-
-
Notifications
You must be signed in to change notification settings - Fork 672
Assertions needed to catch zero NIDs #955
Comments
Whilst running the sytests for #954 against SQLite, I am noticing that the monolith logs are showing a lot of failed It appears to be because @kegsay In SQLite these are wrapped in a transaction in |
Actually it looks like this is also happening sometimes in Postgres mode, just in a different place, so that's fun.
|
So |
All NIDs (event, event type, state block, state snapshot, event state key, room, etc) always start from
1
, but there are sometimes cases where bugs cause us to request0
NIDs. We should put some assertion to catch these bugs as they happen.The text was updated successfully, but these errors were encountered: