docs: add v3.2.0 smoke test plan

docs/SMOKE_TEST_v3.2.0.md

@@ -0,0 +1,257 @@
+# v3.2.0 Smoke Test Plan
+
+A structured walkthrough for the Intel-Aware Edition. Run with EVE Online open
+and at least 2-3 characters logged in. Every test below has an explicit "do X,
+expect Y" so a missing or broken behavior is unambiguous.
+
+Total time budget: ~30 minutes for the happy path, ~60 minutes if you hit edge
+cases. File issues against this checklist as you go.
+
+---
+
+## Setup (one-time, ~5 min)
+
+1. Pull and run from main:
+   ```bash
+   cd ~/projects/Argus_Overview
+   git pull
+   ./run.sh
+   ```
+2. **Settings → Intel** — verify all four controls render:
+   - "Show threat chrome on previews" (master toggle, on)
+   - "Track per-character location" (on)
+   - "Max jumps for tinting" (default 1, range 0-5)
+   - Replay strip count summary
+3. **Overview tab** — click "Import Windows" or use One-Click Import.
+   Expect: thumbnails appear for every running EVE client, each with its
+   character name in the info label.
+4. **Status dock** — verify it appears above the preview grid with one
+   chip per imported character. Each chip shows:
+   - Colored-initials avatar (e.g., "AB" for AreteBoss)
+   - Character name in bold
+   - "—" in the system label slot
+5. **Per-character accent border (PR #70)** — each preview frame should
+   render with a thin 2px border in the same color as the chip avatar.
+   Different characters → different border colors.
+
+If any of the above is missing, stop here — the import path is broken
+and nothing downstream will work.
+
+---
+
+## Test 1 — Per-character accent border at small grid sizes (PR #70)
+
+**Why**: Accent border was added so identity is glanceable when the grid
+gets cramped.
+
+1. Resize the Argus window so previews shrink to ~150x110.
+2. **Expect**: each frame's accent border is still visible. Different
+   characters render in distinct colors.
+3. Restart Argus, re-import.
+4. **Expect**: same character → same color as before. (PR #79 fix —
+   if colors swap on restart, that's a regression.)
+
+---
+
+## Test 2 — Per-character system tracking (PR #65)
+
+**Why**: Each character's system label should update independently from
+their own EVE Local channel log.
+
+1. Pick one of your characters and jump to a different system in EVE.
+2. Wait up to 2 seconds (tracker poll interval).
+3. **Expect**: that character's chip system label updates from "—" to
+   the new system name. Other characters' labels do NOT change.
+4. Jump again to a third system.
+5. **Expect**: chip updates again to the new system name.
+6. Repeat with another character.
+7. **Expect**: each character tracks independently.
+
+**If broken**: most likely the EVE log path detection failed. Check
+`~/.eve/logs/Chatlogs/` for `Local_*.txt` files dated today. If those
+exist but tracking doesn't update, file an issue with the directory
+path.
+
+---
+
+## Test 3 — Threat chrome master toggle (PR #78, your "make it optional" ask)
+
+**Why**: Verify the toggle actually suppresses everything.
+
+1. Click "Test Alert" in the Intel tab. (This fires a WARNING report.)
+2. **Expect (chrome ON)**: every preview frame's border tints orange
+   for ~30s, every chip dot lights up orange.
+3. **Settings → Intel → uncheck "Show threat chrome on previews"**.
+4. **Expect**: all active threat tints clear immediately. No frame
+   borders, no chip dots.
+5. Click "Test Alert" again with chrome OFF.
+6. **Expect**: NOTHING visual changes on previews or chips. But the
+   intel report should still appear in the Intel tab's log.
+7. Re-enable the toggle.
+8. **Expect**: future test alerts tint normally again.
+
+---
+
+## Test 4 — Same-system threat fan-out (PR #66)
+
+**Why**: With per-character tracking on, only the matching character
+should tint.
+
+1. Move character A to system X, character B to system Y. Wait for
+   chip labels to update.
+2. Edit `intel/parser.py` test injection or use the Python console (see
+   "Manual report injection" below) to fire a DANGER report **for
+   system X**.
+3. **Expect**: only character A's frame border + chip dot tint red.
+   Character B stays calm.
+4. **Expect**: character A's frame pulses (border thickens 3→5px for
+   ~600ms then settles).
+5. **Expect**: tint decays to invisible over ~30s.
+
+---
+
+## Test 5 — Jumps-from fan-out + "+Nj" badges (PR #67, #68, #71)
+
+**Why**: Adjacent characters should also tint at reduced intensity with
+a distance badge.
+
+1. With characters in two different systems that are 1 jump apart
+   (use jump info in EVE map):
+2. Fire a DANGER report for the system character A is in.
+3. **Expect**:
+   - Character A's frame: full red border, no badge (same-system).
+   - Character B's frame: half-opacity red border, "+1j" badge near
+     top-left.
+   - Both chips: same dot intensities, "+1j" text next to character
+     B's chip dot.
+4. Move character B 2 jumps away.
+5. Fire DANGER for character A's system.
+6. **Expect**: character B's frame tints at the floor alpha (~40%)
+   with "+2j" badge.
+7. Move character B 3+ jumps away.
+8. Fire same alert.
+9. **Expect**: character B's frame is **untinted** (default
+   `intel.threat_jumps_threshold = 1` excludes it).
+10. **Settings → Intel → set "Max jumps" to 3**.
+11. Re-fire same alert.
+12. **Expect**: character B now tints at the floor alpha with "+3j".
+
+---
+
+## Test 6 — Focus mode (PR #64)
+
+**Why**: Spotlight one preview when something dangerous is happening.
+
+1. **Double-click** any preview frame.
+2. **Expect**:
+   - That frame scales up larger (past the normal 600x450 cap).
+   - All other frames fade to 25% opacity.
+3. Press **Esc**.
+4. **Expect**: focus exits, all frames return to normal.
+5. Double-click frame A. While focused, double-click frame B.
+6. **Expect**: spotlight swaps to frame B; frame A drops to dimmed.
+7. While focused on a frame, fire a DANGER alert for that character's
+   system.
+8. **Expect**: focused frame still pulses, but the dim frames don't
+   fight it for attention. (Visual sanity check.)
+
+---
+
+## Test 7 — Replay strip (PR #72)
+
+**Why**: Combat scrubbing for the "did I see a sabre uncloak?" case.
+
+1. Right-click a preview frame → **"Toggle Replay Strip"**.
+2. **Expect**: a horizontal strip of ~6 thumbnail cells appears at the
+   bottom of that frame. Initially empty (the buffer takes ~5 seconds
+   to fill).
+3. Wait 5+ seconds while the EVE client renders different content
+   (e.g., undock, look around).
+4. **Expect**: the strip cells fill left-to-right with thumbnails of
+   the recent capture history.
+5. Hover the leftmost cell.
+6. **Expect**: the main image label of that frame swaps to the older
+   captured frame. A blue highlight appears on the hovered cell.
+7. Move the mouse off the strip.
+8. **Expect**: main image returns to live capture.
+9. Right-click → **"Toggle Replay Strip"** again.
+10. **Expect**: strip vanishes. Buffer keeps accumulating in the
+    background but isn't visible.
+11. Restart Argus.
+12. **Expect**: the strip you toggled on is restored on launch.
+
+---
+
+## Test 8 — Decay timing (cross-cutting)
+
+**Why**: 30s decay was tuned with no real-world data. Validate it.
+
+1. Fire a DANGER alert.
+2. Stopwatch the time from full intensity to invisible.
+3. **Expect**: ~30 seconds, smooth linear fade.
+4. Subjective: does it linger too long during a busy fleet op? Or
+   disappear too fast? File a tuning note if either.
+
+---
+
+## Manual report injection (advanced)
+
+The "Test Alert" button only fires a generic WARNING. To exercise
+DANGER/CRITICAL or specific systems, paste this into a running Python
+console attached to the app (or modify `_test_alert` temporarily):
+
+```python
+from argus_overview.intel.parser import IntelReport, ThreatLevel
+from datetime import datetime
+
+# In MainWindowV21 instance scope (e.g., self):
+report = IntelReport(
+    system="HED-GP",                      # change to test specific systems
+    threat_level=ThreatLevel.DANGER,      # CLEAR / INFO / WARNING / DANGER / CRITICAL
+    hostile_count=3,
+    ship_types=["sabre", "broadsword"],
+    player_names=[],
+    raw_message="hostiles HED-GP +3 sabre broadsword",
+    timestamp=datetime.now(),
+)
+self.intel_tab.alert_dispatcher.dispatch(report)
+```
+
+For per-character system testing, override the location tracker:
+
+```python
+# Force a character to a specific system without waiting for log parsing
+self.main_tab.window_manager.set_character_system("CharName", "HED-GP")
+self.main_tab.status_dock.set_character_system("CharName", "HED-GP")
+```
+
+---
+
+## Filing issues
+
+For each broken behavior:
+
+```
+**Test #N step Y**
+**Expected**: <copy from the doc>
+**Got**: <what actually happened>
+**Setup**: <character count, theme, FPS, OS>
+```
+
+Use the [issue tracker](https://github.com/AreteDriver/Argus_Overview/issues).
+Tag with `smoke-test-v3.2.0` if you create the label.
+
+---
+
+## Out of scope
+
+These are known limitations, not bugs:
+
+- Test Alert button only fires WARNING; use the manual injection
+  recipe for DANGER/CRITICAL.
+- The `replay_strip_enabled` summary in IntelPanel is read-only —
+  per-character toggles live in the right-click menu.
+- Replay strip uses 800ms throttle; capture rates higher than ~1.25 FPS
+  do not increase the buffer's temporal resolution.
+- Auto-tag → release.yml workflow can't auto-publish GitHub releases
+  without a PAT (debt #4 in the post-v3.2.0 audit).