Behaviour

By default, the widget execute the hardcoded query but one can edit the query to run it's own.

• If the query is empty the default query (everything that is open) is executed
• If the textarea is not empty, its content is interpreted as the query

The columns are still the same as well as the order of the result.

The query is written in TQL, using code mirror with autocompletion and the list of accepted field is available on the right hand side (like expert mode in trackers).

For this story, only "title" and "description" semantics are usable. A query like: title = "foo" will return all artifacts who's title semantic match "foo".

Unlike TQL in the trackers, at this point, TQL search on semantics, not on field names. To avoid any confusion, the initial Runtime Check (see art #10677) will report an error if:

• There is a field called "title" in one of the tracker and this field is not attached to "title" semantic
• At least one tracker doesn't have a "title" semantic defined

Same applies for "description" accordingly.

The TQL search must met the "expression limit" defined by site administrator to avoid too complex queries

Technical

• The grammar defined for TQL can be kept
• Need to implement a new Tree Verification
• Need to implement a new Query Generation
• Warning: how to properly manage errors between Restler and JS