This partially reverts fcf7f7d6. The original intention was to have
no library code reading config. Moving code snippets to another lib
doesn't solve this. We need quick & easy access to a monitoring
backend, on CLI, in other modules and so forth. Duplicating so much
code again and again is not an option.
I prefer the former way of doing things unless we find a better
solution.
Library code should not read configuration files. The new createBackend() method
from Monitoring/Controller takes care of reading backend configuration and error
handling. The Backend constructor now requires a resource instance.
This interface is intended to be used within the data views and monitoring controllers. The monitoring list
and show controllers should only query data views instead of the actual query. Thus a data view is required to
implement the QueryInterface, which will be changed later.
The name is not final because it somehow conflicts with the Queryable interface which is also marked as work in progress.
Since the BaseQuery no longer allows passing the columns to select via
its the constructor, the columns are resolved when set by the
columns() function.
The BaseQuery forwards fetch*() function calls to its datasource with the
query as first parameter since the datasource implements the actual fetching functionality.
There's no interface for fetch*() functions on the datasource level yet.
The previous implementation always created both the count
and the select query.
For readability the property baseQuery has been renamed to select which the
IDO queries must take into account.
This commit breaks the application as it introduces incompatible changes which
will be adopted bit by bit.
Since the filter functionality is subject to change it's removed from the BaseQuery.
The functions setOrderColumns() and getOrderColumns() are deprecated because they're
replaced by functions in the Sortable interface.
Further the Sortable interface now defines the sort constants ASC and DESC and are
thus removed from the BaseQuery. In addition the sort constants are no longer integers
but strings.
The distinct() and isDistinct() functions are removed because they're database specific.
They have been introduced with changes in the PivotTable implementation which should
actually not call distinct blindly anyway.
Eric Lippmann
Thomas Gelf
Johannes Meyer
Michael Friedrich
Alexander Klimov