it depends on what ConfigBean these REST APIs are mapped to. For the most part the REST API to alter the ConfigBeans directly but there are cases (like deployment) where you don't want to just alter the configuration, you want to deploy so it makes more sense to redirect the REST APIs to a command class rather than changing the config bean directly.
which ones were creating issue for you ?
jerome
On Apr 14, 2010, at 11:58 AM, Jason Lee wrote:
> I've been working with the REST API, trying to see what needs to be done to port the Admin Console from AMX to REST, and I've hit a snag. The REST API seems to...wrap? delegate to? the CLI command classes, and pull attribute information from their parameters. These parameters are all lower case. AMX, however, as a layer on top of JMX, returns the JMS properties in mixed case, which is what the console uses (in practical terms, we deal with the data as a Map). I know this was covered a long time ago, but why are we using the command classes and not the ConfigBeans directly?
>
> There are likely more, but here are the solutions for the Console's problems as I see it:
>
> * The REST API is altered to use the ConfigBeans directly. This solves the case issue, but may break any existing clients (are there any)? This is a bit distasteful as it's a fairly significant change to solve the problem of one client. On the flip side, though, I do prefer the mixed case when it comes to reading the output manually, and it might also make it easier to generated "pretty" wrappers should someone be so inclined (e.g., validateatmostonceperiod vs ValidateAtMostOncePeriod could get getValidateatmostonceperiod vs getValidateAtMostOncePeriod).
>
> * Another solution, of course, is to change the admin console, but that will be a lot of tedious, error-prone work. We'd love to be able to reimplement our Handlers and let the pages stand as they are. That, of course, may not be possible, but that's what we'd like to do. :)
>
> * Ken and I briefly batted around some sort of wrapper class, such as a case-insensitive Map implementation. This looked pretty thorny to get right off the cuff. I don't think we've written it off just yet, but it's lower on my list, at least.
>
> Any other thoughts or suggestions? Am I missing something obvious? Can anyone shed some light on why the REST API behaves as it does (or correct my understanding of it)?
>
> --
> Jason Lee
> Senior Member of Technical Staff
> GlassFish Administration Console
>
> Oracle Corporation
> Phone x31197/+1 405-343-1964
> Blog http://blogs.steeplesoft.com
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: admin-unsubscribe_at_glassfish.dev.java.net
> For additional commands, e-mail: admin-help_at_glassfish.dev.java.net
>