Yes, the source doesn't get completely removed instead of only getting flagged if the file is larger than MAX_SIZE_WO_TREE just like in case when there's no full tree provided.
Theoretically you could introduce a new flag for our case which would behave the same - I just don't think it worth it because the problem is the same: we cannot check the finished download against the full tree because it's not available at the time of the checking (cannot be read from the hash database since it's not saved). In my thinking these are similar enough cases to use this flag to solve the problem of flooding.
Yes, the source doesn't get completely removed instead of only getting flagged if the file is larger than MAX_SIZE_WO_TREE just like in case when there's no full tree provided.
Theoretically you could introduce a new flag for our case which would behave the same - I just don't think it worth it because the problem is the same: we cannot check the finished download against the full tree because it's not available at the time of the checking (cannot be read from the hash database since it's not saved). In my thinking these are similar enough cases to use this flag to solve the problem of flooding.