Implement ethereum specific fork detection and POI fallback

[stwiname]

Description

Use a different method for fork detection when chain doesn't have a concrete finalization mechanism. If POI is enabled then we can use that to rewind to the last operation before a fork occurred.

TODO

• Improve error message when POI not enabled. Need to include a way to recover from this
• Set fork method based on whether finalization is supported

Type of change

Please delete options that are not relevant.

• New feature (non-breaking change which adds functionality)

Checklist

• I have tested locally
• I have performed a self review of my changes
• Updated any relevant documentation
• Linked to any relevant issues
• I have added tests relevant to my changes
• Any dependent changes have been merged and published in downstream modules
• My code is up to date with the base branch
• I have updated relevant changelogs. We suggest using chan

[github-actions]

Coverage report

❌ An unexpected error occurred. For more details, check console

Error: The process '/usr/local/bin/yarn' failed with exit code 1

St.❔	Category	Percentage	Covered / Total
🔴	Statements	53.28%	3290/6175
🟡	Branches	66.52%	304/457
🔴	Functions	41.7%	98/235
🔴	Lines	53.28%	3290/6175

Test suite run failed

Failed tests: 1/53. Failed suites: 1/11.

  ● ProjectService › can get a block timestamps

    TypeError: Cannot read properties of null (reading 'unsafe')

      49 |     unfinalizedBlockService: UnfinalizedBlocksService,
      50 |   ) {
    > 51 |     super(
         |     ^
      52 |       dsProcessorService,
      53 |       apiService,
      54 |       poiService,

      at new BaseProjectService (node_modules/@subql/node-core/src/indexer/project.service.ts:52:25)
      at new ProjectService (packages/node/src/indexer/project.service.ts:51:5)
      at Object.<anonymous> (packages/node/src/indexer/project.service.test.ts:28:22)

Report generated by 🧪jest coverage report action from 8c6a6e2

[ianhe8x]

@stwiname I didn't read this PR closely, but this is what I think the best process

1. check parent hash when we add a new unfinalised block
2. if not match, stop fetching loop, create a mark
3. in post block hook, read the mark and start look up the fork
   1. if can not find fork in the block cache, either
   1. use poi and continue search fork
   2. add a mandatory number of blocks to the earliest block we have (like 1000) and rollback
4. on start up, if there are block cache in metadata, check the latest block in the cache before abandoning them if the gap is too large.
5. also in some case (like gap is small), we can keep the block cache without resetting after restart.

[stwiname]

Closing this in favour of #217