Refactorings for schema files and Doctrine_Record builder. More to come but got held up on some things that zYne- must fix first.

--no commit message

--no commit message

The unset functionality is not working for foreignKey relations

Doctrine_Query_Tokenizer::bracketExplode() were case sensitive, and 
changed tokenizer to make them case insensitive.

Introduced a new attribute ATTR_AUTOLOAD_TABLE_CLASSES that can be used to control whether Doctrine should try to autoload custom table class files when looking for them.

The default is set to false because fetching the relations deletes
previously fetched relations =P. This is, if you fetched an aggregated
value (SUM, COUNT), it wont be refreshed, it will be overwritten by the
actual related values.

Consolidated a few things. Initial entry of Inflector class. Moved some methods from Doctrine base class to Doctrine_Lib and Doctrine_Inflector.

--no commit message

More refactorings. Commented out the plugin tests for now. They seem to wreak havoc with a lot of other tests (especially validator tests) and i havent found the reason yet.

--no commit message

Makes it easier to see the dependencies between these two concerns, cache the results of the query parsing process later on, implement & use alternative hydrators to experiment with different algorithms etc. etc.
Other small refactorings.

Refactoring. Started to introduced a clear separation between column names and field names (aka column aliases).\nDoctrine_Connection, DQL Parser classes/methods map field names => column names.\nDoctrine_Hydrate maps column names => field names during hydration.\nColumn names are only stored in Doctrine_Table:: and Doctrine_Table::.\nRelations use column names in 'local'/'foreign'.\nWhen using field names (column aliases) you need to use the column names in 'local'/'foreign' when setting up a relation (hasOne/hasMany), not the field names.\n\n In other words column names are only used to communicate with the database. field names are used everywhere else. the casing of field names does not matter. column names are forced to lower case for portability. If you dont use field names (column aliases) your column names are your field names (and therefore all lowercase).