Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.
Bug 320496 - EMF Compare in fails in Eclipse3.6 passed in eclipse3.5.
Summary: EMF Compare in fails in Eclipse3.6 passed in eclipse3.5.
Status: CLOSED FIXED
Alias: None
Product: EMFCompare
Classification: Modeling
Component: Core (show other bugs)
Version: 1.3   Edit
Hardware: All All
: P1 blocker (vote)
Target Milestone: ---   Edit
Assignee: EMF Compare CLA
QA Contact:
URL:
Whiteboard:
Keywords:
: 330907 (view as bug list)
Depends on:
Blocks:
 
Reported: 2010-07-21 08:30 EDT by Rahul Yadav CLA
Modified: 2013-09-12 09:02 EDT (History)
7 users (show)

See Also:


Attachments
stack trace of exception thrown by emf comapre code (2.36 KB, text/plain)
2010-07-21 08:33 EDT, Rahul Yadav CLA
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description Rahul Yadav CLA 2010-07-21 08:30:03 EDT
Build Identifier:  I20100603-1500

In our product we compare models we were earlier using eclipse3.5 and now moving to eclipse3.6. Eclipse3.6 fails with NPE on opening Compare view on two files this was working absolutely fine with eclipse3.5. I have checked out the source code for EMF compare shipped with eclipse3.5 and 3.6 and on comparison i found that method checkReferencesUpdates(DiffGroup root, Match2Elements mapping) in class org.eclipse.emf.compare.diff.engine.check.ReferencesCheck is modified and a call to checkContainmentReferenceOrderChange() method is added in While loop which was not there in eclipse3.5. This new method is throwing NPE, if i comment this method call code works fine.

Reproducible: Always
Comment 1 Rahul Yadav CLA 2010-07-21 08:33:05 EDT
Created attachment 174841 [details]
stack trace of exception thrown by emf comapre code
Comment 2 Denis Roy CLA 2010-07-21 09:08:41 EDT
Putting this in the EMF bucket, not the forums bucket.
Comment 3 Cedric Brun CLA 2010-07-21 09:43:13 EDT
Thanks for your feedback, we'll fix this issue ASAP.
Comment 4 Rahul Yadav CLA 2010-07-23 09:00:40 EDT
Can you give us some information on when can we expect a fix as we are on verge of release if not...is there any workaround till then?
Comment 5 Laurent Goubet CLA 2010-07-26 05:34:20 EDT
Hi Rahul,

Unfortunately, I can't propose any workaround on this, we'll have to fix the bug before you can use the latest releases. Take note that installing the old version of EMF Compare in 3.6 should work if you are desperate.

Laurent
Comment 6 Laurent Goubet CLA 2010-07-30 04:21:02 EDT
Hi Rahul,

I've fixed the potential NPE; the fix will be available in the next integration build. However I am unsure has to how you got to that point. Could you provide us with the models you were comparing so that I can try and see if there isn't another NPE lurking behind?
Comment 7 Rahul Yadav CLA 2010-07-30 12:35:59 EDT
Thanks for the fix. We have a release in couple of days...Any idea when can we get the integration build?Can we get access to the repository were you have fixed it, may be we can checkout the code and use until we get a integration build.
Comment 8 Laurent Goubet CLA 2010-07-30 12:44:33 EDT
Rahul,

Easiest way to check out the code is to follow the instructions on http://wiki.eclipse.org/Setting_up_a_development_environment_for_EMF_Compare#Using_Project_Set_File_.28PSF.29 .

As for the next integration build, I was supposed to get it done today, but haven't had enough time to. It should be done by tomorrow noon (CET).
Comment 9 Laurent Goubet CLA 2010-08-02 05:40:48 EDT
Rahul,

Sorry for the delay, we've got some trouble with building EMF Compare :/. You can access one of the latest nightly on https://build.eclipse.org/hudson/view/Modeling/job/emf-compare-1.1/lastSuccessfulBuild/artifact/EMF.Compare.downloads/ , but we can't promote these builds until some bugs on Eclipse side are fixed.
Comment 10 Rahul Yadav CLA 2010-09-06 10:32:48 EDT
Laurent,
  Thanks for the quick response and the fix. we are able to use the nightly build code and resolve the concerned NPE issue. 
Is there going to be any official release of EMF compare after this? if yes, when can we expect it.
Actually i am into some more Compare related issues which seems to be due to EMF compare...if there going to be an official release of EMF compare we can give it a try and narrow our investigation.
Comment 11 Laurent Goubet CLA 2010-09-06 10:40:33 EDT
Rahul,

We've switched back to the old build system fro EMF Compare until we've taken care of all the hudson related issues. Either way you can access the latest released builds of EMF Compare on http://eclipse.org/modeling/emf/downloads/?project=compare .

Most notably, both 1.1.1RC sport the fix for this issue (1.1.1 final is part of the eclipse Helios SR1 (3.6.1) release which is due September 24, 2010).
Comment 12 François Rey CLA 2010-10-11 03:40:27 EDT
I don't know is this bug is the same as bug #326194, but I post here because in that other very similar bug many are complaining and nobody seems to care.

Can a release with a working compare be made asap?
==================================================

This issue basically makes PDT useless and it's really unacceptable to not release any working build yet.
Comment 13 François Rey CLA 2010-10-11 03:52:27 EDT
(In reply to comment #12)
> I don't know is this bug is the same as bug #326194, but I post here because in
> that other very similar bug many are complaining and nobody seems to care.
> 
> Can a release with a working compare be made asap?
> ==================================================
> 
> This issue basically makes PDT useless and it's really unacceptable to not
> release any working build yet.

Oops just realized this bug is against EMF and I did not mean to pollute your space. The needed release is from PDT, so ignore my comment.
Comment 14 Cedric Brun CLA 2010-10-14 03:42:35 EDT
Even if it's not related to this bug at all I'm quite taken aback by your tone.

Anyway this bug has been fixed and the SR1 release has been made, closing this bug.
Comment 15 Laurent Goubet CLA 2011-03-10 08:42:12 EST
*** Bug 330907 has been marked as a duplicate of this bug. ***