chore: filter v2 tests push invalid payload #916
Conversation
- TestEmptyPayload - TestEmptyContentTopic - TestEmptyContentTopicEmptyPayload - TestTimestampInFuture - TestZeroTimestamp - TestNegativeTimestamp
Jenkins BuildsClick to see older builds (55)
|
|
I wonder about timestamp field in Waku messages. There is no protection against insider attack. When someone wants to send messages in the past and alter order of messages, we let him do it. Is it difficult to verify common time flow (to have consensus on time) ? What are the trade offs here ? I imagine what would happen, if postal offices and government offices unite to collect fine fees with penalty stating something happened in the past. And I imagine they don't do it, because such attack is complex and slow. Not with Waku, though. Order can be altered quickly and wrong decisions or claims could be made at the same speed. Interested in your thoughts. |
romanzac
requested review from
richard-ramos and
chaitanyaprem
as code owners
This should get sorted out once RLN gets applied to light protocols as well. Since RLN is timestamp dependent that should take care of issues with timestamps (if someone tries to push older message, they get penalized). You can refer to discussion here waku-org/research#45 |
There was a problem hiding this comment.
All the tests documented, none of them actually test anything wrt filter Push, rather only relay publish.
I would expect these to be part of testing relay API than filter.
What is the objective of the tests though if at all we are thinking of for filter protocol?
Changed according to your suggestions. Please have a look. Thanks. |
Description
Batch of tests for Waku Filter v2 - PUSH with invalid payloads
Changes
Tests available to run:
cd go-waku/waku/v2/protocol/filter
go test