This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
|Product:||platform||Reporter:||Jaroslav Tulach <jtulach>|
|Component:||Explorer||Assignee:||_ tboudreau <tboudreau>|
|Issue Type:||DEFECT||Exception Reporter:|
|Bug Depends on:|
Description Jaroslav Tulach 2003-07-09 16:34:22 UTC
Following constract seems pretty fragile and hard to keep when evolving the implementation. Please make it more robust by writing a verification check: The PropertyEditor instance will already be initialized with the initial value, and if it is an instance of ExPropertyEditor, ExPropertyEditor.attachEnv(env) will already have been called. The PropertyEnv instance is passed to allow rendering hints to be passed to the InplaceEditor instance.
Comment 1 _ tboudreau 2003-07-11 07:46:22 UTC
Okay, not a problem. Re hard to keep, well it's mainly scoping for the inplace editor - that the inplace editor will be responsible for presentation of a property editor, and firing changes that may be applied to it, but the infrastructure that instantiates it is responsible for any calls that set the value of the property editor, and that the inplace editor does not decide when or how the value is actually set - it can only communicate that it is ready for this to be done. Value setting code for property editors should be centralized so there is a single contract that can be evolved without forcing changes in anybody's inplace editors.
Comment 2 _ tboudreau 2003-07-14 14:55:57 UTC
Comment 3 Jaroslav Tulach 2003-07-15 09:08:11 UTC
Probably done in CustomInplaceEditorTest. Tim, can you set milestone and tell me which test is doing this?
Comment 4 _ tboudreau 2003-07-15 12:33:03 UTC
It's implicit in all of the tests - they cannot be tested without being set to render a property correctly, which means connect() is called. If connect is broken for an inplace editor, it will definitely be caught by the test that fakes keyboard events.