node and notices that it is over 32mb

This reverts commit b1c1e93d, reversing
changes made to 60be90af.

Instead just keep mempool around a little longer (like pre-refactoring).  This should
reduce large mallocs and remove a memcpy

Refs Tokutek/ft-index#46 Comments, minor renames, reworked bn_data::move_leafentries s.t. splits do not significantly increase memory usage of basement nodes and to be cleaner

now uses dmt iterate_on_range and the dmt builder instead of a for loop

Refs Tokutek/ft-index#46 Killed dmt_functor<> template.  Added dmtwriter_t template parameter to dmt and some documentation

Refs Tokutek/ft-index#46 Add some comments.  Improve dmt::verify() to check for more types of corruption.  More static checks

Refs Tokutek/ft-index#46 Remove ctree/dtree/mnode/cnode/base node/dnode/marks/etc... That is, removed all items that
were designed to allow us to upgrade the dmt later (e.g. fixed-length nodes for tree form) or merge with old omt (e.g. supporting marks)

Use dmt to replace omt in bn_data class for storing leafentries.
Optimization for serial inserts and mempool

osx doesn't support __thread well so scoped_malloc breaks it.  Rather than
fix it, we implement a correct but poorly-performing implementation for
now that will be used on osx machines.

fixes #160

Refs Tokutek/ft-index#159 Calculate klpair size correctly when message application deletes a leafentry

Refs Tokutek/ft-index#159 Calculate klpair size correctly when garbage collection deletes a leafentry

Garbage collection was done on the clones and did not always make it to in-memory.
Also, closing a db and reopening no longer (necessarily) flushes the nodes out of memory; we close the entire environment when we want to ensure we read it.

partial eviction

thread-local buffer when possible.

frwlock

refs Tokutek/mongo#886, change get_last_key to use a callback instead of copying data, so that we don't have
mongo code freeing data that fractal tree allocated.

Add db->get_last_key() and test.
Add test

partition

eviction callback

serialization. Assume the fifo is about 1kb larger than the read buffer
(in practice, the fifo is 5 bytes smaller than the read buffer).

create_no_array() so that deserialization does not malloc/free needlessly.