In testing for WT-5680, I saw address sanitizer failures caused by attempting to copy cursor->value into a tombstone during a cursor->remove operation. As well as being wrong, this is inefficient because we'll never read a value from a tombstone.
In testing for WT-5680, I saw address sanitizer failures caused by attempting to copy cursor->value into a tombstone during a cursor->remove operation. As well as being wrong, this is inefficient because we'll never read a value from a tombstone.