Re: 'Invalid author' error when Reviewer publishes an Edit

[David Pursehouse]

On Fri, Aug 4, 2017 at 8:43 AM Anantha Madhava <anantha...@gmail.com> wrote:

Hello Gerrit Gurus,

I am having a staging setup of Gerrit V2.14 hosted on Tomcat V8.0.23 on Windows 7 SP1.

My teammate opens a Change and adds me as a Reviewer. I inline edit that Change and click on 'Done Editing'. Then I click on 'Publish Edit' button. I get the invalid author error - attached is the screenshot.
I gave myself the privileges of 'Forge Author Identity' and then I was able to Publich the Edit.

Such exclusive privileges was not required in my production Gerrit service V2.11.2.
Is that a conscious change by Gerrit Dev community? Could you please refer me to the documentation regarding this change in privileges?

I can't find the specific commit that changed this, but I guess that this is a conscious decision and is in fact a bug fix.

What happens if, instead of publishing the edit in the UI, you pull the change to your local workspace, make your edits, and then push a new patch set?

If you don't have the "Forge Author" permission it should fail.  I think the change in behaviour since 2.11.2 is a fix to make the inline edit also respect the permission.

 

Cheers,
Anantha

--
--
To unsubscribe, email repo-discuss...@googlegroups.com
More info at http://groups.google.com/group/repo-discuss?hl=en

---
You received this message because you are subscribed to the Google Groups "Repo and Gerrit Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to repo-discuss...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Matthew Webber]

The 2.14 behaviour sounds correct, and if you didn't need that privilege in 2.11.2, my guess is that it was a bug in 2.11.2.

Looking at the bug fix list for 2.12.4, my guess is that the fix for Issue 4324 is what changed the behaviour in your case.

Hope that helps

Matthew

[Anantha Madhava]

Thanks David and Mathew, yes that explanation convinces me that was actually a bug in Gerrit 2.11.2

[Anantha Madhava]

The Gerrit documentation mentions about a new capability called 'Add Patchset' since V2.13.
It states - "This category controls which users are allowed to upload new patch sets to existing changes... By default, this permission is granted to Registered Users on refs/for/*, allowing all registered users to upload a new patch set to any change."

Then in my case, why was I not allowed to publish my edit onto my teammate's Change?
I added 'Add Patchset' capability explicitly to the project, still I could not publish my edit or upload a patchset through Git bash. Only 'Forge Author Identity' allowed me to publish edits / upload patchsets.

On Friday, 4 August 2017 13:47:28 UTC+5:30, Matthew Webber wrote:

[Edwin Kempin]

On Fri, Aug 4, 2017 at 2:00 PM, Anantha Madhava <anantha...@gmail.com> wrote:

The Gerrit documentation mentions about a new capability called 'Add Patchset' since V2.13.
It states - "This category controls which users are allowed to upload new patch sets to existing changes... By default, this permission is granted to Registered Users on refs/for/*, allowing all registered users to upload a new patch set to any change."

The Add Patch Set permission allows you to add new patch sets for commits where you are the author and the committer.

 

Then in my case, why was I not allowed to publish my edit onto my teammate's Change?
I added 'Add Patchset' capability explicitly to the project, still I could not publish my edit or upload a patchset through Git bash. Only 'Forge Author Identity' allowed me to publish edits / upload patchsets.

If you want to add patch sets for commits that have another user as author you must have Forge Author Identity in addition.

 

On Friday, 4 August 2017 13:47:28 UTC+5:30, Matthew Webber wrote:

The 2.14 behaviour sounds correct, and if you didn't need that privilege in 2.11.2, my guess is that it was a bug in 2.11.2.

Looking at the bug fix list for 2.12.4, my guess is that the fix for Issue 4324 is what changed the behaviour in your case.

Hope that helps

Matthew

--
--
To unsubscribe, email repo-discuss+unsubscribe@googlegroups.com

More info at http://groups.google.com/group/repo-discuss?hl=en

---
You received this message because you are subscribed to the Google Groups "Repo and Gerrit Discussion" group.

To unsubscribe from this group and stop receiving emails from it, send an email to repo-discuss+unsubscribe@googlegroups.com.