Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
[dtp-dev] reExecuteScript extension point

DTP-Dev,
Any plans of giving a Derby reference implementation of org.eclipse.datatools.sqltools.result.reExecuteScript extension point ?

I am facing few issues here

1. I can re-execute ad-hoc sql queries (using the OperationCommand instance supplied to the IReExecutionRunnable) coming from either sql scrapbook editor or DTP Launch Configuration for Ad-hoc sql. But for Stored Procedures re-execution OperationCommand doesn't look to be sufficient info. I suppose SP re-execution would need Launch Configuration or ProcIdentifier. A Launch Shortcut could have come handy. There seems to be no association between the sqltools results framework & the debug framework. Is this a design limitation of org.eclipse.datatools.sqltools.result plugin ? Are there workarounds ?

2. From an end-user point of view, I think the execution history is not intuitive. More than one place could be perceived as "history" by a new user
a) Results view
b) whats persisted in the LaunchConfiguration Dialog (Run--> Run... / Debug...)
c) & the tool bar Run / Debug dropdowns is also a 'History'.
Should their History items not be in synch ? Since the DTP Launch Configs make use of Results View to display the results, re-execution from Results View should allow the user to view / modify the Launch Configuration.

3. The reExecuteScript extension point, expects a 'consumer_name' element. To an extender it's not obvious from extension pt. doc. who a consumer is and what it wishes to consume. Internally the ReExecutionRegistryReader would form IReExecutionRunnable from the extension point only if the IResultInstance.OperationCommand.getConsumerName().equals(consumer_name). The sql scrapbook editor hard codes the OperationCommand's consumer name to 'SQL Editor'. The Routine Editor & SPLaunchConfigurationDelegate creates a CallableSQLResultRunnable which hard codes it to 'Routine Editor'.
So these seem to be the only 2 possible literals for consumer_name.
Whats the intention here ?

I think re-execution from history is an important feature, more so because 're-execution' from editors & LCD creates a new row in history, which IMHO is annoying.

Thoughts ?

Regards,
Raghu


Back to the top