- Added data logging
- Fixed problem where we were not always waiting for nested results
- Filtering on the _template property was failing
- Put change tracking reset in .finally()
- Don't using _template to drive class for subSetOf templates
- Tweaks for no database case
- Don't process schema when there is not one
- Eliminate duplicate indexing of schema
- (srksag pr) corrections to table sync
- (srksag pr) index synchronization happens at the table level
- (srksag pr) DDL changes are part of transaction
- Refreshing while passing in a map ignored processing if object being refreshed was in the map
- Added nojoin as a parameter at the same level as fetch: in cascade fetch specs
- Fixed serious bug where one-to-one relationships would get nulled out on refresh if the refresh made them point to a different object
- Fixed issue converting types on filtered consolidation
- Added consolidation of children that are in same table with different filter criteria
- Added serialization of reads to avoid overruning pool
- Added consolidation of duplicate queries
- moving to symver style of number for next round of changes
- Corrected synchrnonization of schema when properties added to extended templates
- Handling of setting null values
- Incorrect logging of errors on txn end
- minor logging changes
- minor logging changes
- Improved logging on end()
- Optimized index creation to only create indexes on one-to-many relationships
- Allow fetches specs in API specified as false to override.
- Syncrhonization broken on an empty database in previous release
- Don't fetch just because of an existing object relationship, fetches must be specified in spec
- Eliminate duplicate indexes
- Correction to schema management
- Handle deadlocks as update conflicts
- Merged pull request for table sync
- Defensive code against null elements in arrays
- Added call to MarkChangedArrayReferences when running with Semotus
- Changed default persistor to fetch: false
- Robust logging via objectTemplate.debugInfo = "io;api"
- Changed the id mapping works in a way that eliminates is dual function as an extra cache
- Fixed problem that broke Mongo
- Made persistor prop touching compatible with new semotus
- changed Number type to Double Precision
- Removed logging
- Fixed touch top object logic
- Improve boolean type conversion on Mongo
- Changed the updates to always fill in parent pointers for child relationships
- fully force refreshes
- More robust boolean handling in Mongo
- Support for enumurating persistor properties for amorphic
- Fixed situation where setting filter types could fail to write out
- Merge schema for subsets
- Preserve sort order which was broken
- Added limit and offset
- Nested dirty sets were not working
- Fixed some flaws in 2.50
- Added support for PersistObjectTemplate.objectMap = {} to prevent object id values from changing when refetched
- Fixed refresh
- self-referential entities not getting a persistor (still needs to be addressed in amorphic)
- changed the way recursive fetches protects against infinite recursion so fetch specs fully honored
- Remove from dirty list objects deleted in a transaction with object.persistDelete
- Implemented pruneOrphans to ensure orphans in a one-to-many relationship disappear
- Added support for filtered collections
- SetDirty defaults to not cascade
- You must set PersistObjectTemplate.enableOrphanHookups to set dangling parent pointers
- Optimize not reprocessing duplicately fetched entities *** Note that this feature means you have to replicate fetch specs for all paths to an object
- Correct touching of top object only for changed objects
- object.cascadeSave
- Handling of null values
- Automatically index foreign keys
- Fixed cascading saves on dirty
- Fixed subclass properties on query
- Transaction support in delete
- Superclasses now to fetch all columns
- Support for native joins
- Added alias to native knex support
- Added limited regex support
- Added native knex support
- Fixed $nin conversion
- Fixed change tracking
- Changed default table inheritence
- Numerous changes around dirty tracking and foreign key handling
- support for query objects using subsetOf in schema
- clean up parent pointers on cascade save for the benefit of data conversions
- Inherit collection and table for subclasses with a schema
- Use table rather than schema in knex to allow schema to be dual-purposed
- Updated boolean handling
- was attempting to join relations in local references
- More fixes to cascade save
- Fixes to cascade save
- fixes to fetch
- removed saved from object which was conflicting with application data
- fixes to deleteFromPersistWithQuery
- fixes to fetch
- Added an isKnex method to template
- Allow table to override forcing subclasses into parent tables
- Fixed cascade save option
- Changed from string to text for column types
- Allow column to be overridden in sub-class
- Fixed error message when missing schema entries during table sync
- Increased size of schema entry
- Cleaned up dependencies
- Added support for getPOJOFromQuery
- Further updates for knex support
- Alpha support for knex
- Add actual template name to the id of an object to allow for different classes reading an object
- Fixed problem introduced in 1.77
- Allow get/fetch by forcing queryOptions in model
- Update conflicts were resetting the version number meaning that the second time they would succeed
- Alias feature to allow duplicate templates for same collection
- Fixes to getFromPersistWithQuery for subdocuments
- Don't create persistors for schemaless templates
- When creating transient objects don't reset transient flag for fetchProperty
- When creating persistor properties preserve toClient/toServer settings from original property
- Fixed 0.1.64 fix
- Added logging for orphaned sub-documents
- Patch for sorting and also making cascades work properly for sub-doc queries
- Throw Update Conflict errors
- Saving other than top object would fail if parent pointer was not first property
- Optimization to not do duplicate processing
- Revamped sub-doc handling such that sub-doc fetches in same doc happen only with crossDocument: true on the children: and parent: schema properties
- Added transient parameter to queries for the benefit of semotus
- Problem referencing one-to-many of cross-document sub-documents to documents
- Allow query options and querys in .fetch call
- Case of cross-document references could lead to two separate objects of the same id
- Make sure duplicate peristors are not injected
- Fixed propFetch() to honor fetch parameters
- made cascade fetch pass through one-to-many relationships
- fixed more subdocument cases
- Fixed subdocument case that was broken in 0.1.46
- Added tests to ensure that init is not called on fetch
- Fixed a problem where sub-documents would not be found if multiple paths to them
- Fixed problem with closure introduce in 0.1.42 for sub-doc refs
- Treat sub-documents more transparently by allowing cross-collection links of sub-documents and direct query of them
- !! Changed the format of schemas (see persist_banking.js)
- Added error for saving an untemplated object
- Was not handling case where references in same collection were not sub-documents because they had a schema children reference
- Persistor object changes were not being synchronized resulting in extra callse to propFetch methods
- Made code not dependent on order of properties
- Make sure data retrieved with older version accounts for changes in the way template names are now