Task 1: Repair the Syntax

Open starter/stage1_syntax_bug.txt.

1. Copy it to a Python file.
2. Run it without editing and record the complete final traceback line.
3. Inspect the reported line and the line before it.
4. Make one focused correction.
5. Retest with 85 and 50.

Expected classifications: alarm and normal.

Check-Off A: Syntax Evidence

Show your partner:

• the original SyntaxError message;
• the exact grammar rule that was violated;
• the corrected program running with both inputs.

State why exception handling would not be the appropriate repair.

Task 2: Reproduce the Conversion Failure

Call parse_reading with:

"T1,23.5"
"T2,error"
"missing-comma"

For each record, write the expected behavior before running. Record the exception type, message, and failing operation for invalid input.

Task 3: Define the Parser Contract

The repaired function must:

• return (sensor_id, numeric_value) for SENSOR,VALUE;
• raise ValueError("expected SENSOR,VALUE") for the wrong number of fields;
• raise ValueError("value must be numeric") when conversion fails;
• never print or request input.

Use validation for structure and a narrow try block for numeric conversion.

Task 4: Diagnose the Missing Sensor

limits = {"T1": 80.0, "T2": 75.0}
limit = limits["T9"]

Decide which contract applies:

1. T9 must exist, so KeyError reveals invalid program data.
2. T9 is optional, so a documented default is meaningful.
3. The caller should report “unknown sensor” and continue.

Implement option 3 in classify_reading without catching unrelated exceptions.

Task 5: Keep Unknown Separate from Normal

Required results:

classify_reading("T1", 70.0, limits)  # "normal"
classify_reading("T1", 85.0, limits)  # "alarm"
classify_reading("T9", 70.0, limits)  # "unknown"

Do not use a default limit that silently classifies an unknown sensor as normal.

Task 6: Reproduce the Empty Average

Call average_reading([]) and read the traceback from the final line upward.

The contract for this lab is:

Return the average of a non-empty sequence; raise ValueError for an empty sequence.

Repair the function without catching ZeroDivisionError after it happens.

Task 7: Correct the Average

The current function divides by len(values) - 1.

Use this evidence table before editing:

Repair the formula and retain all three cases as regression tests.

Task 8: Do Not Hide Unexpected Defects

Search your file for:

• bare except:;
• except Exception:;
• handlers containing only pass;
• try blocks containing unrelated operations.

Replace each with a specific design or let the unexpected exception propagate.

Task 9: Process Several Records

Implement process_records(records, limits) so it:

1. parses each record;
2. skips malformed records after recording an error message;
3. classifies known and unknown sensors;
4. returns (accepted, errors);
5. keeps processing after an expected bad record.

The starter includes starter/sensor_readings.txt as optional data.

Task 10: Regression Tests

Your test set must include:

• one ordinary valid record;
• one boundary reading equal to its limit;
• one nonnumeric value;
• one record with a missing comma;
• one unknown sensor;
• an empty list passed to average_reading.

Every repaired defect needs a test that would fail if the defect returned.

Instructor-Selected Demonstration

Each student independently completes one action:

• classify a new defect;
• annotate a traceback;
• narrow a try block;
• select a specific handler;
• raise ValueError for a violated contract;
• write a boundary or invalid-input regression test.

You may retry once after focused feedback.

Submission Checklist

• The debugging record contains evidence for every stage.
• All supplied function names remain unchanged.
• No bare except, silent pass, or broad except Exception remains.
• Parsing uses a narrow conversion handler.
• Empty input raises the documented ValueError.
• The logic defect is repaired.
• Normal, boundary, and invalid cases are tested.
• Both returned collections from process_records are explained.