A method call fails, but is the line that called it actually wrong? In Python, a stack trace—usually called a traceback—shows the calls that led to an exception. Read it to find where execution failed before deciding what to fix. section 13.1 section 13.1.1
Listing 13.1 Showing a traceback when running some code
class Task:
def __init__(self, title, urgency):
self.title = title
self.urgency = urgency
def _update_db(self):
# update the record in the database
print("update the database")
def update_urgency(self, urgency):
self.urgency = urgency
self.update_db() #1
task = Task("Laundry", 3)
task.update_urgency(4) #2
# output the following error:
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "<stdin>", line 10, in update_urgency
AttributeError: 'Task' object has no attribute 'update_db'
Run this example to see two calls in the traceback: the call to `update_urgency` and the failing call inside that method. section 13.1
Read the traceback from top to bottom. Its first call shows what started the failing sequence; its last call points to the operation that directly raised the exception. The final line gives the exception type and description. Here, `AttributeError` says that `Task` has no `update_db` attribute. The method is actually named `_update_db`, so change the call inside `update_urgency` and run the code again. section 13.1.2 section 13.1.4
When using a traceback to locate a bug:
- Start at the bottom: inspect the exception message and the last call’s method name and line number. section 13.1.4
- If the last call belongs to a dependency, read upward until you reach a call in your own code. section 13.1.5
- Check where the code ran. Console tracebacks identify input as `<stdin>`; a script traceback identifies the file and shows the operation at the reported line. Console line numbers can differ because submitted empty lines are removed. section 13.1.2 section 13.1.3
A traceback is a snapshot taken after execution stops. If you need to inspect each operation as it happens, use a debugger instead. For practice, add methods that call one another and put a failing operation in the last method; the resulting traceback will show a longer sequence of calls. section 13.2.5 section 13.1.6
