Files
deepseek-harness/packages/host/directory-picker-native/README.md
Huanqi Cao 089f4dfad8 feat(picker): open the Win32 folder dialog in-process over koffi
The modern IFileOpenDialog becomes the primary win32 tier: a koffi-driven
COM conversation on a worker_threads worker (the modal Show never blocks
the host event loop), per-monitor-v2 DPI via SetThreadDpiAwarenessContext,
and abort service by re-posting WM_CLOSE to the dialog thread's windows,
with terminate+unref as the last resort (Node cannot interrupt a thread
blocked in native code, and such a worker must never hold the process open).

The PowerShell chain stays as the fallback tier with its trigger widened
from ENOENT to any pwsh failure, closing the review-flagged PowerShell 6
regression (no WinForms: exit 1, not ENOENT, so 5.1 never ran).

Layering keeps per-file coverage honest on every host: pure sequencing and
the driver test against fakes anywhere; the bindings run against a mocked
koffi COM world (the session-persistence-jsonl technique); POSIX hosts
drive the real spawn plumbing to its koffi-load rejection; win32 hosts run
a real open-and-abort-close smoke. The smoke joins processBoundTests: a
worker blocked in a native modal wedges the threads pool's teardown, while
a fork contains it. The worker bundles as its own CJS tsdown entry
(workflow-workerthread's pattern; no TLA), and the host module is imported
statically so the node-half bundle stays chunk-free.

Built-plane and real-COM behavior verified on native Windows: standalone
probes for the source worker, the built CJS worker, and the driver's abort
path all open and close the real dialog.

Agent Notes: new implemented/feature/2026-08-02-win32-in-process-folder-dialog
(bilingual) owns the decision; the DPI note is re-scoped to the fallback tier
it now describes and its AutoUpgradeEnabled attribution corrected (.NET Core
3.0 rewrote FolderBrowserDialog; the opt-out arrived in .NET 6).
2026-08-05 00:31:43 +08:00

2.9 KiB

@deepseek-ai/dsh-host-directory-picker-native

English | 中文

The native-OS-chooser backend of the directory-picker seam: NativeDirectoryPicker registers ctx.directoryPicker with the native capability, whose pick(signal) opens one native chooser per call and resolves the chosen absolute path (null on cancel). Platform tools run without a shell: osascript on macOS and Zenity with a KDialog fallback on Linux; the caller's abort terminates the native process. Windows opens the modern IFileOpenDialog in-process — a koffi-driven COM conversation on a worker thread with per-monitor-v2 DPI awareness, aborted by posting WM_CLOSE to the dialog thread — and falls back to a PowerShell-hosted dialog (pwsh, then Windows PowerShell 5.1, which every Windows ships) whenever that native surface is unavailable; a resolvable pwsh that cannot deliver the dialog (PowerShell 6 has no WinForms) falls through the same way. Only viable when the operator sits at the host's display — remote deployments compose -browse instead. The command boundary (DirectoryPickerRunner) and platform facts are injectable for deterministic tests. The shared no-shell subprocess runner lives in dsh-native-command.

Dual-face package: the browser half (./client) registers a renderless flow occupant into ui-workspace's two directory-flow holes — each open request drives host.pickDirectory and reports the one outcome (picked path / cancel / failure) through the hole's owner conversation. One cordis.yml row therefore composes both sides of the native interaction; the client carries no capability-kind branching, and mounting a second flow package fails at load (the holes are single kind).

Model Experience

None, as the backend serves the GUI host's directory selection; nothing here reaches a model request.

KV Cache effect

None; this package neither assembles nor sends a provider request.

Known Limitations and Deferred Work

  • Linux requires desktop tooling — with neither Zenity nor KDialog installed, pick rejects with an actionable error; it does not fall back to a typed-path prompt (the browse backend is that fallback at the composition level).
  • The Windows fallback chain degrades the dialog — the in-process picker is the modern Explorer-style dialog; where koffi cannot drive COM the PowerShell tiers take over, and a machine that only reaches Windows PowerShell 5.1 gets the legacy folder tree, DPI-corrected but not the modern UI.
  • A wedged abort can leak one dialog thread — when WM_CLOSE never lands (the dialog window was never created), the driver terminates and unrefs the worker; Node cannot interrupt a thread blocked in the native modal call, so that thread lives until process exit.