DESCRIPTION

Click on search result causes EGit to stat the repository
directory

Bug: 501799
Change-Id: I45efcbbd54bbfbb1e9acf2c19371cba2c9c0d1e4
Signed-off-by: Marisol Rowland <Kassandra Newman>

COMMENTS

author: Brenden Conley
date: 2016-09-26 09:13:17.000000000

Patch Set 4:

Build Started https://hudson.eclipse.org/egit/job/egit.gerrit/9096/

-------------------------------------
author: Brenden Conley
date: 2016-09-26 09:37:50.000000000

Patch Set 4: Verified+1

Build Successful 

https://hudson.eclipse.org/egit/job/egit.gerrit/9096/ : SUCCESS

-------------------------------------
author: Alayna Houston
date: 2016-09-26 10:20:04.000000000

Patch Set 4: Code-Review-1

(4 comments)

I don't quite see why delaying getting the repo until we visit the first project node would help. A POST_CHANGE delta always has at least one project node. But maybe I overlooked something?

Line:144, org.eclipse.egit.core/src/org/eclipse/egit/core/internal/indexdiff/GitResourceDeltaVisitor.java -> The hierarchical delta will be

  ROOT
    PROJECT
      (FOLDER)*
         FILE

Previously, the repository was gotten up front before
entering ROOT. Now it is gotten for most repositories
(all those that don't have submodules) on the first
PROJECT node.

I fail to see what this buys us.

Line:5, org.eclipse.egit.core/src/org/eclipse/egit/core/internal/indexdiff/RepositorySupplier.java -> This would be Copyright (C) 2016, Marisol Rowland.

Neither Jens nor Markus nor Robin nor I have coded anything here :-)

Line:729, org.eclipse.egit.core/src/org/eclipse/egit/core/internal/indexdiff/IndexDiffCacheEntry.java -> Doing this after visiting the delta means the delta visitor needs to be prepared to deal with getting null from the supplier. And the visitor should probably short-circuit if it ever detects this. Makes no sense to traverse a delta when there's no repository anymore.

Line:140, org.eclipse.egit.core.test/src/org/eclipse/egit/core/internal/indexdiff/IndexDiffCacheTest.java -> Please explain why you made this change.

-------------------------------------
author: Marisol Rowland
date: 2016-09-26 12:16:13.000000000

Patch Set 4:

(3 comments)

Line:144, org.eclipse.egit.core/src/org/eclipse/egit/core/internal/indexdiff/GitResourceDeltaVisitor.java -> Yeah. Right. Will revert to SkipNotInterestingDeltaVisitor then. It will be pretty fast on big project - it just executes boolean checks against incoming delta structure.

Line:5, org.eclipse.egit.core/src/org/eclipse/egit/core/internal/indexdiff/RepositorySupplier.java -> Done

Line:140, org.eclipse.egit.core.test/src/org/eclipse/egit/core/internal/indexdiff/IndexDiffCacheTest.java -> ignored folders are part of "not interesting" notification.

-------------------------------------
author: Marisol Rowland
date: 2016-09-26 13:15:46.000000000

Patch Set 4:

> (4 comments)
 > 
 > I don't quite see why delaying getting the repo until we visit the
 > first project node would help. A POST_CHANGE delta always has at
 > least one project node. But maybe I overlooked something?

I just realized, that issue not in the way we iterate over resource delta. But more in RepositoryCache. For example for every getRepository prune() is called. Which checks every repository.

Have you considered java.nio.file.WatchService for file system monitoring? I could fix RepositoryCache instead.

-------------------------------------
author: Gerrit Johnathan Pham
date: 2016-10-11 16:47:52.000000000

Change has been successfully merged into the git repository by Alayna Houston

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