# Paladin macro test cases Prepared **7 November 2026**. This is a prospective checklist, not a test report. **Every record below is NOT_RUN. No game client was used to produce this appendix.** Use the [client test plan](PALADIN_CLIENT_TEST_PLAN.md) for the test order, setup, shared branch matrices and evidence requirements. Record each execution with the [test record template](templates/PALADIN_TEST_RECORD.md), applying the relevant rows of the [repository validation matrix](VALIDATION_MATRIX.md). ## Exact content under review The [canonical Paladin pack](../macros/paladin/pack.json) contains **39 records: 36 researched designs and three deprecated historical records**. Every `tested_builds` array is empty. This checklist changes none of those states. - Application source: `f5d4822758cc022f22279eb9b04e11e5b08c97ed`. - Raw pack SHA-256: `70bb134481befab9747905ea351e651814a850d5aca7c5efbd79713c8090fec7`. - Research snapshot: Talents Forever generated 4 November 2026, reviewed 7 November, base build 1.60.1.70755 with separately reported development-note and server-hotfix qualifications. This is **not** an observed test build. Before testing, compare the current pack's hash, each selected record's version and its exact body with this snapshot. A difference requires reviewing the affected cases; do not attribute results from another body/version to these records. Record the actual client build and learned abilities independently. ## Shared groups and interpretation The main plan defines **B** baseline, **S** Shift support, **A** Alt support, **H** hostile mouseover, **P** hybrid and **C** cast/cancel matrices. **B applies to every current record**, in addition to any group shown in the table. A dash means the special workflow needs its own cases, not that baseline checks can be omitted. Each special check below describes an authored branch or a source-supported expectation **to observe in the client**. It does not assert a successful cast, current acquisition, numerical effect, cooldown duration or verified rank. Record the actual result and any discrepancy. A missing spell, talent, helper, target type or test setup is blocked/not run, not a pass. Expected failed-cast cases pass only when the observed selection and failure match the documented case; they do not replace successful intended-cast evidence. Distinguish two kinds of fallback throughout: - An invalid condition, such as a dead mouseover in a `nodead` branch, allows evaluation of the next branch. - A matching but unusable recipient, such as an out-of-range living friendly mouseover, selects that branch. It must not be treated as permission to cast on the next target or choose another spell. ## Current records — all NOT_RUN | Exact macro ID | Version | Extra groups | State | Per-record special checks to observe | | --- | --- | --- | --- | --- | | `paladin-prot-holy-strike` | 1.1.2 | — | NOT_RUN | Living hostile **current target** only; a different hostile mouseover is ignored. Exercise friendly/dead/no target, melee weapon, range and facing conditions. Compare the ordinary ability button with the exact macro from attacks stopped and after target changes, including resource, cooldown and GCD failures. No explicit `/startattack` does not establish native swing behavior. | | `paladin-prot-fury-judgement` | 1.0.3 | H | NOT_RUN | Shift applies Seal of Fury; release Shift and press separately after the actual GCD for Judgement. Test the intended Fury taunt on an eligible enemy, another learned seal and no seal. Confirm the macro does not apply a seal during the unmodified branch, wait, recover a failed taunt or select an alternative recipient. Check modifier tooltip and binding conflicts. | | `paladin-prot-hammer-of-justice` | 1.0.2 | H | NOT_RUN | Its `/stopcasting` conditions match a living hostile mouseover or target. While casting, test a matching enemy with the stun ready and with cooldown/range/immunity failure; cancellation can occur before the failed stun. With no matching hostile branch, check that this command does not deliberately cancel the cast. A stun is not a guaranteed interrupt against an immune enemy. | | `paladin-prot-purify` | 1.0.2 | S | NOT_RUN | Exercise eligible poison/disease effects and a magic-only negative case. A matched ally without an eligible effect must not redirect the action to another ally who needs dispelling. | | `paladin-prot-freedom` | 1.0.2 | S | NOT_RUN | Observe the effect on an eligible recipient and whether it replaces this Paladin's existing blessing. The macro does not inspect movement debuffs or choose who needs Freedom. | | `paladin-prot-blessing-of-protection` | 1.1.2 | S | NOT_RUN | The final `[@player]` makes `/stopcasting` apply on every press, including self fallback and failed attempts. Test while casting with valid friendly recipients, no friendly recipient, range failure, cooldown failure and Forbearance. Record actual recipient eligibility, physical-action restrictions and Forbearance behavior. Self fallback can spend the emergency cooldown on the caster. | | `paladin-prot-lay-on-hands` | 1.1.2 | S | NOT_RUN | The final `[@player]` likewise causes cast cancellation on every press, including failed attempts. Test intended ally, Shift self and no-recipient self fallback. Record mana, healing and cooldown before/after; do not inherit the source's mana consequence as an observed result. | | `paladin-prot-flash-of-light` | 1.0.2 | S | NOT_RUN | Observe cast-time and repeated-press behavior. There is no `/stopcasting` line. Compare ordinary spell behavior without claiming that this macro supplies an immediate cast interruption. | | `paladin-prot-redemption` | 1.0.2 | — | NOT_RUN | Out of combat, a dead friendly mouseover precedes a dead friendly current target. Test in-combat exclusion, living/hostile/no-corpse cases and actual dead-player eligibility. A matching out-of-range, obstructed or ineligible corpse must not redirect to a valid current-target corpse. Confirm the spell's actual acquisition separately. | | `paladin-prot-exorcism` | 1.0.2 | H | NOT_RUN | Test source-supported Undead and Demon recipients plus a wrong-type hostile mouseover ahead of an otherwise eligible target. The wrong-type match must not fall through. This is a damage attempt, not a general-purpose taunt. | | `paladin-prot-templars-bulwark` | 1.0.2 | — | NOT_RUN | Confirm actual learned spell and legal Protection talent allocation. `/stopcasting` is unconditional: test a current cast before a ready attempt and before cooldown, unavailable-talent or Forbearance failures where those setups can be exercised. Record the actual defensive effect and restrictions; the macro does not clear Forbearance or combine defensives. | | `paladin-prot-cancel-protection` | 1.0.2 | — | NOT_RUN | Unmodified press does nothing; Shift removes only the caster's existing Blessing of Protection. Test with no own aura, another player's aura and a different own immunity. It must not cancel another player's blessing, Divine Shield or Forbearance. A helper-applied own BoP can exercise cancellation without claiming the caster learned BoP. | | `paladin-holy-light` | 1.1.2 | A | NOT_RUN | Observe the exact friendly/self recipient, cast-time and repeated-press behavior. There is no authored cast-cancellation command. | | `paladin-cleanse` | 1.1.2 | A | NOT_RUN | Exercise eligible poison, disease and magic effects, plus a non-removable-effect case. A matched recipient with nothing eligible must not redirect to someone else or imply that every debuff is removable. | | `paladin-holy-shock` | 1.1.2 | — | NOT_RUN | Test Alt self, helpful mouseover, **hostile mouseover**, friendly target, hostile target and self fallback in that order. An enemy mouseover ahead of a friendly target must select the enemy branch. Observe shared healing/damage spell availability and failures without treating this as a readiness-based selector. Include the Vigil conflict experiment below. | | `paladin-lights-vigil` | 1.1.1 | — | NOT_RUN | Test living friendly mouseover, then living friendly target, otherwise no authored cast. Hostile mouseover is ignored; there is no modifier/self-fallback branch. Test setup and the later, separately pressed Holy Shock on the intended same recipient. Perform the self-eligibility diagnostic below as an unknown question, not a promised successful self cast. | | `paladin-seal-command-choice` | 1.0.2 | — | NOT_RUN | Default selects learned/talented Seal of Command; Shift selects Seal of Righteousness. Observe replacement of the active seal and modifier tooltip. An unlearned default must not fall back to Righteousness. The player selects the seal; the macro does not inspect procs or target state. | | `paladin-judgement` | 1.0.2 | H | NOT_RUN | Repeat with each actually learned seal intended for use and with no seal. Observe the seal-dependent result. Fury taunt requires Fury; the macro does not select/apply a seal or automatically recover from failure. | | `paladin-repentance` | 1.0.2 | H | NOT_RUN | Confirm actual talent and eligible Humanoid recipient; exercise a wrong-type first match. `/stopattack` is unconditional, including no-target and failed-cast cases. Record control behavior and how existing periodic damage affects it; stopping auto-attack does not remove that damage. | | `paladin-hammer-of-wrath` | 1.0.3 | H | NOT_RUN | Test above and below the source-supported health threshold, and the boundary where it can be measured. A matching above-threshold mouseover must not fall through to an eligible low-health target. Record the client's actual eligibility and error; the macro does not inspect health. | | `paladin-blessing-might` | 1.1.2 | A | NOT_RUN | Observe the intended recipient and replacement/coexistence behavior with this caster's prior blessing. Do not claim a result solely from a tooltip or source description. | | `paladin-blessing-wisdom` | 1.1.2 | A | NOT_RUN | Observe the intended recipient and replacement/coexistence behavior with this caster's prior blessing. | | `paladin-blessing-kings` | 1.1.2 | A | NOT_RUN | Confirm the actual learned spell without inventing a talent requirement. Observe recipient and replacement/coexistence behavior with this caster's prior blessing. | | `paladin-blessing-sacrifice` | 1.0.2 | — | NOT_RUN | Friendly mouseover precedes friendly current target; there is no automatic self fallback. Test an eligible party member and a matched non-party/ineligible friendly ahead of a valid party target. Record actual party/raid eligibility, transferred damage, caster health and blessing replacement. Deliberately selected self is an eligibility observation, not proven support from the absence of a fallback. | | `paladin-greater-might` | 1.1.2 | A | NOT_RUN | Test actual eligible group/raid/class scope with same-class and different-class members. Record Symbol of Kings before/after and a missing-reagent attempt. Any exemption needs actual character/client evidence. The result may affect more than the selected recipient. | | `paladin-greater-wisdom` | 1.1.2 | A | NOT_RUN | Apply the same group/class/reagent cases. Observe what Alt self or automatic self fallback does to eligible members of the caster's class; do not describe it as necessarily single-recipient. | | `paladin-greater-kings` | 1.1.2 | A | NOT_RUN | Apply the same group/class/reagent cases, including any actual exemption and the wider effect of self fallback. | | `paladin-divine-protection` | 1.0.2 | C | NOT_RUN | There is no `/stopcasting`; compare behavior during an existing cast with the ordinary ability. Observe protection, physical-action restrictions and target changes. Check that cancellation leaves Forbearance intact and does not allow an otherwise blocked immunity cast. | | `paladin-divine-shield` | 1.0.2 | C | NOT_RUN | There is no `/stopcasting`. Observe behavior during a cast, immunity, damage restrictions and enemy target changes. Cancellation must be tested independently of Forbearance; it must not be described as clearing that restriction. | | `paladin-righteous-fury` | 1.0.3 | C | NOT_RUN | Observe the aura and actual current threat-mode effect without inheriting numerical source values. Removing the aura is neither an authored taunt nor a command to erase accumulated threat. | | `paladin-holy-shield` | 1.0.2 | — | NOT_RUN | Confirm actual learned talent and shield requirement. Exercise ready/no-channel and a legitimate active channel available to this character; if none is available, retain that case as not run. Test without a shield; the macro does not equip one or guarantee a block. Where actually learned, verify highest-known-rank selection, including source-recorded rank 4 / level 60 / ID `1239570`; source presence alone does not pass acquisition. | | `paladin-swift-judgement` | 1.1.1 | H with Shift | NOT_RUN | Without Shift there is no authored cast. With Shift, hostile mouseover precedes hostile target. Confirm actual learned spell and Bulwark prerequisite. Test intended learned seals and no seal. After the actual GCD, observe ordinary Judgement and Swift cooldowns in both directions; do not describe Swift as a Judgement reset or promise an independent cooldown without the result. | | `paladin-turn-undead` | 1.0.2 | H | NOT_RUN | Exercise the base Undead restriction; test Demons with/without actual passive Turn Evil when those setups are available. The cast name remains Turn Undead. `/stopattack` is unconditional even with no valid recipient or a failed cast. Observe periodic-damage interaction and fleeing behavior in a controlled location. | | `paladin-hybrid-holy-light-judgement` | 1.1.2 | P | NOT_RUN | With Fury active, an enemy selected and a friendly mouseover, observe that the healing branch suppresses the intended Judgement/taunt. Retain dedicated tank taunt access. Record actual range differences and seal-dependent Judgement effect; failed healing must not become an attack. | | `paladin-hybrid-flash-holy-strike` | 1.1.2 | P | NOT_RUN | Test weapon/melee/facing conditions and a matching unlearned or failed healing branch without strike fallback. A ranged healer must not receive an implied ranged strike. Compare actual swing state for successful/failed strikes; no `/startattack` is present. Retain independent tank strike access. | | `paladin-hybrid-holy-shock` | 1.1.2 | P | NOT_RUN | Hostile mouseover is ignored and a friendly current target takes priority. Observe that healing/damage use the same Holy Shock availability. Include self fallback, changed/lost recipients and the separate Vigil follow-up; this key does not remember the Vigil recipient. | ## Cross-record recipient experiment — NOT_RUN Use actual learned Light's Vigil and Holy Shock with their real talent requirements met. Arrange **hostile mouseover A** and **living friendly current target B**, with no modifier held. Use a controlled situation where the hostile attempt is intentional. Record cursor/frame, selected target and actual cast recipient, rather than inferring them from the icon. 1. Press `paladin-lights-vigil` **1.1.1** once. Its authored branch ignores hostile mouseover A and selects friendly target **B**. 2. Press `paladin-holy-shock` **1.1.2** separately, keeping those recipients unchanged. Its authored hostile-mouseover branch selects **A**, not B. 3. Restore ready conditions and the intended setup, then repeat Vigil before a separate press of `paladin-hybrid-holy-shock` **1.1.2**. This healing-first key ignores hostile mouseover A and selects friendly target **B**. 4. Record whether each attempted spell succeeds and whether the intended follow-up effect occurs. The branch expectations above do not prove range, line of sight, spell availability or effect application. 5. Repeat with the pointer moved to a different friendly recipient between presses, and with the friendly recipient lost. Verify the actual new branch; neither key stores or locks the setup recipient. A self or hostile fallback must not be mistaken for a successful same-recipient follow-up. ### Vigil self-eligibility diagnostic — NOT_RUN / outcome unknown The current record does not advertise self eligibility. Deliberately selecting the player or using the player's own frame can still satisfy a friendly unit condition; absence of `[@player]` does not prove that the macro cannot attempt self. Record those explicit attempts and the client's actual result separately. Also test no friendly recipient, where the authored macro has no cast branch. **Do not use UNKNOWN or NOT_RUN as a positive expected result.** A successful cast, rejection or unavailable setup needs its own evidence and review. This diagnostic entry itself grants no advertised self-cast support, tested badge or permission to add a self fallback. ## Historical quarantine — all NOT_RUN These three records are excluded from the ordinary current test pack. Their old bodies and stable IDs are retained for history. A current-looking icon, manual learned-spell confirmation or one surviving alternative does not reinstate the combined historical design. | Exact macro ID | Version | State | Separate research question; no current success promise | | --- | --- | --- | --- | | `paladin-holy-preparation` | 1.0.2 | NOT_RUN | The official Holy talent removal note and absence from the fresh spellbook/descriptor corpus withdraw current source support for Voice of Truth. Existing-character ownership has not been observed here. Capture actual identity/acquisition if investigating; do not assume the default works. Direct Divine Favor remains a separate action when actually learned/talented. **Testing direct Divine Favor, or only the Shift branch, does not validate the historical Voice of Truth / Divine Favor selector.** | | `paladin-blessing-light` | 1.1.2 | NOT_RUN | Preserve the official removal history and literal old body. Do not import/recommend this as a routine current action. Unexpected actual ownership is new acquisition research requiring review, not automatic reinstatement or a passed macro. | | `paladin-blessing-salvation` | 1.1.2 | NOT_RUN | The old blessing was replaced before the later removal of Salvation, Mass Salvation and Revoke. Do not silently substitute any cast name. Its historical self fallback can affect tank threat; investigate only deliberately after actual identity/acquisition is established. Current availability or successful behavior is not promised. | Historical acquisition findings go into a separate research record. Any later proposal to restore or redesign one of these actions needs evidence review and the project's versioning discipline before ordinary current-macro validation. No result in this appendix changes canonical status or fills `tested_builds`.