Vendor the 4 core weights (Regular, Italic, Bold, BoldItalic) of
JetBrains Mono v2.304 under app/assets/fonts/jetbrains_mono/, ~1 MB
total. OFL license checked in alongside. Declared as the
JetBrainsMono family in pubspec so Flutter bundles it automatically.
app/lib/widgets/src/typography.dart exposes clideMonoFamily plus a
platform-ordered fallback chain (SF Mono / Menlo / Monaco →
JetBrains Mono system install / Fira Code / Hack / DejaVu / Liberation →
Cascadia / Consolas → monospace) for web builds and test harnesses
that don't load asset fonts. Terminal widgets pick both up via a
single import.
Co-Authored-By: Claude <noreply@anthropic.com>
test / unit + widget + golden + a11y (push) Successful in 4m34s
test / integration_test (xvfb) (push) Failing after 1m4s
test / bundle smoke (xvfb 5s) (push) Failing after 58s
test / daemon subprocess + web WASM smoke (push) Successful in 2m27s
CLAUDE.md gains a "Decision discipline" guardrail pointing at
decisions/<domain>.md, rewrites every inline ADR link to the new
anchor-link style (ADR 0001 → D-001, etc.), collapses the bottom
"Open questions" section to a one-line pointer, and updates the
parent-project note to reference decisions/architecture.md instead
of the deleted docs/ADRs/.
Three extension slots are reserved with id-only stubs, registered
alongside the existing 21 built-ins in main.dart:
- builtin.decisions — future sidebar for D/Q/R records
- builtin.tickets — future sidebar + kanban workspace
- builtin.claude-control — future .claude/ first-class surface
(Settings / Skills / Agents / Hooks / MCP). Distinct from the
existing builtin.claude PTY-pane stub.
`flutter analyze` clean; builtin + a11y test suites still pass.
Co-Authored-By: Claude <noreply@anthropic.com>
Three tests under app/integration_test/, run against the real built
app via the integration_test package. The load-bearing one is
app_starts_test.dart — boots ClideApp, waits for the root shell to
settle, asserts the three-column layout + welcome tab + statusbar
indicator all render. This is the "tests pass and then app fails to
start" gate that widget tests can't see because they never pump the
real bindings.
theme_picker_test.dart invokes theme.pick, asserts the modal opens,
selects a theme, asserts the modal dismisses.
extension_lifecycle_test.dart disables ipc-status via settings,
asserts its statusbar item unmounts, re-enables, asserts it
reappears.
Co-Authored-By: Claude <noreply@anthropic.com>
Four layers under app/test/:
* kernel/, extension/ — unit tests for every kernel service and
the extension contract. i18n fallback chain gets the full
matrix (exact/lang/default/placeholder, namespace isolation,
unknown namespace, interpolation). Theme resolver and
extension-manager topo sort covered.
* widgets/, builtin/ — widget tests for every primitive and each
Tier 0 built-in. Assertions reach into the Semantics node so
a missing label fails structurally, not visually.
* goldens/ — Alchemist + Ahem font for cross-platform pixel
stability; primitives only (button, tab bar, icon set). No
goldens for compositions — they'd churn through every tier.
* a11y/ — contract-level gate: semantic coverage walks the
built-in catalogue, contrast walks every token pair in every
bundled theme against WCAG-AA thresholds (catches real
regressions: the first run caught `tab.inactive_text` at 2.81
in summer-night), i18n coverage asserts every referenced key
exists in its catalog (exists-in-map, not "key == value" — too
ambiguous for keys that happen to equal their translation).
Helpers under helpers/: `KernelFixture.create()` boots a
KernelServices with in-memory stores + a FakeDaemonClient (no
socket, drivable connected-state), `widget_harness` wraps a
widget in the minimum tree that resolves theme + i18n.
Co-Authored-By: Claude <noreply@anthropic.com>
First real content under app/. Lays the whole Tier 0 foundation in one
commit because the pieces depend on each other circularly (kernel →
extension → widgets → built-ins all reference types from the layer
below); splitting would leave intermediate commits that don't compile.
Key shapes:
* bare WidgetsApp root — no Material, no Cupertino, no Scaffold.
ClideTheme InheritedWidget is the only source of tokens.
* ClideKernel InheritedWidget aggregates 18 services (settings,
project, extensions, theme, panels, events, ipc, commands +
palette + keybindings, clipboard, files, notify, dialog, tray,
secrets, os, net, focus, log, i18n). ExtensionContext exposes
them through a stable interface.
* ClideExtension + sealed ContributionPoint hierarchy (Tab,
StatusItem, Toolbar, Command, TrayItem, LayoutPreset). One
manifest ships N contributions into kernel slots.
* Three-tier theme pipeline: palette (named colors) → semantic
roles → ~60 surface tokens. Defaults at each layer so legacy
palette-only themes produce a complete SurfaceTokens.
* A11y baked in from day one — every interactive primitive wraps
in Semantics(label:, hint:, button:); ensureSemantics() at boot;
theme/contrast.dart helper exposes token pairs for the WCAG
gate in the a11y test suite.
* i18n ported from fframe's L10n pattern (text-driven, namespaced
JSON catalogs, caller-supplied placeholders) with a proper
locale fallback chain (exact → language → default → placeholder)
fframe lacks.
* Four Tier-0 built-ins live (default-layout, welcome, ipc-status,
theme-picker) plus 17 id-reserving stubs so later tiers fill in
without rename churn.
.gitignore extended to cover app/ sub-package artefacts and the
Playwright harness scratch dirs.
Co-Authored-By: Claude <noreply@anthropic.com>
jpmschweitzerandClaude