LongLiteral.to(FloatType) raises OverflowError instead of returning overflow sentinel

[kevinjqliu]

Describe the bug

LongLiteral.to(FloatType) raises OverflowError for integers outside the float range instead of returning FloatAboveMax / FloatBelowMin.

Reproducer

from pyiceberg.expressions import literal
from pyiceberg.types import FloatType

literal(10**39).to(FloatType())
# OverflowError: float too large to pack with f format

Expected behavior

Values above or below FloatType bounds should return the float overflow sentinels, matching DecimalLiteral.to(FloatType) and StringLiteral.to(FloatType).