Garbage collection fixes (again)

[mbutrovich]

This is a subset of the changes of #1349 and a replacement for #1326. It contains only the garbage collection fixes, ignoring the TileGroup freeing and TileGroup compaction features. This PR resolves almost all of the issues identified in issue #1325.

It includes the results of a thorough correctness audit we performed on Peleton's garbage collection system. It includes a whole new test suite for the Transaction-Level Garbage Collector and several important bug fixes to the Garbage Collector and Transaction Manager. The bug fixes themselves are a very small number of the line changes, with the majority being added tests, documentation, and refactoring.

GC Fixes Summary:

• Added 15 tests to transaction_level_gc_manager_test.cpp to handle more complex GC scenarios.
• Currently 5 of these tests still fail for polluting indexes with old keys, but we believe this will require more significant changes at the execution layer to resolve. Will open a new issue describing in detail the remaining problems after this is merged.
• We believe we have resolved all of the tuple-level GC bugs and most of the index bugs.
• GCManager::RecycleTupleSlot allows unused ItemPointers to be returned without going through the entire Unlink and Reclaim process.
• Modified TOTransactionManager to pass tombstones created by deletes to the GCManager. *Modified DataTable's Insert to return the ItemPointer to the GCManager in the case of a failed insert.
• Modified DataTable's InsertIntoIndexes to iterate through indexes and remove inserted keys in the event of a failure.
• Modified GCManager's Unlink function to clean indexes from garbage created by COMMIT_DELETE, COMMIT_UPDATE, and ABORT_UPDATE.

A significant portion of this code was co-written by @dqrs but it was easier for me pull the bug fixes from our TG compaction branch than to remove the undesired changes.

[coveralls]

Coverage increased (+0.3%) to 76.849% when pulling f0130ef89db336712168ca8f72a57a64c7c835fb on mbutrovich:gc_fixes_2 into 196f663d8f04d0a92b878d8efcf799b280c65376 on cmu-db:master.

[mbutrovich]

We passed 78% coverage!

[tli2]

@mbutrovich My theory is that we can increase coverage simply by deleting unused code and simplifying what we have without having to write new test cases for probably a month or so at this point.😅