REGR: exceptions not caught in _call_map_locations #34113
Conversation
| @@ -509,7 +509,7 @@ cdef class PeriodEngine(Int64Engine): | |||
| cdef _get_index_values(self): | |||
| return super(PeriodEngine, self).vgetter() | |||
|
|
|||
| cdef void _call_map_locations(self, values): | |||
| cpdef _call_map_locations(self, values): | |||
There was a problem hiding this comment.
these are still cdef on master, why is this the appropriate fix on 1.0.4?
There was a problem hiding this comment.
why is this the appropriate fix on 1.0.4?
#29232 was a clean-up PR which caused a regression. this is reverting those changes. (applied directly to 1.0.x branch and not touching master, see #34048 (comment))
these are still cdef on master
that's my question to you in the OP.
Could there still be issue on master with the exceptions not being caught?
There was a problem hiding this comment.
Could there still be issue on master with the exceptions not being caught?
I can't reproduce the bugs from #32409. I'm guessing you bisected to find the commit that introduced (reintroduced?) the bug, did you find what commit fixed it?
There was a problem hiding this comment.
yep fixed on master, see #32409 (comment)
but not sure whether the PR that fixed it #31294 should be backported. I can't immediately see the relation.
I see this PR as the safer option as it's a simple revert.
There was a problem hiding this comment.
I see this PR as the safer option as it's a simple revert.
Works for me
black pandasgit diff upstream/master -u -- "*.py" | flake8 --diffNOTE: issue 'fixed' on master. This PR against 1.0.x branch
cc @jbrockmendel Could there still be issue on master with the exceptions not being caught?