Verkle checkpoint db tests

[scorbajio]

This change adds some unit testing to the verkle package and fixes an issue where db stats aren't being updated in the commit function when write operations are being written to the db.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 85.29%. Comparing base (d24ca11) to head (1003bff).
Report is 4 commits behind head on master.

❗ Current head 1003bff differs from pull request most recent head eed4486. Consider uploading reports for the commit eed4486 to get more accurate results

Additional details and impacted files

Flag	Coverage Δ
block	80.79% <ø> (?)
blockchain	?
client	84.23% <ø> (?)
common	?
evm	73.76% <ø> (?)
genesis	99.98% <ø> (?)
statemanager	?
util	?

Flags with carried forward coverage won't be shown. Click here to find out more.

[gabrocheleau]

In the trie package checkpoint, we only add writes for del operations and put operations. Why do we add it here to commit?

[scorbajio]

See bellow for full response: #3407 (comment)

[gabrocheleau]

Nice! Really like the additional tests, and it's great to see you guys jumping in and getting your hands dirty with the verkle work :)

Left one small question concerning the db.writes stats, as there seems to be an inconsistency in the way that we are doing this here vs. in the trie package.

The actual behavior seems to be consistent, but at the trie level, we are incrementing the db.writes count whenever there is a put operations and hasCheckpoints is false. See https://github.com/ethereumjs/ethereumjs-monorepo/blob/master/packages/trie/src/db/checkpoint.ts#L185-L200

I have limited familiarity with that section (i.e. the stats handling) of the code, but shouldn't this be done as low-level as possible (at the level where the actual db writes are happening), to avoid the need to manually increment in multiple places? Curious to hear your thoughts.

Can be merged once that's discussed/resolved.

[scorbajio]

Nice! Really like the additional tests, and it's great to see you guys jumping in and getting your hands dirty with the verkle work :)

Left one small question concerning the db.writes stats, as there seems to be an inconsistency in the way that we are doing this here vs. in the trie package.

The actual behavior seems to be consistent, but at the trie level, we are incrementing the db.writes count whenever there is a put operations and hasCheckpoints is false. See https://github.com/ethereumjs/ethereumjs-monorepo/blob/master/packages/trie/src/db/checkpoint.ts#L185-L200

I have limited familiarity with that section (i.e. the stats handling) of the code, but shouldn't this be done as low-level as possible (at the level where the actual db writes are happening), to avoid the need to manually increment in multiple places? Curious to hear your thoughts.

Can be merged once that's discussed/resolved.

I did some more digging into the discrepancy and added some more tests for the CheckpointDB implementation in the trie package as well: It seems like the issue is coming from the batch function, since if checkpoints already exist, each batch operation is performed using the CheckpointDB's put/del functions, which already increment the write stat, but if checkpoints don't exist, the wrapped db's own batch function implementation is used, which doesn't have access to and does not increment the wrapping class' stats property. I've changed both to now increment that stats in the batch function if checkpoints don't exist.

[gabrocheleau]

thanks for addressing the review! lgtm