A plugin for Flake8 finding likely bugs and design problems in your program. Contains warnings that don't belong in pyflakes and pycodestyle:
bug·bear (bŭg′bâr′) n. 1. A cause of fear, anxiety, or irritation: *Overcrowding is often a bugbear for train commuters.* 2. A difficult or persistent problem: *"One of the major bugbears of traditional AI is the difficulty of programming computers to recognize that different but similar objects are instances of the same type of thing" (Jack Copeland).* 3. A fearsome imaginary creature, especially one evoked to frighten children.
It is felt that these lints don't belong in the main Python tools as they
are very opinionated and do not have a PEP or standard behind them. Due to
flake8
being designed to be extensible, the original creators of these lints
believed that a plugin was the best route. This has resulted in better development
velocity for contributors and adaptive deployment for flake8
users.
Install from pip
with:
pip install flake8-bugbear
It will then automatically be run as part of flake8
; you can check it has
been picked up with:
$ flake8 --version
3.5.0 (assertive: 1.0.1, flake8-bugbear: 18.2.0, flake8-comprehensions: 1.4.1, mccabe: 0.6.1, pycodestyle: 2.3.1, pyflakes: 1.6.0) CPython 3.7.0 on Darwin
B001: Do not use bare except:
, it also catches unexpected events
like memory errors, interrupts, system exit, and so on. Prefer except
Exception:
. If you're sure what you're doing, be explicit and write
except BaseException:
. Disable E722 to avoid duplicate warnings.
B002: Python does not support the unary prefix increment. Writing
++n
is equivalent to +(+(n))
, which equals n
. You meant n
+= 1
.
B003: Assigning to os.environ
doesn't clear the
environment. Subprocesses are going to see outdated
variables, in disagreement with the current process. Use
os.environ.clear()
or the env=
argument to Popen.
B004: Using hasattr(x, '__call__')
to test if x
is callable
is unreliable. If x
implements custom __getattr__
or its
__call__
is itself not callable, you might get misleading
results. Use callable(x)
for consistent results.
B005: Using .strip()
with multi-character strings is misleading
the reader. It looks like stripping a substring. Move your
character set to a constant if this is deliberate. Use
.replace()
or regular expressions to remove string fragments.
B006: Do not use mutable data structures for argument defaults. They are created during function definition time. All calls to the function reuse this one instance of that data structure, persisting changes between them.
B007: Loop control variable not used within the loop body. If this is intended, start the name with an underscore.
B008: Do not perform function calls in argument defaults. The call is performed only once at function definition time. All calls to your function will reuse the result of that definition-time function call. If this is intended, assign the function call to a module-level variable and use that variable as a default value.
B009: Do not call getattr(x, 'attr')
, instead use normal
property access: x.attr
. Missing a default to getattr
will cause
an AttributeError
to be raised for non-existent properties. There is
no additional safety in using getattr
if you know the attribute name
ahead of time.
B010: Do not call setattr(x, 'attr', val)
, instead use normal
property access: x.attr = val
. There is no additional safety in
using setattr
if you know the attribute name ahead of time.
B011: Do not call assert False
since python -O
removes these calls.
Instead callers should raise AssertionError()
.
B012: Use of break
, continue
or return
inside finally
blocks will
silence exceptions or override return values from the try
or except
blocks.
To silence an exception, do it explicitly in the except
block. To properly use
a break
, continue
or return
refactor your code so these statements are not
in the finally
block.
B013: A length-one tuple literal is redundant. Write except SomeError:
instead of except (SomeError,):
.
B014: Redundant exception types in except (Exception, TypeError):
.
Write except Exception:
, which catches exactly the same exceptions.
These have higher risk of false positives but discover regressions that are dangerous to slip through when test coverage is not great. Let me know if a popular library is triggering any of the following warnings for valid code.
B301: Python 3 does not include .iter*
methods on dictionaries.
The default behavior is to return iterables. Simply remove the iter
prefix from the method. For Python 2 compatibility, also prefer the
Python 3 equivalent if you expect that the size of the dict to be small
and bounded. The performance regression on Python 2 will be negligible
and the code is going to be the clearest. Alternatively, use
six.iter*
or future.utils.iter*
.
B302: Python 3 does not include .view*
methods on dictionaries.
The default behavior is to return viewables. Simply remove the view
prefix from the method. For Python 2 compatibility, also prefer the
Python 3 equivalent if you expect that the size of the dict to be small
and bounded. The performance regression on Python 2 will be negligible
and the code is going to be the clearest. Alternatively, use
six.view*
or future.utils.view*
.
B303: The __metaclass__
attribute on a class definition does
nothing on Python 3. Use class MyClass(BaseClass, metaclass=...)
.
For Python 2 compatibility, use six.add_metaclass
.
B304: sys.maxint
is not a thing on Python 3. Use
sys.maxsize
.
B305: .next()
is not a thing on Python 3. Use the next()
builtin. For Python 2 compatibility, use six.next()
.
B306: BaseException.message
has been deprecated as of Python 2.6
and is removed in Python 3. Use str(e)
to access the user-readable
message. Use e.args
to access arguments passed to the exception.
The following warnings are disabled by default because they are controversial. They may or may not apply to you, enable them explicitly in your configuration if you find them useful. Read below on how to enable.
B901: Using return x
in a generator function used to be
syntactically invalid in Python 2. In Python 3 return x
can be used
in a generator as a return value in conjunction with yield from
.
Users coming from Python 2 may expect the old behavior which might lead
to bugs. Use native async def
coroutines or mark intentional
return x
usage with # noqa
on the same line.
B902: Invalid first argument used for method. Use self
for
instance methods, and cls
for class methods (which includes __new__
and __init_subclass__
) or instance methods of metaclasses (detected as
classes directly inheriting from type
).
B903: Use collections.namedtuple
(or typing.NamedTuple
) for
data classes that only set attributes in an __init__
method, and do
nothing else. If the attributes should be mutable, define the attributes
in __slots__
to save per-instance memory and to prevent accidentally
creating additional attributes on instances.
B950: Line too long. This is a pragmatic equivalent of
pycodestyle
's E501: it considers "max-line-length" but only triggers
when the value has been exceeded by more than 10%. You will no
longer be forced to reformat code due to the closing parenthesis being
one character too far to satisfy the linter. At the same time, if you do
significantly violate the line length, you will receive a message that
states what the actual limit is. This is inspired by Raymond Hettinger's
"Beyond PEP 8" talk and
highway patrol not stopping you if you drive < 5mph too fast. Disable
E501 to avoid duplicate warnings.
To enable these checks, specify a --select
command-line option or
select=
option in your config file. As of Flake8 3.0, this option
is a whitelist (checks not listed are being implicitly disabled), so you
have to explicitly specify all checks you want enabled. For example:
[flake8] max-line-length = 80 max-complexity = 12 ... ignore = E501 select = C,E,F,W,B,B901
Note that we're enabling the complexity checks, the PEP8 pycodestyle
errors and warnings, the pyflakes fatals and all default Bugbear checks.
Finally, we're also specifying B901 as a check that we want enabled.
Some checks might need other flake8 checks disabled - e.g. E501 must be
disabled for B950 to be hit.
If you'd like all optional warnings to be enabled for you (future proof
your config!), say B9
instead of B901
. You will need Flake8 3.2+
for this feature.
Note that pycodestyle
also has a bunch of warnings that are disabled
by default. Those get enabled as soon as there is an ignore =
line
in your configuration. I think this behavior is surprising so Bugbear's
opinionated warnings require explicit selection.
Just run:
python tests/test_bugbear.py
MIT
- Ignore keywords for B009/B010
- Silence B009/B010 for non-identifiers
- State an ignore might be needed for optional B9x checks
- Fix error on attributes-of-attributes in except (...): clauses
- Allow continue/break within loops in finally clauses for B012
- For B001, also check for
except ():
- Introduce B013 and B014 to check tuples in
except (..., ):
statements
- Warn about continue/return/break in finally block (#100)
- Removed a colon from the descriptive message in B008. (#96)
- Fix .travis.yml syntax + add Python 3.8 + nightly tests
- Fix black formatting + enforce via CI
- Make B901 not apply to __await__ methods
- allow 'mcs' for metaclass classmethod first arg (PyCharm default)
- Introduce B011
- Introduce B009 and B010
- Exclude immutable calls like tuple() and frozenset() from B008
- For B902, the first argument for metaclass class methods can be "mcs", matching the name preferred by PyCharm.
- black format all .py files
- Examine kw-only args for mutable defaults
- Test for Python 3.7
- packaging fixes
- graduated to Production/Stable in trove classifiers
- introduced B008
- bugfix: Also check async functions for B006 + B902
- introduced B903 (patch contributed by Martijn Pieters)
- bugfix: B902 now enforces cls for instance methods on metaclasses and metacls for class methods on metaclasses
- introduced B902
- bugfix: opinionated warnings no longer invisible in Syntastic
- bugfix: opinionated warnings stay visible when --select on the command-line is used with full three-digit error codes
- bugfix: opinionated warnings no longer get enabled when user specifies
ignore =
in the configuration. Now they require explicit selection as documented above also in this case.
- bugfix: B007 no longer crashes on tuple unpacking in for-loops
- introduced B007
- bugfix: remove an extra colon in error formatting that was making Bugbear errors invisible in Syntastic
- marked as "Beta" in trove classifiers, it's been used in production for 8+ months
- introduced B005
- introduced B006
- introduced B950
- bugfix: don't raise false positives in B901 on closures within generators
- gracefully fail on Python 2 in setup.py
- introduced B004
- introduced B901, thanks Markus!
- update
flake8
constraint to at least 3.0.0
- introduced B003
- bugfix: don't omit message code in B306's warning
- change dependency on
pep8
to dependency onpycodestyle
, updateflake8
constraint to at least 2.6.2
- introduced B306
- bugfix: don't crash on files with tuple unpacking in class bodies
- introduced B002, B301, B302, B303, B304, and B305
- packaging herp derp
- bugfix: include tests in the source package (to make
setup.py test
work for everyone) - bugfix: explicitly open README.rst in UTF-8 in setup.py for systems with other default encodings
- first published version
- date-versioned
Glued together by Łukasz Langa. Multiple improvements by Markus Unterwaditzer, Martijn Pieters, Cooper Lees, and Ryan May.