GeoGit Files

Geogit 0.6.0 CHANGELOG
======================

NOTE: due to the change in the default storage order for tree nodes (i.e. hash function used),
repositories created with versions prior to 0.6.0 are no longer compatible. This compatibility 
breakage, though a bit unfortunate, is trully needed for trees to be built much faster.

Features
--------
* Pass credentials from a web client through GeoServer to GeoGit.
  When using geogit's DataStore with GeoServer, the full user name and email address can be passed
  down as WFS Transaction properties down to the datastore for it to create sensible geogit commits
  with all the needed information besides commit message.

* Add Performance Measurement Framework for benchmarking and logging of heap/non heap memory.
  Run geogit config metrics.enabled true when gogin to assess the time taken to run geogit commands,
  and check for the results in .geogit/log

* Added --fid-attrib <attribute name> option to the shp import command to select an attribute to be
  used as the feature identifier instead of the default shape recno.

* Allow exporting OSM data to PostGIS using a mapping with multiple rules

* Horizontal scale support: MongoDB Implementation for the staging, object, and graph databases.
  Next step is to make it easier to configure the backend of choice.

* Import from and export to Oracle Spatial

* Faster GraphDatabase imlementation based on the BluePrints API (https://github.com/tinkerpop/blueprints/wiki).
  The pre-0.6 Neo4J implementation still lives as a separate module.

* Added a rebuild-graph operation that fixes errors in the graph database.

Improvements
------------

* Performance improvement for building large feaure trees
  When importing a large dataset, the tree nodes are first written to temporary files and then added
  to the tree in the default storage order, making for the tree building process much faster as it has
  to split itself less times into smaller trees.
 
  Also use a FNV-1a <http://en.wikipedia.org/wiki/Fowler%E2%80%93Noll%E2%80%93Vo_hash_function> hash
  function instead of SHA-1 to compute the storage order and bucket by depth of tree nodes.

* Performance and database size improvements
  Performance improvements to the import command, it uses several read threads (up to number of cores /2)
  to read different subsets of the data source (for JDBC based ones) and perform the feature import in parallel.
  The default staging and permanent object stores (based on Berkeley DB Java Edition) now use a single
  writer thread for bulk inserts, reducing the load on the BDB background threads that perform
  checkpointing and log cleaning, and resulting in smaller on-disk databases.

* Improve performance of OSM imports (from pbf)
  Time taken to import from osm files (pbf or xml) reduced dramatically by using a producer-consumer 
  approach to parse and load the features out of the export files, allowing for one single insert to
  the working tree for the whole file, and offloading the node (point) cache to a temporary database
  instead of a smaller cache on memory and tree lookups on cache misses.
  
* "geogit clone" clones into a new directory named after the cloned repository instead of under
  the current directory

* Improved error message when connection to osm server fails.

* Several documentation improvements, including a new theme for the home page and a GeoGit workshop
  with a realistic scenario (http://geogit.org/workshop/)