Roadmap

[imdrasil]

This issue includes all thoughts and information about features that must/should/could be added to Jennifer in a one of the following releases.

This issue is pinned to the top and will be updated. If anyone would like to propose a new feature - create a separate issue with detailed description. If you would like to pick up existing one from this list - also create a separate issue to indicate that fact.

All items that have any progress marked with an associated release which means they are ready but haven't released yet.

Roadmap

Must

• add ability to give multiple field names to a single validates_* validation macro in model class declaration
• add multiple adapter support [0.9.0]
• add through option to has_many and has_one associations
• ensure all migration table builders has all required actions and similar interface (e.g. ChangeTable has not all actions) [0.9.0]
• remove Adapter .exec, .query and .scalar methods [0.9.0]
• add description how to integrate micrate with Jennifer [0.9.0]
• remove Model::Base.primary_field_type [0.9.0]

Should

• add reversible mechanism to Migration::Base in addition to existing up/down
• add complete timestampz postgres data type support
• add native support of CASE-WHEN construction
• consolidate expression model - SQLNode, Statement, Criteria looks a little bit odd ATM and requires consolidation to bring some logic structure to class hierarchy
• add opportunity to define composite primary key
• add opportunity to specify a composite foreign key in association definition
• add :optional key to association macros and make it false by default for belongs_to
• add postgres generated columns support
• remove Resource.build in favor of .new
• add support of models/views without defined primary key
• separate adapters to the separate repos
• add restrict_with_error dependency strategy to association
• add check for cyclic destroy dependency
• add partial index support to postgres adapter
• add USING support to postrges adapter
• add column encoding and collate support

Could

• add opportunity to specify whether migration is executed in pessimistic/optimistic strategy
• consider converting time to UTC on bulk insert for raw insert (array of hashes)
• differentiate data and schema migrations
• consider more flexible way to support roadmap
• add pseudo materialized views for mysql adapter - using table
• add opportunity to switch query to escape arguments instead of bind
• consider sharing connection across fibers instead of connection-per-fiber strategy and add read/write lock mechanism
• make type field of STI models customizeable