Test Albers round-trip away from the poles

[jitendravjh]

Related to #296.

Could not find a bug here. The implementation reproduces every intermediate of Snyder's ellipsoid worked example, m, q, n, C and ρ, with x and y within 5 mm, and no datum throws at the poles.

The pole error is arcsine conditioning. asin(α/qₚ) reaches an argument of one there so the latitude keeps only half its digits, 2.8189985e-6° predicted against 2.8189981e-6° measured at lat -90. α itself is accurate to 1e-16 so nothing upstream is losing precision.

So this only drops the blanket skip in fwdbwd.jl and keeps a narrow one at lat ±90. Albers gets round-trip coverage everywhere else in both float types. Leaving it to you whether that closes #296.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 92.35%. Comparing base (64227d4) to head (ab81bda).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #366   +/-   ##
=======================================
  Coverage   92.35%   92.35%           
=======================================
  Files          40       40           
  Lines        1818     1818           
=======================================
  Hits         1679     1679           
  Misses        139      139           

☔ 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.