Comment 1 Chuck Bridgham 2008-03-11 17:58:30 EDT

Fixes touch EMF2DOMSSERenderer as well to capture saves....

Comment 2 Chuck Bridgham 2008-03-11 18:13:06 EDT

Created attachment 92242 [details]
EMF2DOMSSEAdapter patch

Comment 3 Chuck Bridgham 2008-03-11 18:35:03 EDT

Created attachment 92245 [details]
ResourceSetWorkbenchEditSynchronizer patch

Comment 4 Chuck Bridgham 2008-03-11 18:39:54 EDT

Save events now pass the associated EMFResource instance along to be cached in the recently saved cache.

This way we can differentiate between saves of each EMFResource instance associated with a certain IFile.



The SSEAdapter patch requires the ResourceSetWorkbench... patch applied, as new methods were added.

Comment 5 Nitin Dahyabhai 2008-03-12 17:19:45 EDT

Thanks for the patch, Chuck.  What happens to older code that does not call the new method?

Comment 6 Trung 2008-03-17 14:27:10 EDT

Created attachment 92728 [details]
For 'org.eclipse.wst.common.emfworkbench.integration'

The patch for ResourceSetWorkbenchEditSynchronizer causes a serious error.
Basically, it unloads all Java related resources whenever there is a resource change in the workspace. This results in the inability to locate and load any java resource (class, java files) by EMF utils classes.

I'm still don't understand why other team (especially the JSP team) can pass the smoke test ?

Anyway, here is the patch to end this nightmare.

Comment 7 Trung 2008-03-17 14:30:23 EDT

Created attachment 92729 [details]
For 'org.eclipse.wst.common.emfworkbench.integration'

this new patch has an additional small optimization that I think will be useful for the synchronizer

Comment 8 Chuck Bridgham 2008-03-17 14:54:08 EDT

Thanks for the patches!

I am testing now - 

Comment 9 Carl Anderson 2008-03-17 15:38:11 EDT

*** Bug 222681 has been marked as a duplicate of this bug. ***

Comment 10 Chuck Bridgham 2008-03-17 15:54:09 EDT

So the original fix for this bug still works as intended, and now recognizes related changed resources.  BUT caused a major regression by unloading all JEM resources causing any scenario depending on the JEM reflectoin api's to fail.

After testing, and stepping through the patch - recommend we respin .

Carl is doing a build.

Comment 11 Chuck Bridgham 2008-04-03 11:02:45 EDT

Found an additional regression in 

ResourceSetWorkbenchEditSynchronizer.recentlySavedFilesForceRemove()

Needed to check for null resource, and remove the file from the list as before...

Dropped fix to build

Comment 12 David Williams 2008-04-24 00:43:33 EDT

mass change to add 'contributed' keyword based on bugzilla query, please correct if that's not accurate (by marking patches as obsolete and removing the 'contributed' keyword.