You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the bug #12885 and #12925 indicate that an overflow exception should be thrown, but when I overflow on a LIST instead of a STRING of the same size I get an out of memory error.
using FCW = cudf::test::fixed_width_column_wrapper<int8_t>;
auto s = cudf::make_list_scalar(FCW({1, 2, 3, 4, 5, 6, 7, 8, 9, 10}));
auto col = cudf::make_column_from_scalar(*s, 214748365);
The scary part is that the OMM happens because the size calculation was done as an int and overflowed so we tried to allocate a very large negative allocation. If I overflow twice the allocation succeeds, but I end up writing data outside of what was actually allocated.
using FCW = cudf::test::fixed_width_column_wrapper<int8_t>;
auto s = cudf::make_list_scalar(FCW({1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20}));
auto col = cudf::make_column_from_scalar(*s, 214748365);
The text was updated successfully, but these errors were encountered:
… utility (#13841)
Internal lists functions `make_lists_column_from_scalar` (used by `make_column_from_scalar`) and `generate_list_offsets_and_validities` (used by `concatenate_list_elements`) are updated to use the `make_offsets_child_column` utility to build the offsets from sizes. This utility handles `size_type` overflow when computing an offsets column in a consistent way (i.e. throwing `std::overflow_error` appropriately).
Closes#13833
Authors:
- David Wendt (https://github.com/davidwendt)
Approvers:
- Nghia Truong (https://github.com/ttnghia)
- Bradley Dice (https://github.com/bdice)
- Robert (Bobby) Evans (https://github.com/revans2)
URL: #13841
Describe the bug
#12885 and #12925 indicate that an overflow exception should be thrown, but when I overflow on a LIST instead of a STRING of the same size I get an out of memory error.
The scary part is that the OMM happens because the size calculation was done as an int and overflowed so we tried to allocate a very large negative allocation. If I overflow twice the allocation succeeds, but I end up writing data outside of what was actually allocated.
The text was updated successfully, but these errors were encountered: