Inline editors are mostly adapted to new data access

Save empty page ids as strings, not NULL

pid is part of primary key and so must not be NULL. Otherwise the unique constraint does not work, SQLite effectively treats every NULL as unique.

Rudimentary handling of serial data

Serial data is bound to a page but does not affect versioning

Start unifying schema types

Schemas per se are type agnostic, isLookup property is removed. Data is stored and accessed differently based on how it is entered and retrieved.

The crucial change is i

Start unifying schema types

Schemas per se are type agnostic, isLookup property is removed. Data is stored and accessed differently based on how it is entered and retrieved.

The crucial change is introduction of the composite key of pid and rid. Previous page data utilizes rid = 0 to differentiate itself. Other types, notably lookup, have autoincrementing rid.

Database migration is not implemented yet.

show more ...

Use labels as keys for getData array of Values

This makes accessing the value of a specific column easier, since
otherwise we would have to loop through the array and call getLabel for
every column.

Use labels as keys for getData array of Values

This makes accessing the value of a specific column easier, since
otherwise we would have to loop through the array and call getLabel for
every column.

Tests have been adjusted.

show more ...

show display value in diffs. fixes #236

close result set

better handling of failing sqlite plugin init. fixes #200

show new row when adding to lookup

This renders the newly created row and adds it dynamically to the
existing table.

Merge branch 'lookups' into validationrefactor

* lookups:
test translation feature
implement rendering via sub type in Dropdown
fix displayValue for looked up type, more tests
fixed construc

Merge branch 'lookups' into validationrefactor

* lookups:
test translation feature
implement rendering via sub type in Dropdown
fix displayValue for looked up type, more tests
fixed constructor signature
first test for new Dropdown features
adjusted tests for renamed aliases
Add Lookup Support to Dropdown
column aliases need to differ from real column names
use display value as default renderer
introduces getDisplayValue

show more ...

more Validator refactoring

first start at refactoring the validation mechanism

A Validator is now returned by the AccessTable. It has it's own save
method. The entry.php was split up in entry, edit and revert to separate
the

first start at refactoring the validation mechanism

A Validator is now returned by the AccessTable. It has it's own save
method. The entry.php was split up in entry, edit and revert to separate
the different concerns. Validation is only done in entry and the thanks
to the new Validator mechanism, the once loaded access table is reused
in saving.

This will break all kind of stuff. Fixing that is for the upcoming
commits

show more ...

column aliases need to differ from real column names

Otherwise the grouping is ambigious. In Search() this was the case
already.

begin stremalining raw handling #148

This starts implementing the wanted features in the Value and
AccessTable class. Calling classes and tests aren't adjusted, yet

removed unecessary function calls

some cleanup regarding timestamp setting in AccessTable*

inline editing for lookups works now

Additionally only the edited field not the wohle row is validated now.

renamed SchemaData and SchemaLookupData

This is to reflect that they no longer inherit from Schema

refactor SchemaData into AccessTable with factory

first part of the refactoring for the SchemaData and SchemaLookupData
classes. Both now inherit from AccessTable which provides common methods
and t

refactor SchemaData into AccessTable with factory

first part of the refactoring for the SchemaData and SchemaLookupData
classes. Both now inherit from AccessTable which provides common methods
and two factory methods to automatically create the correct class
depending on the schema.

Both classes no longer inherit from Schema. Instead they contain a
Schema as a member variable. This makes it possible to initialize them
with an existing schema (which is required for the factory methods).

The flow is now like this:

1: init a schema (read info from the database)
2: look at the schema to figure out if it's a lookup or not
3: initialize the right Data object with the schema

show more ...