Ensure FloatRange clamping always returns a float

[sarathfrancis90]

FloatRange(..., clamp=True) could return an int instead of a float when an out-of-range value was clamped to a boundary that had been written as an integer literal.

import click

t = click.FloatRange(0, 1, clamp=True)
print(repr(t.convert("2.5", None, None)))   # 1   (int)  -> should be 1.0
print(repr(t.convert("-2.5", None, None)))  # 0   (int)  -> should be 0.0

FloatRange is declared as _NumberRangeBase[float] / FloatParamType and is documented to return a float, so returning an int here violates the type contract.

Root cause

The constructor stores the boundaries exactly as they were passed in, so FloatRange(0, 1, clamp=True) keeps self.min/self.max as Python ints. FloatRange._clamp returned the stored boundary verbatim for non-open bounds, propagating that int to the caller.

Fix

Coerce the boundary through self._number_class (which is float for FloatRange) before returning it, so a clamped value is always a float. In-range values and IntRange are unaffected.

Tests

Added test_float_range_clamp_returns_float, parametrized over a value above and below the range, asserting type(result) is float. The existing clamp tests use float boundary literals (0.5/1.5) and test_range compares with == (where 1 == 1.0), so they did not catch this. The new test fails on main (AssertionError: <class 'int'> is float) and passes with the change.

fixes #3547

---

• Add tests that demonstrate the correct behavior of the change. Tests fail without the change.
• Add or update relevant docs, in the docs folder and in code.
• Add an entry in CHANGES.rst summarizing the change and linking to the issue.
• Add .. versionchanged:: entries in relevant code docs.