#192 closed without a fix + KB note: the camera-seat stage-op omission is latent, not a defect. Checked the history rather than assuming: the camera seat is NOT back-burnered -- built and live-verified 2026-07-18 (2-seat relay, director locks the goal, camera ship tracks and cuts between authored cameras, ranking window and callsign banner reimplemented the same day, screenshots taken) -- but nothing has exercised it since, no tester ticket mentions it, and reaching it needs a deliberate egg entry (hostType=1 + vehicle=camera). So: the missing texture-stage ops in CameraHUDDrawQuad are real in source, the rendering consequence is unknown, and the screen was screenshot-verified working when written, which is decent evidence it inherits acceptable stage state in practice. Filing it as a bug was over-reach -- a code-audit observation promoted to a defect report without establishing a symptom, on a path nobody uses. Recorded instead as a latent-fragility note beside the camera-seat KB entry, with the six-line fix named for the day someone actually sees solid blocks there. Lesson for the sweep backlog: verified-in-code is not verified-as-broken.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

context/open-questions.md

@@ -70,6 +70,14 @@ authentic path scoped.
   Director's flash logic (10s on/15s off, solid final 30s), live `playerRank` pointers re-sort
   rows as scores change. **Each 128×32 ordinal bitmap packs TWO ordinals side by side**
   ("1st|2nd", "3rd|4th" — why 4 bitmaps serve 8 players): texture rank/2, u-half by parity.
+  ⚠ **Latent, not a filed bug (2026-08-15):** `CameraHUDDrawQuad` (L4VIDRND.cpp ~3088) sets a
+  texture and draws WITHOUT the texture-stage ops that #58 had to add to the reticle plate's
+  draw. D3D9 stage state is global and sticky, so this renders correctly only as long as whatever
+  draws before it leaves stage 0 in a texture-friendly state — it was screenshot-verified working
+  when written, so it does today. If the banner or ranking rows ever appear as SOLID BLOCKS, that
+  is this: fix is the same save/set(`SELECTARG1`)/restore pair. Filed as #192 and closed without a
+  fix — no symptom has ever been observed and the seat has not been used in a session since it was
+  built.
   BT_SHOT now captures AFTER the 2D pass (it used to dump pre-HUD — overlays were on screen
   but invisible to screenshots). Screenshot-verified: "1st MAVERICK" + banner over live
   coverage. Remaining: operator-app camera-seat row helper (hostType not in the roster UI),

docs/tracker/CLOSED.csv

@@ -137,4 +137,5 @@ Missiles still genuinely fly — that part was always real — but now
 177,Queued,✅ DONE,#177 CLOSED -- agrees with you. Era witness + the 4.0 manual both say heat is EXPERT-only; logged your note as independent confirmation.,
 179,Pass,✅ DONE,"Small fix, fast turnaround (#179): Oracle reported the Nova spawning with its rear laser's range pips showing in the forward view (self-corrected after one rear-view toggle) — root-caused and fixed the same evening. The HUD now commits the forward view at spawn, as the machine did. Only the pips were wrong; rear weapons never actually fired forward.",8/14 Oracle - OK
 190,Pass,✅ DONE,"Windows Defender briefly flagged the game as a trojan and deleted it. FALSE POSITIVE, and already fixed by Microsoft: the verdict existed in one definition set (8/14 evening) and was gone by the next (8/15 early hours) - the same file now passes cleanly. No action needed; a normal Windows update clears it. We also proved the binary was untampered by comparing it against a known-good build (same sections, no added imports, identical shipped DLLs).",8/15 - second tester on midnight-updated Defender: no problem with the same 957 build
+192,Queued,✅ DONE,"Closed without a fix - no symptom was ever observed. A code audit found a missing graphics setting on the observer-seat callsign banner, but that screen was verified working with screenshots when it was built and nobody has used it since. Reopen if the observer seat ever shows solid blocks where names should be; the fix is about six lines.",
 AUDIT,Pass,✅ DONE,Armor audit � crouched and standing on all chassis with energy weapons.,(undated) Completed by testers; no defects found.

docs/tracker/OPEN.csv

@@ -151,18 +151,15 @@ NEED FROM YOU: stand still, have someone walk then run a mech past you - does th
 189,P2,Queued,🟠 NEED INFO,"After a generator overheats and shuts down, a weapon left on AUTO never comes back online - even once the generator has cooled. Forcing a reassignment revives it.
 STATUS: Re-described from Lynx's full account (the earlier 'generator will not cool' wording was wrong - cooling works fine). Best suspect: the generator's state stays stale after it cools, so auto-assignment keeps thinking it cannot power anything.
 NEED FROM YOU: does it also happen in EXPERT? Lynx saw it in veteran and says expert behaved normally - if that holds, it narrows this a lot.","8/15 - Lynx - Nova, both PPCs seeked to 4, coolant loops boosted, fire repeatedly (~8 min to force the overheat). Loops keep the PPCs cool but seek-4 energy drives generator heat up until it shuts down. PPC forced to manual; left on auto it does not return once cooled. Comes back on a forced reassignment. Flushing coolant DID bring generator temp down. First seen in veteran; expert behaved as expected. 'Deep mojo about heat transfer in the electrical subsystem' - wants more testing before it is documented."
-191,P3,Queued,⚪ QUEUED,"Other players cannot see your coolant flush cloud - the manual says they should.
-STATUS: Confirmed with a 2-node test; the flush never leaves your machine. Same shape as the arm-wreck fix that just landed, so the repair is known.
-NEED FROM YOU: if you have ever seen ANOTHER mech vent white-blue steam, tell us whether you were in the cockpit or the outside view, and whether your own coolant key was involved.",
-192,P3,Queued,⚪ QUEUED,"On the camera/observer seat, the callsign banner can draw as a solid block instead of text.
-STATUS: Found by code audit - the same bug we fixed for the in-cockpit reticle plate was left in the camera path. Small fix.
-NEED FROM YOU: nothing - if you use the observer seat and see solid blocks where names should be, that is this.",
+191,P3,Queued,🟡 PATCHED - RETEST,"Other players could not see your coolant flush cloud - the manual says they should.
+STATUS: FIXED in 4.11.983. The flush never left your machine; it now ships to everyone, tested across two machines.
+NEED FROM YOU: have someone flush coolant while you watch their mech from outside - you should see bluish steam rise from THEIR mech.",
 193,P1,Queued,⚪ QUEUED,"Once, back in July, everyone froze mid-round about a minute after one player's game died - the round ended early for all six.
 STATUS: Never root-caused, logs lost, and it was never filed until now. Not seen since; several netcode fixes have landed since then.
 NEED FROM YOU: if anyone drops mid-round, note whether the rest of you keep playing normally. One clean observation lets us close this.",
 194,P2,Queued,⚪ QUEUED,"An ammo bay fire may be doing far more damage than it should - possibly enough that it always kills, no matter how little ammo is left.
 STATUS: Found by audit. The maths matches the original game; what is unverified is whether one number means per-round or per-salvo.
-NEED FROM YOU: if you survive (or die to) a bay fire when your bin was nearly EMPTY, tell us - a near-empty bin should not be fatal.","8/15 - Oracle, a nearly empty bin should not be fatal, I have survived these, though there was still additional damage from the explosing ammo."
+NEED FROM YOU: if you survive (or die to) a bay fire when your bin was nearly EMPTY, tell us - a near-empty bin should not be fatal.","8/15 - Oracle, a nearly empty bin should not be fatal if the mech can absorb the damage. I have survived these, though there was still additional damage from the explosing ammo."
 195,P3,Queued,⚪ QUEUED,"With 5 or more pilots, the DEATHS number on the Comm screen can run off the edge of the display.
 STATUS: Measured during an earlier audit and never fixed; a safety clamp hides the worst of it.
 NEED FROM YOU: a screenshot of the Comm pilot list in a full 6-8 player round would confirm how it looks now.",