code: use current excinfo when formatting chained BaseExceptionGroup

[bysiber]

In repr_excinfo, when the current exception e is a BaseExceptionGroup, the code uses the original excinfo parameter instead of excinfo_ (which tracks the current exception as the method walks the __cause__/__context__ chain):

while e is not None and id(e) not in seen:
    seen.add(id(e))

    if excinfo_:
        ...
        if isinstance(e, BaseExceptionGroup):
            traceback = filter_excinfo_traceback(self.tbfilter, excinfo)   # should be excinfo_
            reprtraceback = ReprTracebackNative(
                format_exception(
                    type(excinfo.value),   # should be excinfo_
                    excinfo.value,         # should be excinfo_
                    ...

On the first loop iteration this doesn't matter since excinfo_ starts as excinfo. But when walking the exception chain, excinfo_ is reassigned to ExceptionInfo.from_exception(e) while excinfo still points to the original. So for any chained BaseExceptionGroup, the traceback filtering is applied to the wrong exception and format_exception formats the original exception instead of the current one.

The non-group branch already correctly uses excinfo_:

else:
    reprtraceback = self.repr_traceback(excinfo_)

This fix replaces the three occurrences of excinfo with excinfo_ inside the BaseExceptionGroup branch.

[The-Compiler]

Similarly to your other PRs I commented on, could you please add a regression test to catch this issue?

[bysiber]

Added a regression test — it creates a chained exception where a BaseExceptionGroup is the context of an outer ValueError, and verifies the output correctly shows both exceptions with the right traceback chain.

[nicoddemus]

To improve clarity, let's rename excinfo_ to current_excinfo.