[Bug-Candidate]: Echidna can get stuck while shrinking

[rappie]

Describe the issue:

In certain cases, Echidna can get stuck while shrinking.

Example below:

When setting maxBlockDelay and maxTimeDelay to 1, it is easily shrunk further:

After further research I was able to determine that timestamps/blocks actually don't play a role in the breaking of this invariant at all. So it should have been able to shrink it further.

Code example to reproduce the issue:

The code is private, I can share the repo if you DM me on Slack

Version:

Echidna 2.2.3 (custom build from master)

Relevant log output:

No response

[gustavo-grieco]

I think we need to be more aggressive reducing the block numbers? In this case, this required at least a block each operation?

[rappie]

No, there are no block/timestamp increments necessary at all to break this invariant.

Besides block/timestamp shrinking being slow or getting stuck, there are also uneccesary transactions that don't get shrunk away. It seems to me that there is some deeper issue going on.

[gustavo-grieco]

@rappie can you please re-test the shrinking using the latest master?

[rappie]

Still the same issue :disappointed: