ASCII Luhn validation
DigitTally places synthetic checksum examples alongside your code in a Windows AI coding workflow. Ask a compatible model about digit parity and review its edits, using generated test identifiers rather than real account data.
def luhn_valid(text):
if type(text) is not str or not 2 <= len(text) <= 64:
raise ValueError("luhn.length")
if any(char < "0" or char > "9" for char in text):
raise ValueError("luhn.alphabet")
total = 0
for index, char in enumerate(reversed(text)):
digit = ord(char) - ord("0")
if index % 2:
digit *= 2
if digit > 9:
digit -= 9
total += digit
return total % 10 == 0Connection
Base URL, credentials and a reachable server.
Project
Keep leading zeros and test both parity lengths, rejecting spaces and non-ASCII digits without claiming account validity.
Review before applying
Compare actual behavior with the expected result. If a check fails, investigate it before applying more changes.
def luhn_valid(text):
if type(text) is not str or not 2 <= len(text) <= 64:
raise ValueError("luhn.length")
if any(char < "0" or char > "9" for char in text):
raise ValueError("luhn.alphabet")
total = 0
for index, char in enumerate(reversed(text)):
digit = ord(char) - ord("0")
if index % 2:
digit *= 2
if digit > 9:
digit -= 9
total += digit
return total % 10 == 0Review validation without losing the original text
Validation code has two responsibilities that are easy to conflate: defining acceptable input and deciding what that input means. DigitTally is a Windows AI coding workspace where you can discuss those responsibilities against the source files in your project. Bring a synthetic example, its expected result and the caller that handles rejected input, then ask the model to inspect the current boundary.
The Luhn-check example offers a focused task involving digit strings. Describe how leading zeros should survive and which character set the application accepts before requesting an edit. Ask how the proposed change distinguishes a malformed value from one that fails the checksum. Keeping those outcomes separate helps you review the surrounding form or import code as well as the calculation.
Check the proposal with synthetic fixtures of different lengths, including disallowed characters, and run the relevant project tests. A checksum result is evidence about that calculation, not evidence that an account exists. DigitTally connects to supported OpenAI-compatible or Anthropic endpoints; configure credentials where required. You may choose an unrestricted model if the service offers it, with provider rules, tools and context capacity determined by that connection. Judge the code by the input contract you have established.
Questions about identifier validation work
Why use synthetic identifiers in the conversation?
They can reproduce formatting and checksum behavior without sharing real account details. Choose fixtures that retain the relevant length and leading zeros, and label the expected result so the assistant can examine the code against known cases.
Should digit strings be converted to numbers first?
Review whether conversion would discard information the application needs, such as leading zeros. The illustrated contract treats input as text. Ask the assistant to preserve that contract through validation and into the caller's error handling.
What does a passing checksum establish?
It establishes only that the value satisfies the implemented checksum rule. It does not verify account existence, ownership or permission to transact. Keep those separate application requirements out of claims made about this illustrative validation routine.
Which parity cases need fixtures?
Include both odd and even supported lengths, alongside rejected characters and boundary lengths. Run those cases after reviewing the proposed edit so a change in iteration order does not silently shift which positions are doubled.
DigitTally · Updated 2026-10-25