fix: make block query cutoff properly configurable#181
Conversation
This stack of pull requests is managed by Graphite. Learn more about stacking. |
Evalir
left a comment
There was a problem hiding this comment.
Actually, disregard the other review. This should be called BLOCK_QUERY_CUTOFF_BUFFER
prestwich
left a comment
There was a problem hiding this comment.
seconds is not enough resolution. needs to be milliseconds
aa0fb98 to
48769bd
Compare
d6fa407 to
376d078
Compare
cb227d0 to
376d078
Compare
88c0627 to
c7121a6
Compare
Evalir
left a comment
There was a problem hiding this comment.
I wonder if it's better to explicitly circuit break the building process altogether if the deadline ends up being too close to Instant::now() instead of this being implicit, because it indicates we've ran out of time anyway.
8ac4079 to
e782314
Compare
e782314 to
8199d03
Compare
8199d03 to
7575907
Compare
7575907 to
cd1c230
Compare
| desc = "Number of milliseconds before the end of the slot to stop querying for new transactions and start the block signing and submission process", | ||
| default = 3000 | ||
| )] | ||
| pub block_query_cutoff_buffer: u64, |
There was a problem hiding this comment.
there's a minimum reasonable value here right? Which is that quincey will stop signing. we should document that the minimum exists, and then document the current quincey configuration in the builder section of signet.sh/docs
4b2cd52 to
cb62e32
Compare
cb62e32 to
91df8a8
Compare
Merge activity
|

fix: make block query cutoff properly configurable
At some point in the past, the block query cutoff configuration value was removed and a hard-coded value was added in.
This PR adds the block query cutoff back into the config and uses the config value in the
calculate_deadline function.Builder example environment documentation needs to be updated to reflect this change.