Add life total change logging to game log

[TytaniumDev]

I'm building a tool that uses headless Forge simulations to play commander games and then analyze the logs (more detail in #9844 ). Player life changes are currently not logged, which makes it hard to visualize game states at a high level.

This PR adds life logging. The code was primarily written by Claude but I've verified it and made sure tests were written to validate all the scenarios I could think of.

Summary

• GameEventPlayerLivesChanged is already fired on every life total change but had no visitor in GameLogFormatter — the event was silently discarded
• Adds new GameLogEntryType.LIFE enum value, replacing the // Where's life loss? comment that was next to DAMAGE
• Adds visit(GameEventPlayerLivesChanged) to GameLogFormatter producing log lines like: Life: Ai(1)-Doran Big Butts 40 -> 37
• Adds localized message key lblLogPlayerLifeChange to all language files (English fallback for non-English locales)
• Adds 6 targeted tests in GameLogLifeChangeTest

No behavior changes to existing log entries.

Closes #9844

Example game log output

Combat damage → life change

Two Nest Robbers (2/1 haste) attack an opponent at 4 life:

[COMBAT]       p1 assigned Nest Robber (1) and Nest Robber (2) to attack p2.
[COMBAT]       p2 didn't block Nest Robber (1).
               p2 didn't block Nest Robber (2).
[DAMAGE]       Nest Robber (1) deals 2 combat damage to p2.
[DAMAGE]       Nest Robber (2) deals 2 combat damage to p2.
[LIFE]         Life: p2 4 -> 0

Lifegain from spell resolution

AI casts Lightning Helix (3 damage + 3 life) targeting opponent at 3 life:

[STACK_ADD]    p1 cast Lightning Helix targeting [p2]
[DAMAGE]       Lightning Helix (3) deals 3 non-combat damage to p2.
[LIFE]         Life: p2 3 -> 0
[LIFE]         Life: p1 20 -> 23
[STACK_RESOLVE] Lightning Helix (3) - Lightning Helix (3) deals 3 damage to p2. p1 gains 3 life.

Test results

Existing tests (63/63 passing, no regressions)

[INFO] Tests run: 63, Failures: 0, Errors: 0, Skipped: 0
       -- in forge.ai.simulation.GameSimulationTest

Full build with Checkstyle (all 13 modules passing)

[INFO] Reactor Summary for Forge Parent 2.0.11-SNAPSHOT:
[INFO]
[INFO] Forge Parent ....................................... SUCCESS
[INFO] Forge Core ......................................... SUCCESS
[INFO] Forge Game ......................................... SUCCESS
[INFO] Forge AI ........................................... SUCCESS
[INFO] Forge Gui .......................................... SUCCESS
[INFO] Forge Mobile ....................................... SUCCESS
[INFO] Forge Mobile Dev ................................... SUCCESS
[INFO] Forge .............................................. SUCCESS
[INFO] Forge iOS .......................................... SUCCESS
[INFO] Forge LDA .......................................... SUCCESS
[INFO] Adventure Editor ................................... SUCCESS
[INFO] Forge Android ...................................... SUCCESS
[INFO] Forge Installer .................................... SUCCESS
[INFO] BUILD SUCCESS

Build command: mvn -U -B clean install -P windows-linux

Test plan

• mvn -U -B clean install -P windows-linux passes (includes Checkstyle)
• Existing GameSimulationTest — all 63 tests pass (no regressions)
• Game log shows [LIFE] entries after [DAMAGE] entries in combat scenarios
• Game log shows [LIFE] entries for both damage-caused loss and spell-caused gain (Lightning Helix)
• Existing log lines (Damage:, Combat:, etc.) are unchanged
• Life: lines show accurate before/after values matching the game state

🤖 Generated with Claude Code

[MostCromulent]

I think this is a good addition, but suspect it doesn't need built-in tests to validate the logging capability as long as we're confident it works? i.e. do other GameLogEntryType values have automated tests associated with them?

[tool4ever]

yea imo AI a bit too overeager: a more generic test for logging would be more valuable

[TytaniumDev]

@tool4ever @MostCromulent would you like me to remove the test, or condense it somehow? Or should I try to have the AI write a more generic logging test in this PR (seems out of scope)?

[MostCromulent]

I'd suggest just removing the tests from this PR - we're doing some work on logging separately and can think about a more general test there if needed.

Really the proof is in the pudding on this one - either the life events show up in the log or they don't.

[TytaniumDev]

@MostCromulent done!