Problem

DeepDiff crashes with ValueError or OverflowError when comparing NaN or Inf float values with significant_digits=0:

>>> from deepdiff import DeepDiff
>>> DeepDiff(float('nan'), float('nan'), significant_digits=0)
ValueError: cannot convert float NaN to integer

>>> DeepDiff(float('inf'), float('inf'), significant_digits=0)
OverflowError: cannot convert float infinity to integer

The crash also occurs when NaN/Inf values are nested inside dicts, lists, or other structures.

Root Cause

In number_to_string, when significant_digits=0, the code calls int(number) on the rounded value. For NaN and Inf, round() returns NaN/Inf unchanged, and int() raises because these values cannot be converted to integers.

Fix

Guard the int() conversion with math.isfinite() so that NaN and Inf values pass through without attempting integer conversion. Their string representations ("nan", "inf", "-inf") will compare correctly in the existing string-comparison path.

The fix is a one-line change: if significant_digits == 0 becomes if significant_digits == 0 and math.isfinite(number).

Additional Note

When significant_digits > 0, NaN and Inf do not crash because round(nan, n) returns NaN and the string formatting produces "nan" which compares equal to another "nan". This behavior is preserved.