FEAT: fabs ufunc implementation

[SwayamInSync]

Copilot Summary

This pull request adds support for the np.fabs universal function (ufunc) for the custom QuadPrecision dtype, ensuring it behaves identically to np.absolute for real numbers. The implementation includes both the underlying C++ registration and comprehensive tests for various input values. The release tracker documentation is also updated to reflect this new support.

Feature addition:

• Added the fabs ufunc for the QuadPrecision dtype in the C++ extension by registering it in init_quad_unary_ops. This ensures that np.fabs can be used with quad-precision numbers.

Testing:

• Introduced a new parameterized test test_fabs in test_quaddtype.py to verify the correctness of np.fabs for QuadPrecision scalars and arrays, covering positive, negative, zero, special, and small values. The test also checks that the sign of zero is positive after applying fabs.

Documentation:

• Updated release_tracker.md to mark fabs as supported for both the ufunc and its corresponding Python API, reflecting the new implementation status.

[SwayamInSync]

similar to absolute, just does not handle the complex types (so do we) hence registering absolute as fabs

[juntyr]

LGTM, with the added comment

[SwayamInSync]

Let's take this in as well!