Files

T

Metal AgentandClaude Opus 5 b6dc70f9fa Give the elements that carry meaning something to say

The adapter is only worth having if the call sites exist, so this wires the
ten elements `UX_VERTICAL_SLICE.md` actually names and stops there. The
shell root is an application; the sidebar, its row list and the workspace
are named regions; every workspace panel is a pane carrying its label, its
execution target, its waiting state and whether it owns the keyboard; every
live terminal surface is a terminal; the command palette and the add-panel
chooser are dialogs; the footer's usage strip is a status. Seven elements
gained a stable GPUI id along the way, because an element worth naming to a
screen reader is an element worth keeping state on.

The other 137 divs are left alone deliberately. Fixed-width slots, meters,
chips, borders and painted terminal cell runs have no identity to hang a
node on, and publishing 80x24 nodes a frame would be a tree nobody can
navigate and a cost on every repaint.

No new strings were invented for the screen reader where the model already
had one. `describe()` labels each sidebar row -- that is its caller, so its
`#[allow(dead_code)]` is deleted rather than carried, as decision 0023
requires. The three words in a pane header's corner and the three words its
description leads with now come from one function. The two lines a pane
header prints about where it runs and what its process is doing moved out of
`pane_context` into `pane_header_lines`, because the pane's semantic node is
built one level up in `workspace_panel` and recomputing them there would
have meant two expressions that agree today and disagree after the first
edit. The add-panel chooser's heading and its accessible name are one
constant.

Four labels are new strings, because nothing on screen names these areas:
"Lumbridge", "Workspace", "Sidebar" and "Harness usage". Each is the name
the code and the docs already use for the thing it labels.

The scorecard's accessibility row now reads fail rather than a hedge. The
call sites are wired; the semantics are absent, because the adapter no-ops.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01SPYebLiN2w4TqnHUYGdECq

2026-09-01 13:14:01 -07:00

8.0 KiB

Raw Blame History

Native UI spike scorecard

The GPUI and Floem programs under spikes/ consume the same six-pane fixture. No product framework decision is accepted until both rows contain measurements from macOS and Linux and the hard gates pass.

Criterion	Hard gate	GPUI	Floem
Builds on macOS Apple Silicon	yes	pending	pending
Builds on Ubuntu	yes	conditional	pass
Builds on Omarchy/Arch	yes	pending	pending
Dependency/license closure permits Apache-2.0 distribution	yes	pending	pending
Keyboard navigation + AccessKit tree	yes	fail: keyboard passes and the call sites are wired, but the adapter no-ops on 0.2.2, so no accessibility tree is produced	fail: no AccessKit integration at pinned revision
IME and composed Unicode input	yes	framework API exists; end-to-end pending	editor API exists; end-to-end pending
Isolated system browser child	yes	pending	pending
Cold startup, p50/p95	record	pending	pending
Idle RSS / six-stream RSS	record	pending	pending
Key-to-present p50/p95	record	pending	pending
Six-stream frame p95/p99	record	pending	pending
Release binary/package size	record	pending	pending
Native menu/window/clipboard/drag-and-drop	review	pending	pending
API clarity and maintenance burden	review	pending	pending

The current programs establish dependency, build, launch, and interaction baselines. GPUI now routes a real PTY through a VT engine and keyboard encoder, paints coalesced styled cell runs and cursor shapes, and derives PTY rows/columns from the 60% terminal region inside each responsive one/three/five-panel layout when the window or attached panel count changes. Retained-history page/top/bottom navigation is wired. Panel detach/reattach keeps background surface updates alive and exposes detached sessions in the sidebar. Text selection, mouse modes, a native Markdown editor, and one isolated browser child remain.

Ubuntu baseline — metal, 2026-08-31

Host: Ubuntu 24.04 x86_64, Rust 1.94.1. Both candidates used the same 1280×800 six-pane fixture and were launched under the active X11 desktop.

Observation	GPUI 0.2.2	Floem `778bb5f2`
`cargo check` from downloaded dependency cache	20.76 s, 1.04 GiB max RSS	24.26 s first attempt + 4.06 s after enabling required editor feature, 753 MiB max RSS
Debug executable link	pass with a user-local linker symlink for installed `libxkbcommon-x11.so.0`	pass
Debug executable size	442 MiB	475 MiB
Native window launch and six-pane render	pass	pass
Root workspace CI	8/8 tests, strict Clippy and doctests pass in 10.79 s	shared result

These are cold-development baselines, not startup, release-size, or runtime performance numbers. GPUI's Ubuntu row remains conditional because the host has the runtime libxkbcommon-x11.so.0 but lacks the unversioned development linker name normally supplied by libxkbcommon-x11-dev. Floem current main also requires its editor feature for the style/label modules used by this shell; that feature is now explicit in the spike. Neither build result by itself decides the UI stack.

Accessibility gate and provisional direction

The published gpui 0.2.2 used by the visual shell has focus and keyboard actions but does not expose the current AccessKit element API. An isolated probe therefore pins GPUI directly to Zed commit ce48461eaadd16c65c31f835511ab96bd3b6e746 with its matching Rust 1.97.1 toolchain. The real element wiring compiles with stable accessibility IDs, workspace/pane/terminal roles, selected state, focus tracking, and a needs-input description. The same probe installs real EntityInputHandler and ElementInputHandler implementations and tests marked-text commit, a composed accent, and a multi-codepoint emoji without splitting UTF-8. Three deterministic tests pass. AT-SPI/OS IME on Linux and VoiceOver/IME on macOS remain end-to-end gates.

Where the adapter has got to, and what it does not do

apps/lumbridge/src/a11y.rs is the adapter decision 0017 promised and never wrote, landed as the first stage of decision 0023. Ten elements now declare what they mean — the shell root, the sidebar and its row list, every sidebar row, the workspace region, every workspace pane, every live terminal surface, the command palette, the add-panel chooser, and the footer's usage strip — using the role, label, description, selected-state and stable-identity vocabulary the probe proves exists at Zed ce48461e. sidebar::model::describe, written and tested under decision 0017 and called by nothing for the whole life of that record, is now what labels a sidebar row, and its #[allow(dead_code)] is deleted.

The gate is still failed, and this work does not move it. The adapter's bodies take the value they are given and drop it, because published gpui 0.2.2 has no AccessKit dependency to hand it to. Nothing reaches AT-SPI or VoiceOver; there is no accessibility tree to inspect. What has changed is that the four facts UX_VERTICAL_SLICE.md requires — which pane, its selected state, its execution target, its waiting state — are now derived in one tested place instead of being absent from the code entirely, so stage 5 of decision 0023 is a change to one file rather than a pass over the renderer. Eight unit tests cover the derivation. None of them is an assistive-technology claim.

The pinned Floem revision has keyboard focus and editor IME plumbing but no AccessKit dependency or semantic tree. That fails Lumbridge's accessibility gate without a maintained framework fork or adapter. The provisional direction is therefore GPUI-first using current Zed GPUI behind a narrow Lumbridge UI adapter. The published GPUI shell remains the lightweight interaction and visual benchmark until the dependency/toolchain migration is accepted.

This is not the final cross-platform release decision. Current-GPUI validation on amd-server required a 624 MiB Rust 1.97.1 toolchain, a 691-package lockfile, and 4.4 GiB of debug artifacts across check, test, Clippy, and build. Packaging, license closure, IME, platform assistive technology, browser parenting, macOS, and Omarchy still have to pass.

The comparable Floem shell passes build, test, strict Clippy, live rendering, streaming, mouse selection, and numeric pane selection on metal. Its command palette opens through the visible clickable affordance and uses Floem's native TextInput, but Ctrl/Cmd+K did not reach the handler in the X11 automation probe. That unresolved focus/event-routing behavior is another current strike against adopting Floem; it is recorded rather than hidden behind the clickable fallback.

Measurement semantics

Both renderers retain the same all-deterministic six-surface action stream for comparison. GPUI now layers an unbounded, SQLite-snapshotted panel registry over that benchmark and presents one, three, or five vertical panels, each with its own 20/60/20 context/work/decision composition. The Add Panel chooser creates Terminal, Browser, Markdown, and Review panels beside the selection or reattaches an existing session. Every Terminal panel receives an independent actor-owned VT session; the three seed non-terminal surfaces keep deterministic updates running. Counters separate external PTY batches/lines from total model updates. The GPUI footer's usage zone is no longer a constant string: it renders a lumbridge-core projection for the selected panel's account profile, with a provenance chip and explicit unavailable phrases, and its sidebar attention count and runtime rows are derived from live panel and ledger state. The Floem shell still renders the static footer fixtures, so the two candidates are no longer comparable on that strip. The GPUI footer also reports dispatch-to-element-build p50/p95 over a bounded 256-sample window. It is deliberately not called key-to-present or frame-present latency: neither candidate exposes a reliable public cross-platform post-present callback. True present latency requires an external platform probe.

8.0 KiB Raw Blame History Unescape Escape