test_stochasticNoiseGenerator

Focused unit tests for the pluggable Gaussian noise generators used by the native stochastic integrators (stochasticNoiseGenerator.h). These exercise the public behaviors directly, rather than only through an integrator’s happy path:

• PrescribedGaussianNoiseGenerator: cursor-based FIFO replay, remaining()/clear(), omitted-dZ zero-filling, and the error paths (queue exhaustion, mismatched dW length, undersized explicit dZ).

• RandomGaussianNoiseGenerator: seed repeatability, Wiener-only draws, and sqrt(h) scaling.

test_stochasticNoiseGenerator.test_native_output_buffer_methods_are_not_python_apis()

Expose allocation-bearing return APIs instead of ineffective buffer copies.

test_stochasticNoiseGenerator.test_random_incrementsScaleWithSqrtH()

Each increment is drawn as sqrt(h)*N(0,1); with a fixed seed, halving h scales the draw by sqrt(1/2) exactly (same underlying standard-normal sequence).