
| Key: |
RHQ-693
|
| Type: |
New Feature
|
| Status: |
Accepted
|
| Priority: |
Major
|
| Assignee: |
Unassigned
|
| Reporter: |
Greg Hinkle
|
| Votes: |
0
|
| Watchers: |
0
|
|
If you were logged in you would be able to see more operations.
|
|
|
| Date of First Response: |
21/Jul/08 10:31 PM
|
|
This mechanism allows the plugin to supply overrides for the list of options of a property definition. For example, if your configuring a jms topic and want to give the user a drop down of the datasources that are available to use as persistent stores. Anything were the list of options is completely dynamic or should be restricted by logic given the context. The communications with the plugin should include the entire configuration state, likely after each property is altered or on specified dynamic properties to allow a model similar to the javascript form selection for interdependent fields. (should this also support field disablement?)
|
|
Description
|
This mechanism allows the plugin to supply overrides for the list of options of a property definition. For example, if your configuring a jms topic and want to give the user a drop down of the datasources that are available to use as persistent stores. Anything were the list of options is completely dynamic or should be restricted by logic given the context. The communications with the plugin should include the entire configuration state, likely after each property is altered or on specified dynamic properties to allow a model similar to the javascript form selection for interdependent fields. (should this also support field disablement?) |
Show » |
|
some thoughts / questions:
* if we don't support cascading rules, then it's quite easy to cache the dynamic results after they are gotten from the agent the first time. the page will only be slow on initial load, and after that the user can edit lists/maps as well as perform any other multi-page flows without needing to talk to the agent again.
* is there a way to pre-load this data before the user goes to the config ui page, maybe at import-time, or some other sufficiently advanced time? i'm guessing no because dynamic properties can be based off of just about anything right?
* is there a way to make this a complete server-side implementation, where the user given access to some sort of restricted data model through a simplified, dynamic config property DSL (similar to what the embedded CLI might look like)?
* how should this work if the agent is down, or the manage resource that the agent needs to talk to is down? in this case, i think the UI would either fail to render, or would render an empty list - both of which aren't very helpful...unless of course we could pre-fetch this data when we know it was available, or if the implementation was completely server-side.