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
Unlike other index variables in io.cpp, point_offset is simply defined as an int (see line 178). As a result, I am experiencing crashes when concatenating large pointclouds (points=3,000,000,000) with a large number of fields (15 floats). It crashes on the memcpy on line 192.
At a minimum, I think it should be changed to std::uint32_t, consistent with the point_step variable which seems to be enough to fix my current issue. However, the point_offset value is currently used in a manner where the maximum value is width * height * point_step which can result point_offset values that significantly exceed the other terms. A larger datatype, or a rework of the loop, seems worth considering to prevent this type of crash in the future.
The text was updated successfully, but these errors were encountered:
Unlike other index variables in io.cpp, point_offset is simply defined as an int (see line 178). As a result, I am experiencing crashes when concatenating large pointclouds (points=3,000,000,000) with a large number of fields (15 floats). It crashes on the memcpy on line 192.
At a minimum, I think it should be changed to std::uint32_t, consistent with the point_step variable which seems to be enough to fix my current issue. However, the point_offset value is currently used in a manner where the maximum value is width * height * point_step which can result point_offset values that significantly exceed the other terms. A larger datatype, or a rework of the loop, seems worth considering to prevent this type of crash in the future.
The text was updated successfully, but these errors were encountered: