If the IBAN includes a non-ASCII alphanumeric character and has just the right length, the IBAN validation crashes with a KeyError before validation can take place.

Fictional example: The supposed IBAN code "Bank München-Wiesn GmbH" gets normalized to "BankMünchenWiesnGmbH"; a string that starts with a valid country code ('BA', ie. Bosnia-Herzegovina) and happens to have the same length as Bosnia-Herzegovina's IBAN format (20 characters). Normally this erroneous IBAN would've been rejected as invalid, but Python throws a KeyError when trying to convert 'ü' to an int right before the validation step.

We therefore need to also check if all characters in the IBAN code are within the expected range, namely [a-zA-Z0-9] (strictly speaking, the IBAN's specification range is only [A-Z0-9], but we can be lenient since Python's `int()` is case-insensitive).

closes odoo/odoo#113733

X-original-commit: fc468f5a
Signed-off-by: Josse Colpaert <jco@odoo.com>