Follow-up from #24 and #22. The headless duel (docker/scenarios/combat) covers shield blocks, flinches and staggers. It doesn't yet exercise:
Parries (#19 weapon block): a phase where the player unequips the shield and holds Block, expecting Ashlar block: "Player" ... with weapon lines with damage through > 0.
Stamina rules (#21): a phase that drains the player's fatigue below [Game] attack fatigue threshold and checks that no attack starts. Then a phase below attack fatigue cap that checks the logged attack strength is capped.
ignoreStagger (#22): an NPC-side I.Combat.addOnHitHandler that sets attack.ignoreStagger = true for some hits, checking that those hits don't log a flinch.
Acceptance
SCENARIO=combat docker/ashlar-run.sh prints counts for each of these, and a run shows at least one parry, one refused attack and no flinch on ignoreStagger hits.
Keep it CC0 and Lua only, with no commercial data. The default boot is unchanged.
Follow-up from #24 and #22. The headless duel (`docker/scenarios/combat`) covers shield blocks, flinches and staggers. It doesn't yet exercise:
- **Parries** (#19 weapon block): a phase where the player unequips the shield and holds Block, expecting `Ashlar block: "Player" ... with weapon` lines with damage through > 0.
- **Stamina rules** (#21): a phase that drains the player's fatigue below `[Game] attack fatigue threshold` and checks that no attack starts. Then a phase below `attack fatigue cap` that checks the logged attack strength is capped.
- **ignoreStagger** (#22): an NPC-side `I.Combat.addOnHitHandler` that sets `attack.ignoreStagger = true` for some hits, checking that those hits don't log a flinch.
**Acceptance**
- `SCENARIO=combat docker/ashlar-run.sh` prints counts for each of these, and a run shows at least one parry, one refused attack and no flinch on ignoreStagger hits.
- Keep it CC0 and Lua only, with no commercial data. The default boot is unchanged.
claude
added this to the 2. Feel milestone 2026-11-14 08:50:35 +00:00
The duel now runs in phases: shield block (2–14 s), parry (14–24 s, shield unequipped, Block held with the blade), attack (24–32 s), ignorestagger (32–40 s, the duelist's own I.Combat.addOnHitHandler sets attack.ignoreStagger = true on hits it takes), tired (40–50 s, player fatigue held at 5%) and capped (50–64 s, fatigue held at 25%, Use held 1.5 s for a full wind-up).
Engine: two verbose log lines in character.cpp, Ashlar stamina: <id> too tired to attack (once per press, any actor; the player message box is unchanged) and Ashlar stamina: <id> attack strength X capped to Y. No behavior change.
docker/ashlar-run.sh counts parries, refused attacks, capped attacks, ignoreStagger hits, and duelist flinches between its ignoreStagger on/off markers. Scenario run time is now 150 s by default. docs/ashlar/SCENARIOS.md documents the phases and log lines.
Tests
Build OK, components-tests 1636 passed, openmw-tests 596 passed. No new unit tests: the engine change is logging only.
SCENARIO=combat docker/ashlar-run.sh: 13 player blocks, 6 parries, 13 NPC blocks, 59 flinches, 6 interrupted wind-ups, 8 attacker staggers, 8 refused attacks, 4 capped attacks (0.97–1.0 capped to 0.625, matching the formula at 25% fatigue and cap 0.5; the duelist's handler saw strength=0.63), 5 ignoreStagger hits with 0 duelist flinches in the window (it flinched twice before it).
Done in `70b7ea51e8`.
**What changed**
- The duel now runs in phases: shield `block` (2–14 s), `parry` (14–24 s, shield unequipped, Block held with the blade), `attack` (24–32 s), `ignorestagger` (32–40 s, the duelist's own `I.Combat.addOnHitHandler` sets `attack.ignoreStagger = true` on hits it takes), `tired` (40–50 s, player fatigue held at 5%) and `capped` (50–64 s, fatigue held at 25%, Use held 1.5 s for a full wind-up).
- Engine: two verbose log lines in `character.cpp`, `Ashlar stamina: <id> too tired to attack` (once per press, any actor; the player message box is unchanged) and `Ashlar stamina: <id> attack strength X capped to Y`. No behavior change.
- `docker/ashlar-run.sh` counts parries, refused attacks, capped attacks, ignoreStagger hits, and duelist flinches between its `ignoreStagger on`/`off` markers. Scenario run time is now 150 s by default. `docs/ashlar/SCENARIOS.md` documents the phases and log lines.
**Tests**
- Build OK, components-tests 1636 passed, openmw-tests 596 passed. No new unit tests: the engine change is logging only.
- `SCENARIO=combat docker/ashlar-run.sh`: 13 player blocks, **6 parries**, 13 NPC blocks, 59 flinches, 6 interrupted wind-ups, 8 attacker staggers, **8 refused attacks**, **4 capped attacks** (0.97–1.0 capped to 0.625, matching the formula at 25% fatigue and cap 0.5; the duelist's handler saw `strength=0.63`), **5 ignoreStagger hits with 0 duelist flinches** in the window (it flinched twice before it).
- Default boot is unchanged and clean.
Feel still needs Eric's playtest (tracked on #5).
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Follow-up from #24 and #22. The headless duel (
docker/scenarios/combat) covers shield blocks, flinches and staggers. It doesn't yet exercise:Ashlar block: "Player" ... with weaponlines with damage through > 0.[Game] attack fatigue thresholdand checks that no attack starts. Then a phase belowattack fatigue capthat checks the logged attack strength is capped.I.Combat.addOnHitHandlerthat setsattack.ignoreStagger = truefor some hits, checking that those hits don't log a flinch.Acceptance
SCENARIO=combat docker/ashlar-run.shprints counts for each of these, and a run shows at least one parry, one refused attack and no flinch on ignoreStagger hits.Done in
70b7ea51e8.What changed
block(2–14 s),parry(14–24 s, shield unequipped, Block held with the blade),attack(24–32 s),ignorestagger(32–40 s, the duelist's ownI.Combat.addOnHitHandlersetsattack.ignoreStagger = trueon hits it takes),tired(40–50 s, player fatigue held at 5%) andcapped(50–64 s, fatigue held at 25%, Use held 1.5 s for a full wind-up).character.cpp,Ashlar stamina: <id> too tired to attack(once per press, any actor; the player message box is unchanged) andAshlar stamina: <id> attack strength X capped to Y. No behavior change.docker/ashlar-run.shcounts parries, refused attacks, capped attacks, ignoreStagger hits, and duelist flinches between itsignoreStagger on/offmarkers. Scenario run time is now 150 s by default.docs/ashlar/SCENARIOS.mddocuments the phases and log lines.Tests
SCENARIO=combat docker/ashlar-run.sh: 13 player blocks, 6 parries, 13 NPC blocks, 59 flinches, 6 interrupted wind-ups, 8 attacker staggers, 8 refused attacks, 4 capped attacks (0.97–1.0 capped to 0.625, matching the formula at 25% fatigue and cap 0.5; the duelist's handler sawstrength=0.63), 5 ignoreStagger hits with 0 duelist flinches in the window (it flinched twice before it).Feel still needs Eric's playtest (tracked on #5).