Loothing

This is a live hotfix for master looters who had the response and vote data in the session, but could not see it in the council table.

Fixed

• Master looters now always see council-table responses and vote counts. The council table no longer depends on only one session-level ML check when deciding whether to show candidate responses, vote counts, voter names, or notes. If the session, canonical ML, or current global ML state says you are the master looter, Loothing gives you the full ML view.
• ML observer mode now handles reload and handoff edge cases more safely. During brief state disagreement after reloads, ML handoffs, or session restore, the actual master looter is no longer treated like a restricted observer.
• Reload restore no longer lets stale cached ML state hide or leak council data. Loothing now marks restored ML state as unverified until the live group check runs, aligns restored session ML with the latest MLDB/global identity, and avoids using stale cached globals for sensitive vote visibility.
• Remote observer permissions are respected without mutating local saved settings. Regular observers now use the ML's transmitted observer permissions for visibility checks instead of falling back to their own local observer settings.
• ML Observer Mode is now deterministic when the ML is also on council. Being on the observer list by itself no longer downgrades an ML council member, but enabling ML Observer Mode still intentionally prevents that ML from casting council votes while keeping full visibility and ML controls.
• Observer roster state survives reload restore. Explicit observers and open-observation settings are now cached with the reconnect state so observers keep the correct council-table access until the next full sync arrives.
• Hidden vote-count privacy now applies consistently in results views. When vote counts are hidden, the Results panel no longer exposes per-candidate vote totals, percentages, ranking-by-votes row order, or ranked-choice round counts to non-ML viewers.
• Session item rows now use the same ML visibility decision as the council table. During reload or handoff edge cases, item-row vote counts, action buttons, and context-menu actions no longer disagree with the rest of the session UI.

Diagnostics

• /lt diag now reports the visibility decision directly. The player section now includes whether the active session thinks you are ML, whether ML visibility is currently verified, which source granted ML visibility, and whether the UI is allowed to show responses, vote counts, and voter identities. This makes future reports much easier to diagnose from a single pasted diagnostics block.