New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Add context manager for the "try: ... except: pass" pattern #60010
Comments
It is a somewhat common pattern to write: try: To search examples in the standard library (or any other code base) use: $ egrep -C2 "except( [A-Za-z]+)?:" *py | grep -C2 "pass" In the Python2.7 Lib directory alone, we find 213 examples. I suggest a context manager be added that can ignore specifie exceptions. Here's a possible implementation: class Ignore:
''' Context manager to ignore particular exceptions'''
def __init__(self, *ignored_exceptions):
self.ignored_exceptions = ignored_exceptions
def __enter__(self):
return self
def __exit__(self, exctype, excinst, exctb):
return exctype in self.ignored_exceptions The usage would be something like this: with Ignore(IndexError, KeyError):
print(s[t]) Here's a real-world example taken from zipfile.py: def _check_zipfile(fp):
try:
if _EndRecData(fp):
return True # file has correct magic number
except IOError:
pass
return False With Ignore() context manager, the code cleans-up nicely: def _check_zipfile(fp):
with Ignore(IOError):
return bool(EndRecData(fp)) # file has correct magic number
return False I think this would make a nice addition to contextlib. |
Hmm, the __exit__ method was doing exact matches by exception type, so KeyError wouldn't match LookupError or Exception. There are probably a number of ways to fix this, but it may be easiest to use the builtin exception catching mechanisms: class Ignore:
''' Context manager to ignore particular exceptions'''
def __init__(self, *ignored_exceptions):
self.ignored_exceptions = ignored_exceptions
def __enter__(self):
return self
def __exit__(self, exctype, excinst, exctb):
if exctype is not None:
try:
raise
except self.ignored_exceptions:
return True |
Why not just: issubclass(exctype, self.exception_types)? |
Yes, something along those lines would be *much* better: class Ignore:
''' Context manager to ignore particular exceptions'''
def __init__(self, *ignored_exceptions):
self.ignored_exceptions = ignored_exceptions
def __enter__(self):
return self
def __exit__(self, exctype, excinst, exctb):
return exctype and issubclass(exctype, self.ignored_exceptions) |
I'd just write it with @contextmanager. Making it easier to cleanly factor out exception handling is one of the main reasons that exists. @contextmanager
def ignored(*exceptions):
"""Context manager to ignore particular exceptions"""
try:
yield
except exceptions:
pass While the class based version would likely be fractionally faster, the generator based version is more obviously correct. |
I think I'm -1 on this. This saves two lines, but makes the code less explicit, it can't be used for try/except/else or try/except/except, it requires an extra import, the implementation is simple enough that it doesn't necessary need to be in the stdlib, it might encourage bad style, and it's slower. Some of these downsides are indeed somewhat weak, but the upside of saving two lines is quite weak too IMHO. |
I think the zipfile example is really a bad example. IMO, it would try: i.e. there shouldn't be a pass statement at all in this code, and the if can be dropped whether you use try-except or with. |
(Note: I'm not yet convinced this is a good idea. I'm definitely considering it, though) As with many context managers, a key benefit here is in the priming effect for readers. In this code:
the reader doesn't know that (A, B, C) exceptions will be ignored until the end. The with statement form makes it clear before you start reading the code that certain exceptions won't propagate: with ignored(A, B, C):
# Whatever I'm not worried that it makes things less explicit - it's pretty obvious what a context manager called "ignored" that accepts an arbitrary number of exceptions is going to do. One other thing it does is interact well with ExitStack - you can stick this in the stack of exit callbacks to suppress exceptions that you don't want to propagate. |
The "focus" is mostly on what it's being executed rather than what it's being ignored though. "Do this operation and ignore these exceptions if they occur"
It's still understandable, but while I'm familiar with the semantics of try/except, I wouldn't be sure if e.g. this just ignored those specific exceptions or even their subclasses without checking the doc/code.
This seems a good use case. |
If this is desirable then I think it would be better as a classmethod of Exception: with KeyError.trap():
do_something() |
While the classmethod version has some appeal, it doesn't extend well to handling multiple exception types. I'm -0 on this, in any event. I think the original code is more clear. Why force people to learn (or recognize) a second idiom for something so simple? |
New changeset 406b47c64480 by Raymond Hettinger in branch 'default': |
FTR, Raymond and I discussed this on IRC and I gave it a +1 before he committed it. The advantage the callable form has over a class method is that it readily scales to ignoring multiple exception types in a single with statement. |
Hi Raymond, This is a brilliant idea, but before it hits the streets, couldn't you possibly consider extending it with a kwarg to control the depth of the exception stack? The use case I have for that are snippets like this:
Or else I could write this as with ignored(ValueError, TypeError):
a()
with ignored(ValueError, TypeError):
b()
with ignored(ValueError, TypeError):
c() ... but either way it looks bad. This looks a bit better to me: with ignored(ValueError, TypeError, depth=3):
a()
b()
c() If you deem this to be unacceptably unpythonic, then please ignore my suggestion. |
Actually, please disregard my idea. It's way to dangerous, especially in the case of multiple exceptions to ignore :-( |
Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.
Show more details
GitHub fields:
bugs.python.org fields:
The text was updated successfully, but these errors were encountered: