P1D.4: unified COMTRADE cursors, realtime scrub and SIGRA-style harmonics

[masarray]

Field-driven P1D.4 continuation of the COMTRADE investigation workstation. Keep this PR unmerged until the same field COMTRADE record used for the SIGRA comparison is re-tested.

P1D.4 fixes now in the branch:

• one snap authority for the upper investigation ruler and Time Signals waveform cursors;
• waveform-to-ruler drag mirrored at pointer cadence so both surfaces read as one cursor identity;
• Phasor uses one dedicated P cursor; Harmonics uses one dedicated H cursor;
• P/H live analysis is coalesced at WPF composition cadence with stale-result rejection after mouse-up;
• accented CFG digital labels are recovered using strict UTF-8 with deterministic Latin-1 fallback;
• trigger-origin correction, digital-event timing, cached waveform geometry and bounded full-record rendering remain intact;
• Harmonics workstation redesigned for SIGRA-style engineering comparison: all checked analog channels are shown together as stacked strips, H0…H10 share aligned order columns, H1 is the 100% reference, each order exposes % fundamental plus RMS engineering value, and one selected harmonic order applies across all channels;
• H0/DC uses the native ArdIrec spectrum DC component; no managed DSP/DFT is introduced;
• multi-channel harmonic scrub is calculated in one native-gate batch and reuses the existing per-channel/frame cache to preserve realtime responsiveness.

Exact-head qualification:

• de302869f6437c8eb96e1c1303e57e0d5de1c71c — 4/4 green.
• Build ARSAS #2454 ✅
• COMTRADE integration Add auditable Sampled Values evidence bundles #109 ✅
• SV evidence #1464 ✅
• Windows Installer #831 ✅

Field-test focus against SIGRA / ligne_1:

1. C1/C2 upper marker and waveform line remain aligned while dragging fast.
2. Snap is identical from either drag surface.
3. Phasor has only P and follows scrub without stale-frame flicker.
4. Harmonics has only H and updates all checked analog channels together.
5. Harmonic strips show aligned H0…H10, H1=100%, and readable % + RMS labels comparable to SIGRA.
6. Compare numeric harmonic values at the same reference time (for the supplied SIGRA screenshot, H = -9.9 ms).
7. Digital labels such as Déclenchement preserve accents.
8. TRG remains exactly at the corrected 0 reference.

Do not merge until field comparison is accepted.

[masarray]

Exact-head validation — P1D.3 first slice

Validated head: 99d5c33255232131f9dcad84961968643ded8712

All qualification gates are green on this exact head:

• ✅ Build ARSAS #2410
• ✅ Validate COMTRADE viewer integration Make MMS file-transfer failures directly visible and copyable #86
• ✅ Validate ARSAS Windows installer #808
• ✅ Validate SV evidence bundles #1420

Field-retention policy: keep this PR draft / unmerged until the same protection COMTRADE record is re-tested for cursor drag, left/right pan, dense digital-event readability, and the new Harmonics view.

Expected field differences vs P1D.2D:

• C1/C2 no longer rebuild all waveform geometry while dragging;
• visible digital-edge snap lookup uses a prebuilt sorted index + binary search;
• pan is a translated preview during drag and commits once on release;
• digital transition glyphs and lane metadata are density-suppressed;
• Harmonics starts its P1D.3 workstation treatment while keeping ArdIrec native DFT as the sole calculation authority.

[masarray]

Exact-head qualification is complete for 99d5c33255232131f9dcad84961968643ded8712.

Green gates:

• Build ARSAS #2410 — success
• Validate COMTRADE viewer integration Make MMS file-transfer failures directly visible and copyable #86 — success
• Validate ARSAS Windows installer #808 — success
• Validate SV evidence bundles #1420 — success

Field-test package from installer run #808:

• artifact ARSAS-1.6.35-win-x64-installer
• artifact id 10136722555
• artifact digest sha256:f67f978e6db37cc18fa582cefc33cd5fac707797c9b0e849347e91911739b781
• setup EXE SHA-256 d0ffe7fed1b734b9669eb068ced2da4ae6e49bf06d135968a003f500bd666d80

PR remains draft intentionally until this exact build is exercised against the real protection COMTRADE field record. Focus the field check on C1/C2 drag smoothness, pan responsiveness, digital transition readability/snap, Phasor display, and Harmonics display.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: e8870d2773

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Defer harmonics until the H cursor can be resolved

If Harmonics is selected while the initial disturbance load is still running, _harmonicCursorMilliseconds is initialized but the reference timestamp arrays are not yet available, so this fallback silently analyzes C1 or the record center instead of H. The load completion only raises a navigation event, whose shell handler synchronizes the ruler without refreshing harmonics, leaving a result that can remain associated with the wrong frame while the UI presents H as its reference. Avoid calculating until H resolves, or refresh harmonics when the disturbance reference arrays become available.

Useful? React with 👍 / 👎.