reject int overflow of rows * columns in SparseFieldMatrix

[dxbjavid]

SparseFieldMatrix computes its hash key as row * columns + column in int arithmetic, but unlike OpenMapRealMatrix its constructor never checks that the dimension product fits in an int, so a sparse matrix such as 100000 by 100000 is accepted and the key wraps for cells beyond the first 2^31. Two distinct cells can then share a key, for example setEntry(42950, 0) silently overwrites the value stored at (0, 32704), which seems worse than a failure since nothing signals the corruption. This adds the same NumberIsTooLargeException guard that MATH-679 gave OpenMapRealMatrix, along with a test that fails on master. AI tooling was used to help prepare this change.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 87.17%. Comparing base (f554608) to head (285c3c1).
⚠️ Report is 351 commits behind head on master.

Additional details and impacted files

@@             Coverage Diff              @@
##             master     #331      +/-   ##
============================================
+ Coverage     86.54%   87.17%   +0.62%     
+ Complexity     9787       89    -9698     
============================================
  Files           532      499      -33     
  Lines         35516    33469    -2047     
  Branches       6194     5835     -359     
============================================
- Hits          30738    29176    -1562     
+ Misses         3518     3167     -351     
+ Partials       1260     1126     -134     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.