SIMD#83 - Allow conflicting entries on replay side

[Huzaifa696]

Problem

Currently self-conflicting batches are not allowed on the replay side whereas SIMD#83 requires this support.

Summary of Changes

• Introduced a flag in every batch to distinguish between self-conflicting and disjoint batches
• Bypass rebatching for self-conflicting batches (for now)

Related to:
#1025

[Huzaifa696]

This PR is currently a WIP.
It's complete in terms of code changes to support self-conflicting batches in rebatch_and_execute_batches (old flow). Our initial testing shows that schedule_batches_for_execution (new flow) is supporting self-conflicting batches.

We are currently working on testcases and benchmarking. Pushed a little in advance to get your feedback.

[Huzaifa696]

Incorporated unit tests for rebatching logic in rebatch_and_execute_batches to ensure self-conflicting batches are not rebatched. Also added a unit test for schedule_batches_for_execution (new flow) for self-conflicting batches and it seems to handle this case.

After SVM changes (to allow self-conflicting batches), further tests can be added to test execution of self-conflicting batches in process_batches.

Is rebatching of self-conflicting batches required?

[steviez]

This PR is more than a year old and I believe the implementation has been completed elsewhere (SIMD 83 looks to be coming up on the testnet feature activation queue) so going to close this PR