traceback: py313 updates#12032
traceback: py313 updates #12032
Uh oh!
There was an error while loading. Please reload this page.
Conversation
This comment has been minimized.
This comment has been minimized.
stdlib/traceback.pyi Outdated
| @overload | ||
| defformat_exception_only(exc: BaseExceptionGroup[BaseException], /, *, show_group: bool=False) ->list[str]: ... | ||
| @overload | ||
| defformat_exception_only( | ||
| exc: Unused, /, value: BaseExceptionGroup[BaseException], *, show_group: bool=False | ||
| ) ->list[str]: ... | ||
| @overload | ||
| defformat_exception_only(exc: BaseException|None, /, *, show_group: Literal[False] =False) ->list[str]: ... | ||
| @overload | ||
| defformat_exception_only( | ||
| exc: Unused, /, value: BaseException|None, *, show_group: Literal[False] =False | ||
| ) ->list[str]: ... |
There was a problem hiding this comment.
I don't think we need - or should - special case show_group. If show_group is True, but the exception is not an exception group, it will be printed as normal. If it's False, but it is an exception group, only a single exception will be printed - and that is fine. I think it was a deliberate design decision that this works either way, but cc @iritkatriel.
Also, format_exception_only now has a colorize option, defaulting to False, which has weirdly being hidden in the kwargs.
There was a problem hiding this comment.
I agree we don't need to special-case show_group. I added some special-casing because it seemed to me like it was a case similar to others in typeshed, where we alert users to code that wouldn't necessarily generate an exception at runtime, but would be useless. AFAIK, there's no possible reason to pass show_group=True if the object being passed in is not known to be an ExceptionGroup instance, so it seems indicative of a probable bug if a user does that -- therefore something that a type checker could usefully complain about.
Having said all that, however, I also wasn't sure that this was worth the added complexity here, so I'll remove the special-casing for show_group.
Also,
format_exception_onlynow has acolorizeoption, defaulting toFalse, which has weirdly being hidden in thekwargs.
Yes, but I discussed in the PR description why I was deliberately not adding the private _colorize or colorize keyword arguments to any of the traceback functions in this PR. See e.g. python/cpython#111567 (comment)
There was a problem hiding this comment.
AFAIK, there's no possible reason to pass
show_group=Trueif the object being passed in is not known to be anExceptionGroupinstance, so it seems indicative of a probable bug if a user does that -- therefore something that a type checker could usefully complain about.
I have to admit that my understanding is limited, but I think the point is that if you have a BaseException you don't know whether it's a group or not. And maybe you want to format all group members in case it's a group:
try: ... # might raise an exception group or single exceptionsexceptBaseExceptionasexc: format_exception_group(exc, show_group=True)There was a problem hiding this comment.
yeah, that's a fair point
According to mypy_primer, this change has no effect on the checked open source code. 🤖🎉 |
2 participants
A bunch of functions also have new
**kwargsparameters in Python 3.13. I deliberately haven't added these, as the only additional keyword argument you should be passing is_colorize-- and that's a private implementation detail that deliberately isn't exposed.