Fix syntax error positions being offset when comments precede brackets

[Copilot]

validate_syntax() used token-list indices for position reporting, but tokenize() strips comments first — so any non-token characters before a bracket caused the reported char position to be wrong.

Example: >comment[ would report char 2 (index in token list) instead of char 9 (original source position).

Changes

• tokens.py — Added internal _tokenize_with_positions() that returns (Token, original_char_index) tuples, preserving each token's position in the original source string.
• interpreter.py — validate_syntax() now accepts position-annotated tokens and uses original_char_index for error messages; run() uses _tokenize_with_positions() for both validation and token extraction.
• tests/test_interpreter.py — Added parametrized cases with comments before brackets to assert correct original-source positions are reported.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.07%. Comparing base (65287d3) to head (0654aae).
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #72      +/-   ##
==========================================
+ Coverage   99.04%   99.07%   +0.03%     
==========================================
  Files          10       10              
  Lines         209      217       +8     
  Branches       31       33       +2     
==========================================
+ Hits          207      215       +8     
  Partials        2        2              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.