Skip to content

Latest commit

 

History

History
294 lines (220 loc) · 17 KB

File metadata and controls

294 lines (220 loc) · 17 KB

Programa agent notes

Initial setup

Run the setup script to initialize submodules and build GhosttyKit:

./scripts/setup.sh

Local dev

After making code changes, always run the reload script with a tag to build the Debug app:

./scripts/reload.sh --tag fix-zsh-autosuggestions

By default, reload.sh builds but does not launch the app. The script prints the .app path so the user can cmd-click to open it. Pass --launch to kill any existing instance and open the app automatically:

./scripts/reload.sh --tag fix-zsh-autosuggestions --launch

By default, every --tag shares one DerivedData directory (programa-shared), so switching tags or re-running the same tag is a warm incremental build instead of a cold one. Tags still get their own bundle id, app name, socket, and log — only the build directory is shared. Pass --isolated to give a tag its own DerivedData directory again (today's old default), e.g. when you need two tags to build fully independently:

./scripts/reload.sh --tag fix-zsh-autosuggestions --isolated

Command-line builds via reload.sh also pass COMPILER_INDEX_STORE_ENABLE=NO to xcodebuild, since a CLI build doesn't need Xcode's editor indexing. This doesn't affect indexing when the project is opened in Xcode.app directly.

reload.sh prints an App path: line with the absolute path to the built .app. Use that path to build a cmd-clickable file:// URL. Steps:

  1. Grab the path from the App path: line in reload.sh output.
  2. Prepend file:// and URL-encode spaces as %20. Do not hardcode any part of the path.
  3. Format it as a markdown link using the template for your agent type.

Example (shared DerivedData, the default). If reload.sh output contains:

App path:
  /Users/someone/Library/Developer/Xcode/DerivedData/programa-shared/Build/Products/Debug/Programa DEV my-tag.app

Claude Code outputs:

=======================================================
[Programa DEV my-tag.app](file:///Users/someone/Library/Developer/Xcode/DerivedData/programa-shared/Build/Products/Debug/Programa%20DEV%20my-tag.app)
=======================================================

Codex outputs:

=======================================================
[my-tag: file:///Users/someone/Library/Developer/Xcode/DerivedData/programa-shared/Build/Products/Debug/Programa%20DEV%20my-tag.app](file:///Users/someone/Library/Developer/Xcode/DerivedData/programa-shared/Build/Products/Debug/Programa%20DEV%20my-tag.app)
=======================================================

Never use /tmp/programa-<tag>/... app links in chat output.

After making code changes, always use reload.sh --tag to build. Never run bare xcodebuild or open an untagged Programa DEV.app. Untagged builds share the default debug socket and bundle ID with other agents, causing conflicts and stealing focus.

./scripts/reload.sh --tag <your-branch-slug>

If you only need to verify the build compiles (no launch), use a tagged derivedDataPath:

xcodebuild -project GhosttyTabs.xcodeproj -scheme programa -configuration Debug -destination 'platform=macOS' -derivedDataPath /tmp/programa-<your-tag> build

When rebuilding GhosttyKit.xcframework, always use Release optimizations:

cd ghostty && zig build -Demit-xcframework=true -Dxcframework-target=universal -Doptimize=ReleaseFast

reload = build the Debug app (tag required). Pass --launch to also kill existing and open:

./scripts/reload.sh --tag <tag>
./scripts/reload.sh --tag <tag> --launch

reloadp = kill and launch the Release app:

./scripts/reloadp.sh

reloads = kill and launch the Release app as "Programa STAGING" (isolated from production Programa):

./scripts/reloads.sh

reload2 = reload both Debug and Release (tag required for Debug reload):

./scripts/reload2.sh --tag <tag>

For parallel/isolated builds (e.g., testing a feature alongside the main app), use --tag with a short descriptive name:

./scripts/reload.sh --tag fix-blur-effect

This creates an isolated app with its own name, bundle ID, socket, and derived data path so it runs side-by-side with the main app. Important: use a non-/tmp derived data path if you need xcframework resolution (the script handles this automatically).

Before launching a new tagged run, clean up any older tags you started in this session (quit old tagged app + remove its /tmp socket/derived data).

Debug event log

All debug events (keys, mouse, focus, splits, tabs) go to a unified log in DEBUG builds:

tail -f "$(cat /tmp/programa-last-debug-log-path 2>/dev/null || echo /tmp/programa-debug.log)"
  • Untagged Debug app: /tmp/programa-debug.log

  • Tagged Debug app (./scripts/reload.sh --tag <tag>): /tmp/programa-debug-<tag>.log

  • reload.sh writes the current path to /tmp/programa-last-debug-log-path

  • reload.sh writes the selected dev CLI path to /tmp/programa-last-cli-path

  • reload.sh updates /tmp/programa-cli and $HOME/.local/bin/programa-dev to that CLI

  • Implementation: vendor/bonsplit/Sources/Bonsplit/Public/DebugEventLog.swift

  • Free function dlog("message") — logs with timestamp and appends to file in real time

  • Entire file is #if DEBUG; all call sites must be wrapped in #if DEBUG / #endif

  • 500-entry ring buffer; DebugEventLog.shared.dump() writes full buffer to file

  • Key events logged in AppDelegate.swift (monitor, performKeyEquivalent)

  • Mouse/UI events logged inline in views (ContentView, BrowserPanelView, etc.)

  • Focus events: focus.panel, focus.bonsplit, focus.firstResponder, focus.moveFocus

  • Bonsplit events: tab.select, tab.close, tab.dragStart, tab.drop, pane.focus, pane.drop, divider.dragStart

  • GhosttyApp.logBackground background events log to a separate companion file with -bg inserted before .log, e.g. /tmp/programa-debug-<tag>-bg.log (override with PROGRAMA_DEBUG_BG_LOG)

Regression test commit policy

When adding a regression test for a bug fix, use a two-commit structure so CI proves the test catches the bug:

  1. Commit 1: Add the failing test only (no fix). CI should go red.
  2. Commit 2: Add the fix. CI should go green.

This makes it visible in the GitHub PR UI (Commits tab, check statuses) that the test genuinely fails without the fix.

Debug menu

The app has a Debug menu in the macOS menu bar (only in DEBUG builds). Use it for visual iteration:

  • Debug > Debug Windows contains panels for tuning layout, colors, and behavior. Entries are alphabetical with no dividers.
  • To add a debug toggle or visual option: create an NSWindowController subclass with a shared singleton, add it to the "Debug Windows" menu in Sources/programaApp.swift, and add a SwiftUI view with @AppStorage bindings for live changes.
  • When the user says "debug menu" or "debug window", they mean this menu, not defaults write.

Pitfalls

  • Custom UTTypes for drag-and-drop must be declared in Resources/Info.plist under UTExportedTypeDeclarations (e.g. com.splittabbar.tabtransfer, com.darkroom.programa.sidebar-tab-reorder).
  • Do not add an app-level display link or manual ghostty_surface_draw loop; rely on Ghostty wakeups/renderer to avoid typing lag.
  • Typing-latency-sensitive paths (read carefully before touching these areas):
    • WindowTerminalHostView.hitTest() in Sources/WindowTerminalHostView.swift: called on every event including keyboard. Keyboard events (.keyDown/.keyUp/.flagsChanged) take an early switch currentEvent?.type fast path; everything else — all pointer events, and an ambiguous/nil currentEvent — must fall through to the full divider/sidebar/drag routing below it. Do not add work to that keyboard fast path.
    • NSWindow.programa_sendEvent in Sources/WindowSwizzles.swift: runs for every event. The hit-view context it caches is only ever read for pointer-down events, so it is computed only for those — do not restore an unconditional hit-test here (#183).
    • TabItemView in ContentView.swift: uses Equatable conformance + .equatable() to skip body re-evaluation during typing. Do not add @EnvironmentObject, @ObservedObject (besides tab), or @Binding properties without updating the == function. Do not remove .equatable() from the ForEach call site. Do not read tabManager or notificationStore in the body; use the precomputed let parameters instead.
    • TerminalSurface.forceRefresh() in GhosttyTerminalView.swift: called on every keystroke. Do not add allocations, file I/O, or formatting here.
  • Terminal find layering contract: SurfaceSearchOverlay must be mounted from GhosttySurfaceScrollView in Sources/GhosttyTerminalView.swift (AppKit portal layer), not from SwiftUI panel containers such as Sources/Panels/TerminalPanelView.swift. Portal-hosted terminal views can sit above SwiftUI during split/workspace churn.
  • Submodule safety: When modifying a submodule (ghostty), always push the submodule commit to its remote main branch BEFORE committing the updated pointer in the parent repo. Never commit on a detached HEAD or temporary branch — the commit will be orphaned and lost. Verify with: cd <submodule> && git merge-base --is-ancestor HEAD origin/main. Note: vendor/bonsplit is NOT a submodule — it is vendored in-tree (MIT, from the manaflow-ai fork); edit and commit it like any other source directory.
  • All user-facing strings must be localized. Use String(localized: "key.name", defaultValue: "English text") for every string shown in the UI (labels, buttons, menus, dialogs, tooltips, error messages). Keys go in Resources/Localizable.xcstrings with translations for all supported languages (currently English and Japanese). Never use bare string literals in SwiftUI Text(), Button(), alert titles, etc.
  • Shortcut policy: Every new Programa-owned keyboard shortcut must be added to KeyboardShortcutSettings, visible/editable in Settings, supported in ~/.config/programa/settings.json, and documented in the keyboard shortcut and configuration docs.

Test quality policy

  • Do not add tests that only verify source code text, method signatures, AST fragments, or grep-style patterns.
  • Do not add tests that read checked-in metadata or project files such as Resources/Info.plist, project.pbxproj, .xcconfig, or source files only to assert that a key, string, plist entry, or snippet exists.
  • Tests must verify observable runtime behavior through executable paths (unit/integration/e2e/CLI), not implementation shape.
  • For metadata changes, prefer verifying the built app bundle or the runtime behavior that depends on that metadata, not the checked-in source file.
  • If a behavior cannot be exercised end-to-end yet, add a small runtime seam or harness first, then test through that seam.
  • If no meaningful behavioral or artifact-level test is practical, skip the fake regression test and state that explicitly.

Socket command threading policy

  • Do not use DispatchQueue.main.sync for high-frequency socket telemetry commands (surface.report_*, surface.ports_kick, status/progress/log metadata updates).
  • For telemetry hot paths:
    • Parse and validate arguments off-main.
    • Dedupe/coalesce off-main first.
    • Schedule minimal UI/model mutation with DispatchQueue.main.async only when needed.
  • Commands that directly manipulate AppKit/Ghostty UI state (focus/select/open/close/send key/input, list/current queries requiring exact synchronous snapshot) are allowed to run on main actor.
  • If adding a new socket command, default to off-main handling; require an explicit reason in code comments when main-thread execution is necessary.

Socket focus policy

  • Socket/CLI commands must not steal macOS app focus (no app activation/window raising side effects).
  • Only explicit focus-intent commands may mutate in-app focus/selection (window.focus, workspace.select/next/previous/last, surface.focus, pane.focus/last, browser focus commands).
  • All non-focus commands should preserve current user focus context while still applying data/model changes.

Testing policy

Never run tests locally. All tests (E2E, UI, python socket tests) run via GitHub Actions or on the VM.

  • E2E / UI tests: trigger via gh workflow run test-e2e.yml (see programa-hq CLAUDE.md for details)
  • Unit tests: xcodebuild -scheme programa-unit is safe (no app launch), but prefer CI
  • Python socket tests (tests_v2/): these connect to a running Programa instance's socket. Never launch an untagged Programa DEV.app to run them. If you must test locally, use a tagged build's socket (/tmp/programa-debug-<tag>.sock) with PROGRAMA_SOCKET=/tmp/programa-debug-<tag>.sock
  • Never open an untagged Programa DEV.app from DerivedData. It conflicts with the user's running debug instance.

Ghostty submodule workflow

Ghostty changes must be committed in the ghostty submodule and pushed to the Darkroom Engineering ghostty fork. Keep docs/ghostty-fork.md up to date with any fork changes and conflict notes.

cd ghostty
git remote -v  # origin = upstream, darkroom = fork
git checkout -b <branch>
git add <files>
git commit -m "..."
git push darkroom <branch>

To keep the fork up to date with upstream:

cd ghostty
git fetch origin
git checkout main
git merge origin/main
git push darkroom main

Then update the parent repo with the new submodule SHA:

cd ..
git add ghostty
git commit -m "Update ghostty submodule"

Release

Single lane: every commit on main that passes the CI workflow is automatically built, signed, notarized, and published as the latest GitHub release via .github/workflows/release.yml (triggered by workflow_run on CI completing with conclusion: success, on branches: [main]). There is no nightly/beta channel — if something ships broken, fix it forward on main and the next green CI run auto-ships the fix. Auto-ship builds get a monotonic build number derived from the run ID AND a distinct user-visible version — the committed major.minor with the patch replaced by the workflow run number (e.g. 0.4.213) — both injected into Info.plist at build time, never committed. They publish to a single, reused rolling GitHub release (titled with the effective version) that is overwritten each ship and marked "latest" — so the releases page stays clean (exactly one rolling entry, nothing else) and releases/latest/download/* always resolves to the newest green build. Every ship is therefore distinguishable in the about box and on the releases page.

Each ship also seals a rolling-candidate-<build> draft release as its build-specific payload (the versioned DMG/EXE and dSYMs). Candidates never leave draft state — draft releases are invisible on the public releases page and to releases/latest — so they never add a second entry. After promoting a candidate's assets into rolling, the reconciler deletes every older candidate draft, keeping exactly the just-promoted one around as a private rollback archive (retention 1); download it with gh release download rolling-candidate-<build> --repo darkroomengineering/programa (requires collaborator access, since it is a draft).

Milestone marketing-version bumps (e.g. 0.15.00.16.0) are git tags only — they do not create a GitHub release. Bump, tag, and let the next auto-ship pick up the new major.minor:

./scripts/bump-version.sh          # bump minor (0.15.0 → 0.16.0)
./scripts/bump-version.sh patch    # bump patch (0.15.0 → 0.15.1)
./scripts/bump-version.sh major    # bump major (0.15.0 → 1.0.0)
./scripts/bump-version.sh 1.0.0    # set specific version

This updates both MARKETING_VERSION and CURRENT_PROJECT_VERSION (build number). Then update CHANGELOG.md, which is the source of truth for the changelog, commit, and optionally tag as a milestone marker:

git tag vX.Y.Z
git push origin vX.Y.Z

The tag is a marker in git log/git tag only; it does not trigger a build or a release. The next push to main (or the same commit, once CI goes green) ships it through the normal auto-ship lane.

Notes:

  • Requires GitHub secrets: APPLE_CERTIFICATE_BASE64, APPLE_CERTIFICATE_PASSWORD, APPLE_SIGNING_IDENTITY, APPLE_ID, APPLE_APP_SPECIFIC_PASSWORD, APPLE_TEAM_ID.
  • The rolling release carries appcast.xml, programa-macos.dmg, programa-windows.exe, and the Sparkle enclosures programa-macos-<build>.dmg for the newest builds (keep window in scripts/sparkle_enclosure.js; older ones are pruned after each promotion). The appcast must point at rolling, not the candidate: GitHub serves no assets from a draft, and a candidate URL 404s for every auto-updating client. dSYMs and the versioned EXE live only on the candidate draft.
  • README download button points to releases/latest/download/programa-macos.dmg.
  • Versioning: bump the minor version for milestone tags unless explicitly asked otherwise.
  • Changelog: update CHANGELOG.md; it is the source of truth for the changelog.
  • workflow_dispatch on release.yml still runs a dry-run build that uploads an artifact instead of publishing.