chore: consolidate parallel-session WIP (HOLD — do not push)

Local snapshot of intermingled in-flight work, committed to unblock the encode
refactor (a clean ffmpeg_win.rs for the vbv-dedup follow-on). These hunks span
the same files and can't be cleanly split here; the commit bundles three
distinct workstreams that each belong in their own PR:

  - logging rework (~43 files: level re-tiering, structured fields, `?e`,
    hot-path flood latches)
  - conflicting-host detection (detect.rs + detect/{linux,windows}.rs + wiring
    in main.rs/mgmt.rs/Cargo.toml/docs/packaging)
  - standby-sink DWM-stall attribution (windows/display_events.rs + capture/
    vdisplay wiring)

NOT verified as a combination. NOT to be pushed until the refactor is done and
these are re-verified and reorganized into their proper per-workstream PRs.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
2026-07-16 12:42:53 +02:00
parent d466e3e2b2
commit 11045a0f70
62 changed files with 1595 additions and 214 deletions
+6 -3
View File
@@ -234,9 +234,12 @@ pub struct DisplayPolicy {
/// untouched.
#[serde(default)]
pub ddc_power_off: bool,
/// EXPERIMENTAL (Windows): after an `Exclusive` isolate deactivates the physical monitors,
/// additionally DISABLE their PnP device nodes (persistently, so a standby monitor/TV whose
/// hot-plug events re-arrive stays disabled) and re-enable them at restore. Targets the same
/// EXPERIMENTAL (Windows): DISABLE physical monitors' PnP device nodes for the stream's
/// duration (persistently, so a standby monitor/TV whose hot-plug events re-arrive stays
/// disabled) and re-enable them at teardown. Two selectors: the monitors an `Exclusive`
/// isolate deactivated, plus — in ANY topology — external monitors that are connected but not
/// part of the desktop (the standby TV that was never active, whose input auto-scan /
/// instant-on HPD cycling re-probes the link every few seconds). Targets the same
/// "connected-but-dark head" periodic-stutter class as [`Self::ddc_power_off`], but at the
/// Windows-reaction level: a disabled devnode's wake events trigger no PnP arrival, no CCD
/// re-evaluation, no DWM invalidation. A crash-recovery journal re-enables leftovers on host