Community
Participate
Working Groups
Mylyn Reviews Connector: Gerrit 2.4.0.v20140903-1455 1. Open the review at https://git.eclipse.org/r/#/c/33476/ 2. In the review editor, scroll to Patch set 7 and expand: No content is retrieved. Tested on both Windows 8.1 and Ubuntu 14.04.
This review has the same problem: https://git.eclipse.org/r/#/c/32350
I can reproduce. I wonder if it has something to do with the fact that neither of those reviews has a patch set 1. Marc-Andre, do you know why those reviews start numbering their patch sets at numbers other than 1? Have earlier patch sets somehow been removed or was patch set 4 the initial push to https://git.eclipse.org/r/#/c/33476/ ?
Yeah, maybe we're assuming somewhere that patch sets can never be non-consecutive. I know that it frankly never occurred to me that that might not be the case. It looks like the contents of patch set 4 are actually from patch set 7, which would match that theory. Same thing with 32350 -- it goes to 11, which is exactly how many patches there are.
Interesting! Those patch sets were drafts so that should be a big clue.
So, my guess is that it is in PatchSetContentRemoteFactory or in the ui association in ReviewSetContentSection, since the other metadata like commit # is correct. But I can't see anything obviously wrong at an initial glance.
(In reply to comment #3) > It looks like the contents of patch set 4 are actually from patch set 7, which > would match that theory. Same thing with 32350 -- it goes to 11, which is > exactly how many patches there are. This appears to be correct. Good find! Marc-Andre, which patch sets were drafts?
I guess the first few patch sets were drafts and those are hidden from users who don't have permission.
(In reply to Sam Davis from comment #6) > (In reply to comment #3) > > It looks like the contents of patch set 4 are actually from patch set 7, which > > would match that theory. Same thing with 32350 -- it goes to 11, which is > > exactly how many patches there are. > > This appears to be correct. Good find! > > Marc-Andre, which patch sets were drafts? I added you to the review (33476) and Miles as well. Patch 1, 2, 3 are drafts.
I still can't see those patch sets, but being able to see them might make this harder to investigate.
Created attachment 248550 [details] mylyn/context/zip
The problem is that we're using the patch set number to index an array which contains only the non-draft patch sets. This will be fixed by https://git.eclipse.org/r/#/c/36532/
Thanks very much for the contribution, Nicholas!