fix: keep every bound in a chained comparison

[MohammedAlkindi]

Issue #, if available: none filed.

Description of changes:

Python parses a < b < c as one Compare node holding every operator, but visit_Compare read only ops[0] and comparators[0]. Every bound after the first was discarded, so the emitted OpenQASM silently implemented different logic than the source. On main if 0 <= n < 10 with n = 1000 emits h __qubits__[0]; unconditionally; it now emits x.

The fix walks the chain and folds the pairs with ag__.and_, the helper the upstream logical_expressions pass already uses. Single and mixed chains are unchanged.

Testing done:

494 passed, 2 xfailed. Reverting only the source and keeping the two new tests leaves exactly those two failing, at 492 passed. ruff is clean apart from five pre-existing unused-noqa findings present either way.

Merge Checklist

General

• Read the CONTRIBUTING doc
• Used the commit message format from CONTRIBUTING
• Updated documentation (not applicable)

Tests

• Added tests that prove the fix is effective
• Tests are not configured for a specific region or account

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (bcc6723) to head (a02c0d6).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff            @@
##              main      #124   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           52        52           
  Lines         2290      2296    +6     
  Branches       266       268    +2     
=========================================
+ Hits          2290      2296    +6     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[rmshaffer]

Thank you for taking the initiative to fix this! I think it's in the right direction; just left a couple comments that I'd appreciate if you could address.

[MohammedAlkindi]

All three are in 9121ce0, which I pushed without commenting, so this still shows as blocked.

Docstring rationale removed, and the standalone test is now the i/j lines inside test_comparison_ops_py.

Both chains you named are covered, using your examples. Reverting comparisons.py to main and keeping the new tests:

4 failed, 2 passed
  test_comparison_chained
  test_comparison_chained_multiple_operators      # 4 < a <= b <= c < 8
  test_comparison_chained_with_equality_operator  # 4 < a < b == c < 8
  test_comparison_ops_py

Restoring it: 6 passed. So both do depend on the change.

ruff check src clean. Local only: no Actions run has appeared on this PR, just readthedocs.

[MohammedAlkindi]

Correction to my last line, I read the rollup too quickly. The four workflows did run on 9bfbaa8 and all passed. On 9121ce0 they are sitting at action_required:

Python package                    action_required
Check code format                 action_required
Code Freeze                       action_required
Check long description for PyPI   action_required

So they are waiting on your approval to run, not missing. Sorry for the noise.

[MohammedAlkindi]

All three addressed, thanks for the review.

Worth flagging from the validation you asked for: 4 < a < b == c < 8 is not merely working already. On main it compiles to a > 4 alone — visit_Compare reads ops[0] and comparators[0] and discards the rest of the chain, so the other three bounds are dropped with no error. That test is covering a real silent drop rather than confirming existing behaviour.

Both chain tests now assert the full expected IR instead of substrings, since a substring check passes even when the operands are folded together wrongly. test_comparison_ops_py also gained a three-operator chain whose failing bound is the last one.

[MohammedAlkindi]

This and #125 are both approved and green, but main has moved one commit since (#126, the codecov-action bump), so both now show as behind. Should I update the branches, or is anything else needed before merging?