2005/12/19 12:19:52 fs 1.1.2.14: #i53095# extended implementation to work with an XIntrospectionAccess component
2005/11/02 12:17:17 fs 1.1.2.13: #i10000# exception specifications
2005/11/02 11:43:43 fs 1.1.2.12: #i10000# exception specifications
2005/10/25 12:09:28 fs 1.1.2.11: #i53095# cache the enum converters
2005/10/25 07:13:12 fs 1.1.2.10: #i53095# knitting lose ends (amongst others, make the handlers available as service)
2005/10/24 08:41:53 fs 1.1.2.9: start making the handlers full-fledged components, with using a new infrastructure replacing extensions/source/inc/componentmodule.*
2005/10/17 14:09:32 fs 1.1.2.8: #i53095# some cleanup of remaining TODOs
2005/10/13 13:01:04 fs 1.1.2.7: #i53095# introduce an XObjectInspector/Model
2005/10/11 13:29:22 fs 1.1.2.6: #i53095# phase 3:
introduced XPropertyHandler and XObjectInspectorUI
same open issues as in previous phase
(plus probably some more, since not everything is tested, yet :-\)
2005/09/05 07:41:51 fs 1.1.2.5: #i53095# phase 3, part 1: introduced XPropertyControl and relatives,
describing one control in the ObjectInspector, responsible for one
property
known issues:
- rebuildPropertyUI can cause problems now: If the user clicks into
the control for property A, which causes property B to be committed,
which causes the UI for property A to be rebuilt, then this will
crash currently. Reason: rebuildPropertyUI now synchronously replaces
the VCL-Window of the rebuilt control, which is exactly the one
which is still in some MouseButtonDown-handler.
possible solutions:
- see if rebuiltPropertyUI can be obsoleted - handlers should be able
to just obtain the XPropertyControl from the PropertyUI, and
re-initialize the control. Shouldn't they?`
- make one of the steps in the chain (mouse-click, handler-call,
rebuildPropertyUI-callback) asynchronous.
2005/08/18 12:44:31 fs 1.1.2.4: #i53095#, phase 2
moved (nearly) all property handling to dedicated handlers, the controller is
now simply managing a set of handlers
open issues for making the property browser completely generic:
- target page for a property - at the moment, the pbrw uses form-specific
knowledge
- relative position of properties. Again, the pbrw uses the OPropertyInfoService
which is not generic
- isComposeable for a given property. Also OPropertyInfoService-dependent ATM
- help ids of pages and the pbrw as a whole. They're hard-coded at the moment
other open issues:
everything in the code which is tagged with TOD/UNOize. Those are items which
do not immediately hinder phase 3 (real UNOization, i.e. definition of new
UNO interfaces for the handlers, the controller, and so on), but need to be
addressed in phase 4 (knit lose ends)
2005/08/12 16:30:12 fs 1.1.2.3: - more fine-grained control in the IPropertyBrowserUI which elements
to enable or disable
- moved designing the SQL command into a dedicated handler
- some more reactions on actuating properties move to dedicated handlers
- *nearly* completed implementation of the "composed browser UI", which
collects and combines UI change requests (IPropertyBrowserUI)
(still missing: proper auto-firing)
2005/08/10 15:41:46 fs 1.1.2.2: #i53095#
get rid of nearly all [1] the implementations in OPropertyBrowserController::Clicked,
and move them to a FormComponentHandler
[1] still to migrate:
- browsing for events (needs a dedicated event property handler)
- handling for clicking the button of the Command property - this
is kind of asynchronous, and IPropertyHandler is not yet prepared for this
2005/08/09 14:00:02 fs 1.1.2.1: #i53095# phase 1:
- don't use strings to transver values between controls and introspectee, but Anys
- first version of a dedicated property handler for form-component-related properties
(not yet completed)
known regressions over previous phase:
- handlers for events not yet implemented, thus some assertions
- click handlers for form-component-related properties do not yet work,
thus the browse buttons mostly do not work
