New stable build (#615)

The new stable build (#616) introduces many fixes and some new features. See the complete change log below.

The new build is now available for download both from our build server and from nuget. See our download page for more details.

New features:

• Incremental backups
• Boosting is now supported for querying and indexing
• Will automatically handle casts for indexes
• Better handling of errors from RavenDB, will not hide them in json text any more
• Added the ability to randomly sort values
• Easier way to handle dynamic fields, with the new CreateField method.
• Added ability to defer database commands to be called on SaveChanges()
• Added a way to get the current user from a trigger

BREAKING CHANGES:

• Moving the methods that expose indexing behavior to a separate namespace. That way, they don't pollute the query namespace.
• Moving to a more strict behavior for existing backups. Backup will FAIL for full backups if destination directories exist.
• Removing support for contains from linq API, will generate a big error with explanation why.

Studio:

• Can delete all docs in a collection
• Fix a bug when typing a not valid query an error dialog was appear.
• Show a busy indicator when documents are loading
• Can search for documents in the
• Supporting api key for the studio
• now ctrl+click on any navigation will open in new tab
• Add mapped results button to index edit
• Fix a NRE. In case of a projection Metadata can be null, and ToJson assumes Metadata is always not null.
• Fix spatial query not sending parameters
• Add the current URI to the error message
• Store the URL of the request that is faulted.
• Convert SecurityException to ServerNotFound exception.
• Logs are now sorted by date descending
• Ctrl+Click on the Studio will open in a new tab.
• Better errors in the studio.

Client API:

• Implement DeleteByIndexAsync
• Optimizations - don't create copy of the strings we send to the server, send them directly
• Added IDatabaseCommands.Head() method to retrieve a document's metadata (JsonDocumentMetadata).
• Fix a bug where nested aggressively caching operation wasn't allowed
• Adding better error handling for when we can't parse json from the server
• Properly translate a custom identity property in a base type
• Filter out OR, AND etc that might create issues parsing during Search
• Will automatically handle _Range postfix if the values are numeric and there isn't a _range postfix for range queries
• fixed anonymous arrays in expression string builder
• Moved to Reading Json directly, rather than the string, better error handling that way
• Avoid calling TryEnlist if we don't have to, will avoid full trust issues
• Better support for handling equality of guids to strings
• Fixing an issue where multi load would optimize itself on the wire and will give different results when running remote vs. local

Indexing & Searching:

• Adding a way to access the metadata in strongly typed indexes. MetadataFor().
• Fixing #410 - reserved words in indexes
• Fixed aliases in projection using Linq chained method syntax
• Supporting boosting when using Linq Search API
• Support boosting on search phrases
• Better error handling when the facet document is not found
• Adding support for boosting a value during indexing
• Allow to set boost level over the entire document
• We can now specify boosting on a per field basis
• Fixed an issue where some Enumerable extension methods couldn't be bound
• Fixing an issue where we doubly handled reverse
• Can generate complex indexes with the appropriate lambda casting
• Adding a method that allows us to access the entity as a document. AsDocument().
• Fixing an issue with OrderBy followed by a Search
• Add a helper method to dynamically store fields in an index, without explicitly using the Lucene.Net.Document.Field class.
• Fixing an issue where we have an issue with multiplication of decimals would generate an error
• Fixing an issue where we weren't comparing to null properly
• Fixing an issue where very complex dynamic query wouldn't hash properly

Mvc Integration:

• Fixing an issue where we would generate a recursive check for ajax requests

Server:

• Can get the database size using: GET /database/size
• Reduce the size that we use for transactional logging
• Reduced the size that the db would grow by default, smaller db dir sizes
• Allow customizing the index storage path and logs path for each tenant
• Better way to handle disposal of database resources when running in IIS.
• Safe handling of errors during disposal
• Handle the scenario where an idle database cannot be properly disposed in a nicer fashion
• Fix typo: Non-Authoritive-Information to Non-Authoritative-Information
• Support both "~\" and "~/" for configuration of relative paths.
• Better handling of concurrency problems when shutting down the server
• Added a way to configure esent circular log
• Fixing things so we can register the nlog config properly for http endpoint in IIS
• Support copying indexes on incremental backups
• Don't send an etag for the case where the Studio.xap cannot be found (force the browser to always recheck)
• Ensure that we can't get into duplicate index errors because of concurrency issues
• Allow parsing of more free form [[ ]] elements
• Fixing an issue where we couldn't read an inflight transaction document using Munin
• Allow to define tokens that have read / write permissions per db, not globally
• Better handling of disposal concerns when running inside IIS
• Better logging for start/end on IIS
• Add support for env variables in paths
• Better output for when we are running in memory
• Fixing incremental backups support
• Can now support administrator in OAuth
• Auth bundle now support admin users

Licensing:

• Allow to specify license string using Raven/License
• Allow to specify license path using Raven/LicensePath
• Won't do multiple license checks when running using OEM
• Don't run licensing for each database on the server, run it only once
• Avoid publishing presence again if we get notification within the publish limit

Replication:

• Will not replication recovery information attachments
• Added a way to explicitly avoid replication for certain documents / documents
• Adding option to control transitive replication (default is OFF)
• Fixing an issue where deleting a document with the replication bundle active wouldn't physically remove it from the index

Smuggler:

• Adding smuggler option to skip pushing indexes