#47 Stage 1: move struct App + ShellApp impl out of src/gtk/ into a backend-neutral src/app.rs

[JDonaghy]

Stage 1 of #47 (native macOS GUI), split out because it is the only stage this fleet can build and verify. #47 stays open as the tracking issue; stages 2–4 (the macOS wrapper itself, MacDriver) need an Apple SDK / objc2 toolchain that no machine in this fleet has.

Blocked on quadraui#699 (Backend Rc-handle symmetry) and on the pin bump in #767 — both must land first.

What

Move struct App, its impl App blocks and impl quadraui::ShellApp for App out of src/gtk/mod.rs into a backend-neutral src/app.rs, leaving src/gtk/mod.rs with only run(), build_shell_config() and genuinely GTK-only helpers.

This is the concrete next step of the north star in GOALS.md: the backends shrink to thin event-to-engine wiring. src/gtk/mod.rs is 7,684 production lines — 38% of both backends' remaining mass — and the 2026-09-03 audit found it is already mostly portable.

Why it is mostly mechanical

From the #47 re-audit (44882e9, full text in PLAN.md → "#47 re-audit findings"):

• Of the ~6,700 lines spanning struct App → impl ShellApp for App, only ~40 lines touch gtk4:: / gio:: / pangocairo:: / glib:: directly.
• map_gtk_key_name, gtk_key_name_to_quadraui, map_gtk_key_with_unicode look GTK-specific but take and return plain &str / quadraui::UiEvent — zero gtk4 dependency. They move verbatim.
• setup_gtk_clipboard is already #[cfg]-branched for target_os = "macos" internally and never touches gtk4. Moves verbatim.

The four platform-typed fields

Field	Type	Disposition
settings_monitor	Option<gio::FileMonitor>	#[cfg(feature = "gui")], ~1 call site
window	Option<gtk4::Window>	#[cfg(feature = "gui")], ~15 call sites, all window-chrome (resize/maximize/CSD)
css_provider	Option<gtk4::CssProvider>	#[cfg(feature = "gui")], ~3 call sites, colorscheme reload
backend	Rc<RefCell<GtkBackend>>	Switch to whatever shape quadraui#699 lands on — this is the blocker

Re-derive all of these at pickup; src/gtk/mod.rs churns heavily and the counts above were measured at 44882e9.

Files

• src/gtk/mod.rs
• src/app.rs (new)
• src/lib.rs

Acceptance

• cargo build --no-default-features and cargo build --features gui both green.
• cargo test --features gui,test-support green — in particular the GtkDriver black-box suite in src/gtk/testing.rs, which exercises the moved dispatch paths through production code.
• src/gtk/mod.rs production line count drops materially; state the before/after using python3 scripts/prod_lines.py src/gtk src/tui_main src/render.rs in the PR.
• No new #[cfg(feature = "gui")] inside src/app.rs beyond the three fields named above — if more is needed, that is a finding to report, not to route around (Platform-Neutrality Rule).

Not in scope

src/macos/, MacDriver, anything requiring an Apple toolchain. Pure refactor: no behaviour change, so no new black-box test is required — say so explicitly in the PR per CLAUDE.md's testing bar.

[JDonaghy]

Coordinator Assignment

Machine: dell64
Repo: vimcode
Files: (unspecified)
Do not touch: (nothing else in flight)

Briefing

Issue #785: #47 Stage 1: move struct App + ShellApp impl out of src/gtk/ into a backend-neutral src/app.rs

Stage 1 of #47 (native macOS GUI), split out because it is the only stage this fleet can build and verify. #47 stays open as the tracking issue; stages 2–4 (the macOS wrapper itself, MacDriver) need an Apple SDK / objc2 toolchain that no machine in this fleet has.

Blocked on quadraui#699 (Backend Rc-handle symmetry) and on the pin bump in #767 — both must land first.

What

Move struct App, its impl App blocks and impl quadraui::ShellApp for App out of src/gtk/mod.rs into a backend-neutral src/app.rs, leaving src/gtk/mod.rs with only run(), build_shell_config() and genuinely GTK-only helpers.

This is the concrete next step of the north star in GOALS.md: the backends shrink to thin event-to-engine wiring. src/gtk/mod.rs is 7,684 production lines — 38% of both backends' remaining mass — and the 2026-09-03 audit found it is already mostly portable.

Why it is mostly mechanical

From the #47 re-audit (44882e9, full text in PLAN.md → "#47 re-audit findings"):

• Of the ~6,700 lines spanning struct App → impl ShellApp for App, only ~40 lines touch gtk4:: / gio:: / pangocairo:: / glib:: directly.
• map_gtk_key_name, gtk_key_name_to_quadraui, map_gtk_key_with_unicode look GTK-specific but take and return plain &str / quadraui::UiEvent — zero gtk4 dependency. They move verbatim.
• setup_gtk_clipboard is already #[cfg]-branched for target_os = "macos" internally and never touches gtk4. Moves verbatim.

The four platform-typed fields

Field	Type	Disposition
settings_monitor	Option<gio::FileMonitor>	#[cfg(feature = "gui")], ~1 call site
window	Option<gtk4::Window>	#[cfg(feature = "gui")], ~15 call sites, all window-chrome (resize/maximize/CSD)
css_provider	Option<gtk4::CssProvider>	#[cfg(feature = "gui")], ~3 call sites, colorscheme reload
backend	Rc<RefCell<GtkBackend>>	Switch to whatever shape quadraui#699 lands on — this is the blocker

Re-derive all of these at pickup; src/gtk/mod.rs churns heavily and the counts above were measured at 44882e9.

Files

• src/gtk/mod.rs
• src/app.rs (new)
• src/lib.rs

Acceptance

• cargo build --no-default-features and cargo build --features gui both green.
• cargo test --features gui,test-support green — in particular the GtkDriver black-box suite in src/gtk/testing.rs, which exercises the moved dispatch paths through production code.
• src/gtk/mod.rs production line count drops materially; state the before/after using python3 scripts/prod_lines.py src/gtk src/tui_main src/render.rs in the PR.
• No new #[cfg(feature = "gui")] inside src/app.rs beyond the three fields named above — if more is needed, that is a finding to report, not to route around (Platform-Neutrality Rule).

Not in scope

src/macos/, MacDriver, anything requiring an Apple toolchain. Pure refactor: no behaviour change, so no new black-box test is required — say so explicitly in the PR per CLAUDE.md's testing bar.

[JDonaghy]

Coordinator: Assignment Failed

Machine: dell64
Status: failed
Exit code: 143
Duration: 5m 58s
Log: /home/john/.coord/logs/d6e3eb334df4.log

[JDonaghy]

Coordinator: Assignment Complete

Machine: precision
Status: done
Exit code: 0
Duration: 14m 48s
Log: /home/john/.coord/logs/6132d904ba9c.log

[JDonaghy]

Coordinator: Assignment Complete

Machine: precision
Status: done
Exit code: 0
Duration: 2m 16s
Log: /home/john/.coord/logs/7b3079723033.log

[JDonaghy]

Coordinator Assignment

Machine: precision
Repo: vimcode
Files: (unspecified)
Do not touch: (nothing else in flight)

Briefing

You are fixing a failed smoke test for issue #785: [retry] #47 Stage 1: move struct App + ShellApp impl out of src/gtk/ into a backend-neutral src/app.rs

The previous worker created branch issue-785-47-stage-1-move-struct-app-shellapp-impl. You are already on that branch.
Do NOT start over — work from the existing code.

What was done

The previous worker's changes are already committed on this branch.
Run git fetch origin && git log --oneline origin/develop..HEAD to see what was done.
Run git diff origin/develop...HEAD to see the full diff.

Test failure

cargo test: 1 failed (gtk::testing::vscode_dimming::breadcrumb_path_paints_dimmer_than_editor_body_text), 2562 passed

Guidance

Fix the failing tests and push.

Rules

• Do NOT start over or rewrite from scratch
• Fix the specific test failures
• Commit your fixes and push with git push origin HEAD

[JDonaghy]

⏱ Needs attention

Machine: precision
Assignment: 80a5e320b96f
Issue: #785
Reason: Running too long

Running 45.1m, past the 45m threshold for type='work'.

Detection + surfacing only — nothing was killed or reassigned. A human should take a look; use coord log <id> to see what it's doing, or coord retry/coord stop to intervene.

[JDonaghy]

Coordinator: Assignment Complete

Machine: precision
Status: done
Exit code: 0
Duration: 56m 32s
Log: /home/john/.coord/logs/80a5e320b96f.log