fix: avoid integer overflow in browser average interval/ease columns

[krMaynard]

Linked issue (required)

Fixes #5207

Summary / motivation (required)

The browser's average-interval column computed intervals.iter().sum::<u32>() * 86400 in u32. Intervals are in days, so a single card with an interval beyond ~49,710 days (reachable via a large maximum-interval setting, "set due date", or add-ons) overflowed u32 — panicking in debug builds and rendering a wildly wrong span in release.

The average-ease column had the same shape: eases.iter().sum::<u16>() overflows u16 once ~27 cards are averaged in notes mode.

The fix extracts the two averages into small helpers that accumulate in a wider type (u64 for intervals, u32 for eases).

Steps to reproduce (required, use N/A if not applicable)

1. Set a card's interval to a very large value (e.g. via "Set Due Date" with a large offset, or a high maximum interval).
2. Select it in the Browser with the Average Interval column shown.
3. In a debug build this panics; in release the displayed span is wildly wrong.
4. For ease: select ~30+ cards in notes mode with the Average Ease column shown.

How to test (required)

Checklist (minimum)

• I ran ./ninja check or an equivalent relevant check locally.
• I added or updated tests when the change is non-trivial or behavior changed.

Details

Adds unit tests covering the overflow-triggering inputs for both helpers. The full CI workflow ran against this exact commit: https://github.com/krMaynard/anki-fork/actions/runs/30141307409

Build, lint, and test passed on Linux, macOS, and Windows, and format and minilints passed. The workflow's only failure was the Linux SARIF upload step (Resource not accessible by integration), after the substantive Linux checks had passed.

Before / after behavior (optional)

Before: overflow panic (debug) / garbage value (release) for large intervals or many-card ease averages. After: correct averages via wider accumulation.

Risk / compatibility / migration (optional)

Low risk; display-only computation in the browser table.

UI evidence (required for visual changes; otherwise N/A)

N/A (no visual layout change; corrects an incorrectly-computed cell value)

Scope

• This PR is focused on one change (no unrelated edits).