Bug report
Bug description:
(I've tagged this as a "docs" issue because help() is arguably a part of the documentation. Also, some of the docs entries have a similar lack of details. I suppose that could be dealt with separately.)
The help() text for (at least some) builtin exceptions doesn't show the type-specific signature for __init__() (or __new__()), instead showing a generic one with (self, /, *args, **kwargs).
Here's an example (in the REPL) 1:
>>>help(ExceptionGroup)| ...
|__init__(self, /, *args, **kwargs)
|Initializeself. Seehelp(type(self)) foraccuratesignature.
| ...
(The help() entry for __new__() is likewise unhelpful.)
There is a similar lack of information in the docs, though a handful of entries have at least some signature information.
All builtin exceptions (best effort catalog):
| exception type | expected* (approximately)
(incl. additional forms) | docs
have
sig |
|---|
Exception
BaseException | __init__(self, /, *args) | |
ExceptionGroup
BaseExceptionGroup | __init__(self, msg, excs, /) | X† |
| StopIteration | __init__(self, value=None, /, *args) | |
| SystemExit | __init__(self, code=None, /, *args) | |
| ImportError | __init__(self, /, name=None, path=None, name_from=None) | |
| OSError | __init__(self, errno, strerror, filename=None, winerror=None, filename2=None, /) | X† |
| ^^^^^ | __init__(self, arg=None, /) | X† |
| ^^^^^ | __init__(self, /, *args) | ‡ |
| NameError | __init__(self, /, *args, name=None) | |
| AttributeError | __init__(self, /, *args, name=None, obj=None) | |
| SyntaxError | __init__(self, msg, details, /) | X† |
| ^^^^^ | __init__(self, msg=None, /, *args) | ‡ |
| UnicodeEncodeError | __init__(self, encoding, obj, start, end, reason, /) | |
| UnicodeDecodeError | __init__(self, encoding, obj, start, end, reason, /) | |
| UnicodeTranslateError | __init__(self, obj, start, end, reason, /) | |
| -- all others (?) -- | -- same as Exception/BaseException -- | |
* I derived expected signatures from the content of Objects/exceptions.c.
† The doc entries for OSError, ExceptionGroup, and SyntaxError imply that they each take keyword arguments but they do not.
‡ The doc entries for OSError and SyntaxError do not include the alternate signature.
CPython versions tested on:
CPython main branch
Operating systems tested on:
No response
Bug report
Bug description:
(I've tagged this as a "docs" issue because
help()is arguably a part of the documentation. Also, some of the docs entries have a similar lack of details. I suppose that could be dealt with separately.)The
help()text for (at least some) builtin exceptions doesn't show the type-specific signature for__init__()(or__new__()), instead showing a generic one with(self, /, *args, **kwargs).Here's an example (in the REPL) 1:
(The
help()entry for__new__()is likewise unhelpful.)There is a similar lack of information in the docs, though a handful of entries have at least some signature information.
All builtin exceptions (best effort catalog):
(incl. additional forms)
have
sig
BaseException
__init__(self, /, *args)BaseExceptionGroup
__init__(self, msg, excs, /)__init__(self, value=None, /, *args)__init__(self, code=None, /, *args)__init__(self, /, name=None, path=None, name_from=None)__init__(self, errno, strerror, filename=None, winerror=None, filename2=None, /)__init__(self, arg=None, /)__init__(self, /, *args)__init__(self, /, *args, name=None)__init__(self, /, *args, name=None, obj=None)__init__(self, msg, details, /)__init__(self, msg=None, /, *args)__init__(self, encoding, obj, start, end, reason, /)__init__(self, encoding, obj, start, end, reason, /)__init__(self, obj, start, end, reason, /)Exception/BaseException--* I derived expected signatures from the content of Objects/exceptions.c.
† The doc entries for
OSError,ExceptionGroup, and SyntaxError imply that they each take keyword arguments but they do not.‡ The doc entries for
OSErrorandSyntaxErrordo not include the alternate signature.CPython versions tested on:
CPython main branch
Operating systems tested on:
No response
Footnotes
FYI, I don't mean to pick on
ExceptionGroup. It happened to be the one for which I first noticed the problem. ↩