--Omit ntoreturn and ntoskip from the getMore command log line.
--Add nreturned to the getMore command log line.
--Upconvert OP_GET_MORE profiler entries to match getMore cmd.
Includes the following changes to profiler entries for queries:
-- The query field will contain find command parameters, e.g.
{find: <coll>, filter: {...}}.
-- Renames 'nscanned' to 'keysExamined' and 'nscannedObjects' to 'docsExamined'.
-- Renames 'scanAndOrder' to 'hasSortStage'.
-- Removes the top-level ntoreturn and ntoskip fields. Skip, limit, and
batchSize info will be present in the 'query' field instead.
The goal of the refactoring is to clarify the control flow of the batch
execution of inserts, and in so doing to cleanly handle PageFaultExceptions,
properly handle demotion to secondary, correctly profile individual document
insertions within the batch, and correctly attribute lock hold and acquisition
time to those individual insertions.
It is also intended that this refactoring will resolve memory leaks in the
insert code path, by more carefully tracking ownership of WriteErrorDetail
objects throughout the WriteBatchExecutor class.
