fix(client): count spell colors in dominant-color detection (shard variant names)

[boskodev790]

Problem

The dominant-color helpers in viewmodel/dominantColor.ts (getDominantManaColor, getDeckDominantColor) key their shard→color lookup by MTG symbols (W/U/B/R/G) and split each shard on "/" to catch hybrids like "W/U".

But the engine does not serialize ManaCost shards as symbols — it serializes them as Rust variant names: "White", "Red", "WhiteBlue", "PhyrexianRed", etc. This is the same convention costLabel.ts documents on SHARD_ABBREVIATION ("Converts Rust ManaCostShard variant names to MTG abbreviations") and that keywordProps.ts relies on, and it's the format used by the shared gameObjectFactory and every other viewmodel/game test (shards: ["Red"], ["White"], ["WhiteBlue"]).

Because "White".split("/") is ["White"] and SHARD_TO_COLOR["White"] is undefined, no nonland card's mana cost ever contributed a color. Dominant-color detection collapsed to a lands-only signal, so components/board/BattlefieldBackground.tsx ignored spell colors entirely (a mono-red spellslinger deck with basic Islands would tint blue).

The previous dominantColor.test.ts masked this: its two spell-cost cases fed the fake ["W","W"] / ["R"] format — the lone place in the repo disagreeing with the factory and every other test.

Fix

Bridge each shard through the canonical SHARD_ABBREVIATION map (variant name → symbol) before splitting on "/", then look up each colored half — mirroring exactly how keywordProps/costLabel already interpret shards. This makes mono, hybrid (WhiteBlue → W/U → White + Blue), and Phyrexian (PhyrexianRed → R/P → Red) shards all count correctly, while colorless/snow/X halves are ignored.

Tests

viewmodel/__tests__/dominantColor.test.ts:

• Corrected the two spell-cost cases to the real engine format (["White","White"], ["Red"]). With the pre-fix symbol-keyed lookup these now return null/wrong color — proven fail-before, pass-after.
• Added a hybrid case: ["WhiteBlue"] + ["Blue"] → Blue wins (hybrid contributes to both White and Blue).
• Added a land+spell case where the spell color must win, which would silently mis-resolve if spell shards were dropped.

Verified with vitest: all 11 tests pass after the fix; reverting only the source change fails exactly the 3 spell-color tests (discriminating).

Model: claude-opus-4-8

[gemini-code-assist]

Warning

You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again!

[superagent-security]

🚨 Contributor flagged. Click here for more info: Superagent Dashboard