bugfix: TSMT failed to verify empty word for depth 64. #212
Conversation
When a prefix is pushed to the depth 64, the entry list includes only the values different than ZERO. This is required, since each block represents a 2^192 values. The bug was in the proof membership code, that failed to handle the case of a key that was not in the list, because the depth is 64 and the value was not set.
|
Kudos, SonarCloud Quality Gate passed!
|
| // length | ||
| if self.is_value_empty() | ||
| && get_common_prefix_tier_depth(key, &self.entries[0].0) < self.path.depth() | ||
| { | ||
| return false; | ||
| } |
There was a problem hiding this comment.
I don't really understand why we have an entry [(key, EMPTY_WORD)] for empty leaves as an exception (meaning, even though the value is empty, there is an entry in the list). But I didn't want to change it, since that would be a large change.
There was a problem hiding this comment.
We need this entry to figure out which node the path is for (i.e., we use the key from the tuple to figure out the index. There may be a way to work around this (either by introducing a separate field for the index, or by some other means), but I haven't thought about the best way to do it.
| // length | ||
| if self.is_value_empty() | ||
| && get_common_prefix_tier_depth(key, &self.entries[0].0) < self.path.depth() | ||
| { | ||
| return false; | ||
| } |
There was a problem hiding this comment.
We need this entry to figure out which node the path is for (i.e., we use the key from the tuple to figure out the index. There may be a way to work around this (either by introducing a separate field for the index, or by some other means), but I haven't thought about the best way to do it.
When a prefix is pushed to the depth 64, the entry list includes only the values different than ZERO. This is required, since each block represents a 2^192 values.
The bug was in the proof membership code, that failed to handle the case of a key that was not in the list, because the depth is 64 and the value was not set.