Commit Graph

2 Commits

Author SHA1 Message Date
enricobuehler 845a97601d feat(pf-frame,pf-win-display): leaf primitives for the lid-closed first-frame fix
Three leaf-crate additions the IDD-push capturer (pf-capture, plan §W6 C6)
builds on — committed ahead so the capture-crate extraction and the HID
compose kick can land on top:

- pf-frame session_tuning::DisplayWakeRequest — RAII PowerCreateRequest/
  PowerSetRequest(PowerRequestDisplayRequired + SystemRequired), the
  service-grade 'someone is watching this screen' assertion (visible in
  powercfg /requests), held for a capture session so the console cannot
  drop into display-off mid-stream. Object-lifetime, unlike the
  thread-bound ES_* flags in on_hot_thread. Prevention only: no power
  request turns an already-off display back on — that wake is input's
  job (the virtual-mouse compose kick).

- pf-win-display win_display::desktop_bounds() — the virtual-desktop
  bounds as the union of every ACTIVE CCD path's source rect. From the
  CCD database (global), NOT GetSystemMetrics (a per-session view), so
  a non-console-session host still aims HID absolute coordinates at the
  console's real layout.

- pf-win-display console_session_mismatch() — the session guard from
  3d9b3290, copied into the leaf so pf-capture reads it as a peer
  instead of reaching into the orchestrator (relocation authored by the
  W6 extraction session).

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-17 11:17:29 +02:00
enricobuehler ccc4b08d45 refactor(host/W6.2): extract the Windows display-topology cluster into the pf-win-display leaf crate
windows/{win_display,monitor_devnode,display_events}.rs move into
crates/pf-win-display: the CCD/GDI path-activation + mode-set + HDR
advanced-colour + source-rect helpers, the PnP monitor-devnode enable/disable
lever, and the WM_DISPLAYCHANGE / device-arrival watch. The coming pf-capture
crate's IDD-push capturer consumes all three; the host's pf-vdisplay backend
consumes win_display + monitor_devnode. A leaf lets both depend on them as a
PEER instead of the capturer reaching back into the orchestrator (plan §W6).

win_display's one external tie (crate::vdisplay::Mode) becomes the underlying
punktfunk_core::Mode; the cluster is otherwise self-contained (pf-paths for the
state file, serde_json for it, windows). pub(crate) items bump to pub at the
boundary; win_display carries a module-level allow(missing_safety_doc) to keep
the pre-carve behavior (the FFI helpers were pub(crate) unsafe fn with prose
safety docs — the lint only fires once they're pub, and this is an internal
publish=false leaf). The host imports the three modules at its crate root, so
every crate::{win_display,monitor_devnode,display_events}::* path is unchanged.

Verified: Linux clippy -D warnings (leaf empty + host
nvenc,vulkan-encode,pyrowave --all-targets); Windows clippy -D warnings
(pf-win-display --all-targets + host nvenc,amf-qsv --all-targets) Finished exit 0.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-17 10:14:17 +02:00