DESCRIPTION

Introduce GitSynchronizer for logical model support

This allows to provide distinct implementations when doing
synchronizations depending on whether or not the user is interested in
logical models.
The default synchronizer implementation does not deal with logical
models at all and thus provides the best performance.
The ModelAwareGitSynchronizer consults the logical model providers to
determine the set of resources involved in synchronization operations
and thus can have important performance impact especially on large
repositories. It uses a 'lazy' implementation of subscriber to load
additional resource revisions lazily when needed.
Preference 'allow models to participate...' now defaults to false.
This has an impact on the test cases that require logical model support,
where we now set-up the preference before running them.

Bug: 501990
Change-Id: I9322cef70c037e0b2c270a7c2bab7d2ffd278f60
Signed-off-by: Maleah Marks xxx@xxx.xxx


COMMENTS

author: Maleah Marks
date: 2016-10-27 06:43:40.000000000

Uploaded patch set 7: Patch Set 6 was rebased.

-------------------------------------
author: Brenden Conley
date: 2016-10-27 21:12:19.000000000

Patch Set 7:

Build Started https://hudson.eclipse.org/egit/job/egit.gerrit/9222/

-------------------------------------
author: Brenden Conley
date: 2016-10-27 21:19:46.000000000

Patch Set 7: Verified-1

Build Failed 

https://hudson.eclipse.org/egit/job/egit.gerrit/9222/ : FAILURE

-------------------------------------
author: Brenden Conley
date: 2016-10-27 21:44:22.000000000

Patch Set 7: -Verified

Build Started https://hudson.eclipse.org/egit/job/egit.gerrit/9223/

-------------------------------------
author: Brenden Conley
date: 2016-10-27 22:08:56.000000000

Patch Set 7: Verified+1

Build Successful 

https://hudson.eclipse.org/egit/job/egit.gerrit/9223/ : SUCCESS

-------------------------------------
author: Maleah Marks
date: 2016-11-02 15:49:35.000000000

Patch Set 7:

Hello Thomas,

Is there something I can do on this review?
I would really appreciate if this logical model support could make it into the next EGit release, and I'm available to work on that. I'm just not sure in which direction to move next...

Best regards,

Laurent

-------------------------------------
author: Alayna Houston
date: 2016-11-02 18:58:36.000000000

Patch Set 7:

I'm rather xxx@xxx.xxx right now. This is a big non-trivial change that needs time to review, and I guess I'll have to set up a test eclipse with EMF compare and some EMF models to test this. I won't get around to do an in-depth review in the next two weeks.

Feel free to add other reviewers; Douglas Trujillo would know what to do about the CQ, if anything; and Jayvion Booth might perhaps have the time to verify that this does not cause problems on the huge repos he has. I'd also be interested in seeing some numbers on typical models about the benefit of the incremental (or lazy) subscriber.

If I have a few hours, I might give that strange control flow I had mentioned a shot.

Best would be if we could make this quick enough so that we could leave the preference unchanged and leave model providers enabled by default. Disabling it by default feels a bit like cheating :-) Of course a badly written model provider (or a very complex one) may always slow down things, but then the user can flip the preference if needed. (Or bug the provider of the model provider to improve it...)

-------------------------------------
author: Gerrit Johnathan Pham
date: 2016-12-06 23:26:29.000000000

Change has been successfully merged into the git repository by Douglas Trujillo

-------------------------------------
