New atlas_standalone scene/script: attach-or-spawn boot (one REAL
connect_to_sim attempt at SR_PORT-or-9876 — a separate throwaway TCP
probe was proven by live run to kill a pre-accept-loop server via
broken handshake pipe; never abandon a connected socket), else spawn
--port 0 via new ServerProcess.start_with_pipe + LISTENING:{port}
stdout parse, retry against the resolved port. Reader role wired end
to end: protocol.encode_startup_message optional role param (empty
omits the wire key — byte-identical for all existing callers),
sim_bridge.connection_role suppresses the post-handshake
RequestAllSettings auto-send, hud_groups skips AutoPause/AutoResume
sends for readers (all three would otherwise burn Reader violation
strikes per the T-1130 matrix — endorsed by Oscar).
Generic implant host per D-254 SS3: the shell instantiates ALL
registered implant apps; implant_app_manifest gains
available_in_companion (opt-out, default true) and
implant_registry.instantiate_all a standalone filter param (default
preserves hud.gd behavior byte-identically). Boot order is
instantiate_all THEN open_app (reverse renders a permanently black
window — app_changed fires with no listener; matches hud.gd's order).
Owned-server lifecycle: _exit_tree stops a spawned child, attached
servers survive companion close. Known engine limitation documented:
raw SIGTERM bypasses all Godot notifications and orphans a spawned
server; WM close paths verified clean.
Live-verified: spawn-mode (301 systems rendered from systems.db over
the wire), attach-mode, two simultaneous readers, clean shutdown with
zero orphan processes. 14 new gdUnit tests (port/LISTENING parsing).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Three architectural review comments implemented per Tyre's proposals:
#3 — scene_path consumption (full, Option A)
- ImplantRegistry.instantiate_all(parent) loads and instantiates all
registered apps with a declared scene_path. Manifests with empty
scene_path are treated as metadata-only (silently skipped).
- ImplantRegistry.get_app_instance(app_path) returns the live instance.
- hud.tscn no longer direct-instances AtlasApp or EconomicsApp — an
AppsContainer Control holds the registry-managed children.
- hud.gd._ready() calls ImplantRegistry.instantiate_all($AppsContainer).
- main.gd drops @onready vars for atlas_app/economics_app; looks up
both from the registry at the top of _ready().
#6c — schema_version on manifest
- ImplantAppManifest: @export var schema_version: int = 1 (first field).
- ImplantRegistry: const CURRENT_SCHEMA_VERSION := 1; tiered check in
_scan() — older-than-current emits print_verbose, newer-than-current
emits push_warning, both proceed best-effort.
- apps/atlas/app.tres + apps/economics/app.tres: schema_version = 1.
#10 — absorb _on_screen_changed boilerplate into ImplantApp
- Base class gains _screens: Dictionary, _current_screen_id: String,
register_screen(id, screen), current_screen_id(), and a real default
_on_screen_changed implementation that handles leave+hide+enter+show
with has_method guards and same-screen-replace detection.
- atlas_app: deletes _current_screen_id, _get_screen(),
_on_screen_changed() override; on_install() collapses to construct →
setup → wire → register_screen(id, screen) per screen. Preserves
direct screen refs for atlas-specific signal wiring and method calls.
- economics_app: deletes same scaffolding; on_install() reduces to three
lines (construct overview_screen, register_screen, nav.set_default).
Also: replaced Resource.get(name, default) dict-style calls with direct
property access on typed ImplantAppManifest reads (2-arg get() is
Dictionary-only; causes "Too many arguments" parse errors on Resources).
_is_valid_manifest gained a Resource type guard and a property-exists
check.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Addresses 6 mechanical issues from PR review:
- ImplantAppManifest: drop dead display_name and icon_path fields.
default_key carries a TODO noting its future migration to a keybinds
manifest (input concern in app manifest is a layering violation,
tracked explicitly).
- default_mode wire format is now a String ("fullscreen" / "insert") for
mod-author discovery. ImplantRegistry parses via _MODE_MAP, caches the
resolved int in _resolved_modes, and exposes get_resolved_mode(app_path).
main.gd reads the resolved int directly instead of re-parsing.
- Extract client/ui/implant/widgets/system_index.gd (class_name SystemIndex,
static get_sorted_systems). Removes duplicated star_map_data.json loader +
sort lambda from atlas_app and economics overview_screen.
- ImplantApp.on_insert_deactivated() default auto-closes only when the app
is active in INSERT mode. FULLSCREEN apps no longer spuriously close on
insert state changes.
- tooling/db/sqlite-query and sqlite-exec: --help output goes to stderr
(exit 0). Keeps stdout reserved for JSON payloads so JSON-parsing
callers can't get silently corrupted.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Addresses blocker comments from PR review:
- Delete star_map.gd and star_map.tscn — dead implant/map/starchart
HudGroups registration that should have landed with the atlas
unification (D-191 criterion 1)
- Remove test_star_map_is_accessible_from_insert_ui and
test_star_map_scene_exists from test_sprint30.gd — D-191 supersedes
the insert-UI accessibility pattern
- ImplantRegistry._scan() now detects default_key collisions
(first-wins with push_warning) and validates default_mode against
HudGroups.Mode enum (skip + warn on invalid)
star_map_data.json remains — still used by atlas_app and
economics_app for system index lookups.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
jpmschweitzerandClaude Fable 5