S2Cell.GetDistanceToEdge test failure

[jmr]

https://github.com/google/s2geometry/actions/runs/20560316982/job/59050110468?pr=453

[ RUN      ] S2Cell.GetDistanceToEdge
/home/runner/work/s2geometry/s2geometry/src/s2/s2cell_test.cc:759: Failure
The difference between expected_min.radians() and actual_min.radians() is 4.2146848233670653e-08, which exceeds 3e-8, where
expected_min.radians() evaluates to 3.1415926535897931,
actual_min.radians() evaluates to 3.1415926114429449, and
3e-8 evaluates to 2.9999999999999997e-08.
Google Test trace:
/home/runner/work/s2geometry/s2geometry/src/s2/s2cell_test.cc:746: Iteration 559

[  FAILED  ] S2Cell.GetDistanceToEdge (376 ms)

The comment says the max error is "about 3e-8", so the tolerance could just be too low.

s2geometry/src/s2/s2cell_test.cc

Lines 758 to 759 in fd727eb

	// Max error for S1ChordAngle as it approaches Pi is about 3e-8.
	EXPECT_NEAR(expected_min.radians(), actual_min.radians(), 3e-8);

@ericveach