[BUG] Potential NumberFormatException in myAtoi.java with non-ASCII Unicode digits Description

[lionel-w2]

Description

There is a bug in the myAtoi implementation where certain Unicode digit characters (e.g., Arabic-Indic digit '١' U+0661) pass the Character.isDigit() check but cause the parsing logic to fail.

Specifically, when the input contains these characters, the current logic identifies them as digits but fails to accumulate them into a valid numeric string, leading to an empty string being passed to Integer.parseInt(), which throws a NumberFormatException.

Steps to reproduce

Pass a string containing non-ASCII digits, such as "١" (Arabic-Indic digit 1), to the myAtoi function.

The code executes the following sequence:

• Character.isDigit('١') returns true.

• The logic proceeds to parse but fails to handle the non-ASCII mapping.

• Integer.parseInt("") is eventually called (or a similar empty-string conversion).

Result: java.lang.NumberFormatException: For input string: ""

Excepted behavior

The function should follow the "parse as much as possible" rule:

• If non-ASCII digits are not supported, they should be treated as non-digit characters.

• The function should return 0 or the successfully parsed integer part instead of throwing an unhandled exception.

Screenshots

Additional context

No response

[Rimjhim6]

please assign it to me

[lionel-w2]

please assign it to me

Hi @alxkm @Hardvan, @Rimjhim6 is going to handle this. Could you please assign it to her? Thanks!