reverting back to same snapshotId causes ganache-core to freeze

[ewingrj]

Expected Behavior

I can call evm_revert with the same snaphostID multiple times

Current Behavior

It works correctly the first time, but will cause testrpc to freeze up if reverting to the same id a 2nd time

Possible Solution

Steps to Reproduce (for bugs)

1. create a snapshot
2. rollback to the snapshot you created
3. optionally modify state
4. rollback to the snapshot you created again
5. ganache never returns a response and will not respond to any additional requests

Context

Your Environment

• Version used: 7.0.0-beta.0
• Environment name and version (e.g. PHP 5.4 on nginx 1.9.1):
• Server type and version:
• Operating System and version: osx
• Link to your project:

[mikeseese]

Thanks for reporting this! We're going to try to verify this still happens, but I noticed that you're running 7.0.0-beta.0. This build was deprecated when we decided to not continue on the 7.x series yet. Please upgrade to the latest version (ganache-cli@6.1.8).

Thanks!

[konstantinzolotarev]

Yep got same bug.
And also more details on this:

When you revert any snapshot using eth_revert after this you could kill ganache using another evm_revert with any parameter. Here is "advanced logs"

Logs in gist

This logs are from fresh ganache installation.

Tried on:

• ganache-cli@6.1.8
• ganache-cli@6.2.1
• Ganache for Mac v1.2.2

And also in 6.2.1 response format for evm_snapshot changed from 0x01 to 0x1

[davidmurdoch]

@konstantinzolotarev thanks for providing the additional info. As far as the evm_snapshot response change goes, we now (correctly) return an RPC Quantity hex encoded string instead of an RPC Unformatted Data hex encoded string.

[AlexanderPoschenrieder]

Still happening v6.2.5.

[mortimr]

@AlexanderPoschenrieder, @tcoulter mentionned that this behavior is intended here
As snapshots are getting "consummed" after a revert, I just recreate a snapshot after each of my reverts and it's working fine (using this for contract testing)

[balexander4]

Documentation has been updated on this to clarify what is occurring. Just awaiting a the PR approval before closing this issue.

[honestbonsai]

@ewingrj Closing as we believe this has been fixed. If that's not the case, feel free to reopen.