Add pow and powi support

[orejevich]

This PR adds proper handling of comptime_float and comptime_int to std.math.pow and std.math.powi.

Changes:

• pow now supports comptime_float by internally casting to f128 during computation
• powi now supports comptime_int directly, with manual overflow and underflow checks
• Matches existing behavior for runtime types, including special cases like 0^0 and negative exponents

Testing:

• Added new tests for comptime_int and comptime_float
• Verified all local tests pass for modified files

Let me know if anything should be adjusted — happy to tweak if needed!