroring Ghostty's TerminalGlassView) — at 0.85 alpha on dark themes, a translucent window read as near-opaque the moment it lost focus. Keeping the appearance identical focused/unfocused looks better.

What

• Remove the tintOverlay and its tintProperties (0.35 light / 0.85 dark constants) from MactermGlassView, plus the backgroundOpacity scaling that existed only to bound the tint.
• Remove WindowAppearance.syncKeyStatus and its windowDidBecomeKey/windowDidResignKey call sites (delegate forwarding unchanged).
• Remove the now-orphaned NSColor.luminance/isLightColor/adjustingSaturation helpers.
• MactermGlassView's doc comment now records the deliberate divergence from Ghostty so the tint isn't reintroduced as a "missing feature".

The non-glass paths never varied with focus, so this only affects the macOS 26 glass appearance. Any residual inactive treatment the system applies to NSGlassEffectView itself is outside our control.

Testing

• mise run format / lint / test all green.
• Verified visually in the debug build: translucent glass window now renders identically focused and unfocused.

🤖 Generated with Claude Code

Summary by CodeRabbit

• Style
  • Simplified window glass appearance handling.
  • Removed the inactive-window tint overlay, resulting in more consistent glass styling regardless of window focus.
• Bug Fixes
  • Window appearance no longer changes through separate key-status synchronization when windows gain or lose focus.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: ec3d4aeb-5bb1-4e15-ba84-758aa0d7c6d8

📥 Commits

Reviewing files that changed from the base of the PR and between 491fd1b and 5deeb94.

📒 Files selected for processing (2)

• Macterm/Views/MainWindow.swift
• Macterm/Views/WindowAppearance.swift

💤 Files with no reviewable changes (1)

• Macterm/Views/MainWindow.swift

---

📝 Walkthrough

Walkthrough

Window appearance removes inactive-glass tint overlay state and key-window synchronization. Delegate key-state callbacks now only forward notifications, while glass configuration uses style, tint, opacity, and corner radius.

Changes

Window appearance synchronization

Layer / File(s)	Summary
Simplify glass appearance updates Macterm/Views/WindowAppearance.swift, Macterm/Views/MainWindow.swift	MactermGlassView no longer maintains an inactive tint overlay or key-window state, and key-state delegate callbacks no longer invoke appearance synchronization.

Estimated code review effort: 2 (Simple) | ~10 minutes

Possibly related PRs

• thdxg/macterm#178: Modifies the same key-window-dependent glass tint logic in WindowAppearance.swift.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and concisely summarizes the main change: preserving opacity when focus changes.
Description check	✅ Passed	The description covers why, what changed, and testing/verification, with only minor heading differences from the template.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch claude/macterm-opacity-focus-d2a4f5

---

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

Window-state benchmark

State	Metric	main@491fd1bda	this branch	Δ
focused	CPU %	1.20	1.00	-17%
	Memory (RSS MB)	107.8	108.8	+1%
	CPU ms/s (powermetrics)	10.9	9.1	-17%
	Wakeups/s (powermetrics)	187.4	189.3	+1%
unfocused	CPU %	1.10	0.90	-18%
	Memory (RSS MB)	112.2	113.4	+1%
	CPU ms/s (powermetrics)	10.3	8.1	-22%
	Wakeups/s (powermetrics)	174.2	176.1	+1%
minimized	CPU %	0.20	0.20	+0%
	Memory (RSS MB)	112.5	113.6	+1%
	CPU ms/s (powermetrics)	1.9	1.5	-18%
	Wakeups/s (powermetrics)	79.4	79.5	+0%
workload-focused	CPU %	2.40	2.20	-8%
	Memory (RSS MB)	163.4	164.8	+1%
	CPU ms/s (powermetrics)	23.5	21.4	-9%
	Wakeups/s (powermetrics)	250.1	268.9	+8%
workload-unfocused	CPU %	2.50	2.10	-16%
	Memory (RSS MB)	163.7	165.1	+1%
	CPU ms/s (powermetrics)	24.1	18.9	-22%
	Wakeups/s (powermetrics)	260.8	266.2	+2%
workload-minimized	CPU %	0.30	0.30	+0%
	Memory (RSS MB)	163.7	165.2	+1%
	CPU ms/s (powermetrics)	3.3	2.6	-19%
	Wakeups/s (powermetrics)	148.2	149.4	+1%

Reported value is the median of 3×10s windows per state (splitting the window and taking the median keeps one co-scheduled spike from skewing a state); CPU % is the process CPU-time delta over a window. Runs land on different shared runners, so treat small deltas as noise — 🔺/🔻 marks changes ≥25% that also clear the metric's absolute noise floor (CPU % ≥0.5, Memory (RSS MB) ≥25, CPU ms/s ≥5, Wakeups/s ≥50); CPU deltas off a noise-dominated baseline aren't flagged (CPU % baseline ≥1.5, CPU ms/s baseline ≥15). Flagged changes add the benchmark:regression / benchmark:improvement label.