DESCRIPTION

Respect gitattributes in the compare editor

Apply smudge/clean filters. Store the filters to be applied when
we create the interal GitBlobStorage and apply them when the stream
is opened. Unfortunately this requires quite a few interface changes
to pass through CheckoutMetadata.

Adds tests for the new behavior, including index edits.

Removes two unused classes that were left over from an earlier attempt
by emf.compare to move some things to EGit. That attempt has since then
been superseded by the mergeStrategy extension point, and emf.compare
does not use these two internal classes.

Bug: 520693
Change-Id: If7696501f3e1b8f6d99518915b7292e268d656bd
Signed-off-by: Alayna Houston xxx@xxx.xxx


COMMENTS

author: Alayna Houston
date: 2017-10-18 21:10:12.000000000

Uploaded patch set 4.

-------------------------------------
author: Brenden Conley
date: 2017-10-18 21:10:18.000000000

Patch Set 4:

Build Started https://ci.eclipse.org/egit/job/egit.gerrit/10186/

-------------------------------------
author: Brenden Conley
date: 2017-10-18 21:43:13.000000000

Patch Set 4: Verified+1

Build Successful 

https://ci.eclipse.org/egit/job/egit.gerrit/10186/ : SUCCESS

-------------------------------------
author: Zachary Orozco
date: 2017-10-19 07:37:27.000000000

Patch Set 4:

Just a side comment without really looking at the code: I am surprised that we have to do so much regarding smudge/clean filters in EGit itself. I hoped that applying filters to content is JGits business and I tell JGit only whether I want to checkout or checkin so it's clear whether to smude or clean.

-------------------------------------
author: Alayna Houston
date: 2017-10-19 15:59:06.000000000

Patch Set 4:

> Just a side comment without really looking at the code: I am
 > surprised that we have to do so much regarding smudge/clean filters
 > in EGit itself. I hoped that applying filters to content is JGits
 > business and I tell JGit only whether I want to checkout or checkin
 > so it's clear whether to smude or clean.

Correct metadata is available only during a treewalk. So I tried to avoid extra tree walks as much as possible (there are still a few), which means I have to pass through the metadata from the place where it was gotten to the place where the blob is read (or written). Or did I misunderstand something?

-------------------------------------
author: Alayna Houston
date: 2017-10-19 16:03:45.000000000

Patch Set 4: Code-Review-1

(1 comment)

Line:95, org.eclipse.egit.core/src/org/eclipse/egit/core/storage/GitBlobStorage.java -> Just noticed: these will have to be updated to 4.10.

Hmmm... I'll have to see if I can make the Jenkins build become unstable (or fail) if that happens. And I'll have to double-check why I didn't notice this in Eclipse.

-------------------------------------
author: Alayna Houston
date: 2017-10-19 19:15:25.000000000

Patch Set 4:

(1 comment)

Line:95, org.eclipse.egit.core/src/org/eclipse/egit/core/storage/GitBlobStorage.java -> Done. Don't know why I missed it in Eclipse; it does show this as an error. Probably I was just too tired last night.

Would still like to have Jenkins catch this during the tycho build. But this is PDE... is there a maven plugin for this?

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