Community
Participate
Working Groups
I accidentally had a JPA 2.0 persistence.xml in a project that was set to a JPA 1.0 provider. As a result, there were a couple IllegalArgumentExceptions in the .log from BaseJpaPlatformUi.getResourceUiDefinition() whenever I browsed to the JPA content node of the project. I found my problem by debugging and fixed it via the provider setting. I'm not sure any user would have been able to figure this out, but at least we should not be throwing uncaught exceptions. Instead, can we either catch the exception or just log a short error message from this method?
We're looking at some minor changes that could enhance the usability here, since this isn't an unlikely error state to be in. Trying to come up with a fix that would be suitable for RC2. Thanks for reporting this Tim.
Retitling this and splitting the error logging into another bug, as that is potentially destabilizing for RC2. (see bug 313632)
Created attachment 169233 [details] proposed patch Changed the validation for GenericRootContextNode to check for unsupported as well as unrecognizable (invalid) content. Placed the validation error on the file itself if the file exists. Made similar changes for orm.xml validation. Made notes in BaseJpaPlatformUI with respect to bug 313632.
One further note. The patch gives user feedback in the case he falls into this state, but we decided not to address the error logging issue. The exception is caught and logged, so we don't have spare exceptions flying around. But removing the exception required adding a null implementation of the ResourceUiProvider or various other null checking that could cause destability.
I've tested this patch, looks good to me
* Explain why you believe this is a stop-ship defect. Or, if it is a "hotbug" (requested by an adopter) please document it as such. This problem is likely to cause quite a bit of user confusion, possibly to the point where they may recreate a project to get out of the described error state. It is also likely that users will encounter this problem as Tim did. * Is there a work-around? If so, why do you believe the work-around is insufficient? Work around is to change the facet version and platform, but the user likely won't figure this out. * How has the fix been tested? Is there a test case attached to the bugzilla record? Has a JUnit Test been added? Fix has been manually tested by Paul, Karen, and myself. * Give a brief technical overview. Who has reviewed this fix? See comment 3. Karen and I have reviewed the fix. * What is the risk associated with this fix? Medium Low. Generally we wouldn't want to add validation at this point in the release, but this seems to be a particularly nasty error state that is not currently detected in a meaningful way to the user. Given this risk, extra testing has been performed.
Well, if Tim did it, it must be easy to make this error :) ... so, I'm good.
Patch committed and released last night.
verified in RC2
