undo BlockRecord cache insert, when DB fails

[arvidn]

We currently have two caches on top of the block store database.

1. The BlockStore itself keeps an LRU of recently requested FullBlock objects.
2. The Blockchain class keeps recent BlockRecords in a cache (recent, as in close to the peak)

When we fail to insert a block in the DB, we undo the insertion into the block cache in BlockStore, but we currently don't undo the insertion of the block record into the block-record cache in Blockchain.

This PR fixes that.

Additionally, this PR rolls back any additions and removals from ForkInfo, if they have been added while validating the block.

[wjblanke]

aok

[wjblanke]

earle had issues so holding off for now

Dismissing because he doesn't actually want this merged yet

[coveralls-official]

Pull Request Test Coverage Report for Build 7039010402

Warning: This coverage report may be inaccurate.

We've detected an issue with your CI configuration that might affect the accuracy of this pull request's coverage report.
To ensure accuracy in future PRs, please see these guidelines.
A quick fix for this PR: rebase it; your next report should be accurate.

• 13 of 13 (100.0%) changed or added relevant lines in 2 files are covered.
• 1 unchanged line in 1 file lost coverage.
• Overall coverage increased (+0.1%) to 90.361%

---

Files with Coverage Reduction	New Missed Lines	%
tests/util/setup_nodes.py	1	98.31%

Totals
Change from base Build 7038638915:	0.1%
Covered Lines:	93668
Relevant Lines:	103612

---

💛 - Coveralls

[wjblanke]

aok