You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
There are currently no solid test for testing the speed of a node sync. Since the block building is currently very slow, I think the most sane approach is that we build a fixture of 50 blocks that are stored on disk.
We can then run tests in the future, where we manipulate the time of anvil such that they are valid, and afterwards start a node that is using the setup and simply measure the time it takes it to sync.
An extension of this test should also allow us to test a scenario where the pending chain is pruned, without needing the full state to be handled correct, and merely see if the archiver behaves as intended.
The text was updated successfully, but these errors were encountered:
Fixes#8536 by introducing an initial setup for testing synching.
Adds a new "spam" contract that uses a large amount of data, such that
we can see the differences between "noop" transactions and more complex
things.
Adds a new `l1StartTime` option to the setup such that we can do an
instant "warp" to make the execution of past stored blocks much simpler.
While the intention was to have a lot of blocks to see synching over a
longer period, building that with non-trivial transactions will take
until I retire.
![Aging Matt Damon
GIF](https://media4.giphy.com/media/GrUhLU9q3nyRG/giphy.gif)
The test is **NOT** set to run in CI explicitly as it is a "measure"
test more than anything else, and without there being a target it is a
bit weird to restrict it 🤷
There are currently no solid test for testing the speed of a node sync. Since the block building is currently very slow, I think the most sane approach is that we build a fixture of 50 blocks that are stored on disk.
We can then run tests in the future, where we manipulate the time of anvil such that they are valid, and afterwards start a node that is using the setup and simply measure the time it takes it to sync.
An extension of this test should also allow us to test a scenario where the pending chain is pruned, without needing the full state to be handled correct, and merely see if the archiver behaves as intended.
The text was updated successfully, but these errors were encountered: