[chsami]

Problem

Client error telemetry group 2413 (PVirewatchKiller 1.0.4, Microbot 2.6.32): 3 occurrences across 1 report, all on 2026-10-16.

java.lang.NullPointerException: Cannot invoke "net.runelite.api.coords.WorldArea.contains(net.runelite.api.co...
    at net.runelite.client.plugins.microbot.virewatch.PVirewatchScript.lambda$run$3:22
    at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run

Root cause

PVirewatchKillerPlugin.fightArea starts out null. It is only assigned in onGameTick once startingLocation has been captured, or in onConfigChanged. shutDown() sets it back to null. The script loop runs on its own scheduler and called plugin.fightArea.contains(...) without a null check. That throws before the first GameTick after start, and again after shutdown clears the field. The overlay's NPC outline block had the same unguarded dereference, so it hit an NPE on every frame between startUp() adding the overlay and the first tick.

Change

• PVirewatchScript: each iteration reads the local player location, fightArea and startingLocation once, and returns early while any of them is null (the hasFightState helper). The rest of the iteration uses those locals.
• PVirewatchScript: the "away from start" check used != on WorldPoint, which compares identity and so was always true. It now uses equals() (the isAwayFromStart helper). This is a small behaviour change: walkTo is now skipped when the player is already on the start tile. Before, it was called with distance 0, which did nothing.
• PVirewatchKillerOverlay.render: reads plugin.fightArea into one local per frame. Both the fight-area tile loop and the NPC outline block use that local and run only when it is non-null. A shutDown() mid-frame can therefore no longer cause an NPE.
• Plugin version bumped to 1.0.5.

Validation

• ./gradlew build -PpluginList=PVirewatchKillerPlugin test --tests '*PVirewatchScriptTest*': BUILD SUCCESSFUL
• PVirewatchScriptTest: 5 tests, 0 failures. Covers the null fight area or start location before the first tick, the missing local player, and the equals-based start tile comparison.

Known gaps

• The tests cover only the static helpers. The script loop and render() depend on Microbot statics and the NPC cache, so they are not unit-tested and have not been run against a live client.
• onGameTick still calls getLocalPlayer() without a null check. No telemetry reports point there, so it is out of scope.

🤖 Generated with Claude Code

https://claude.ai/code/session_01LjtEMzQHGuq4M1Mpqk9pPV